Eurotelcorama 05 300104

DAIWA
DAIWA INSTITUTE OF RESEARCH LTD.
30th January 2004 (No. of pages: 109)
EuroTelcorama
Issue 5: Euro wireline 2004, fragmentation is the name of the game
James Enck
(+44) 20 7597 8455 「日本語可」
james.enck@dir.co.uk
Jacqueline Millan
(+44) 20 7597 8460
jacqueline.millan@dir.co.uk
th
(Priced as at market close on 28 January)
We explore emerging issues which we think may drive line loss and service
substitution for the European incumbents. We believe current industry and market
forecasts may not fully encompass the range of risks.
Among the topics discussed in this report, we examine information service demand
and consumption patterns in a number of markets, and also along age, gender, life
cycle and internet experience lines. We attempt to relate our findings to some
difficulties which service providers may encounter as adverse European
demographic trends unfold.
We also isolate and discuss the developing gap between access and services, which
we see as being driven by a combination of factors – demographic/generational,
technological, behavioral, political and regulatory.
In such a complex and uncertain environment, we think stock selection requires a

different approach, and in conclusion we introduce a multi-factor approach to
rating, which incorporates many of the themes discussed in this report. As a result
of putting our coverage universe through this process, we have taken the following
rating actions:
• We are upgrading Telefonica from rating 3 [NEUTRAL] to rating 2

[OUTPERFORM], and Telecom Italia ordinaries from rating 4
[UNDERPERFOM] to rating 3 [NEUTRAL].
• We are retaining our 3 [NEUTRAL] rating on BT Group, Deutsche Telekom,
Telecom Italia savings shares, and TeliaSonera.
• We are downgrading our ratings on Portugal Telecom from rating 2
[OUTPERFORM] to 3 [NEUTRAL], and France Telecom, KPN and Swisscom
from rating 3 [NEUTRAL] to rating 4 [UNDERPERFORM].
IMPORTANT
Important disclosures are provided on the last page of this report.
Global Equity Research

Contents
Section Pages
Executive summary 3-4

Introduction 5
1. The state of the sector 6 - 13
Industry revenue appears to be growing ahead of GDP, and average consumer
and business expenditure on the PSTN seems to be stabilizing. As ADSL
penetration grows, there is a stronger case emerging for its role as a revenue
replacement tool. However, MoU per channel on the PSTN continue to trend
downward, and mobile MoU per channel looks sluggish to us. A significant
proportion of minutes are simply disappearing.
2. A tale of two markets 14 - 18

An examination of French and UK ARPU reveals striking similarities in
composition despite differing usage patterns. Fewer people have mobile phones in
France, but use them more on average, though data usage is dramatically below
UK levels. Viewing the mobile and fixed markets as one, it is clear that the number
of channels per capita is rising steadily in both countries, but voice usage seems to
have peaked and is trending down.
3. Understanding the customer 19 – 41

Internet usage apparently tends to become more practical with greater experience,
though average usage does not appear to increase. More mature users and
broadband users show a heightened tendency to self-provision services. Many of
the most popular applications are wholly independent of the access provider, and
generate no incremental revenue for telcos.
4. Europe – the demographic timebomb 42 - 50

The nations of Europe are to see their populations getting significantly older, and
in some cases, smaller, over the next 50 years. Europe is also likely to become
more female. More divorce and less marriage means average household size will
decline, though household creation may increase, with many being single person
households. What are the implications for telcos?
5. Challenges for the telcos – the access layer 51 - 73

The telcos face a range of challenges to their position of power in access:
technological, behavioral, regulatory and political We highlight in particular the
rise of Big Broadband, FTTH initiatives driven by governmental agencies and
citizens’ action groups, which have goals and investment horizons intrinsically at
odds with those of telcos.
6. Challenges for the telcos – the service layer 74 - 95

A range of new applications and social trends should continue to drive a wedge
between access and services. They include VoIP in all its various flavors, social
networking communities and invitation-only communications, and darknets.
7. A multi-factor approach to stock-rating 96 – 109
We introduce an approach to rating stocks which seeks to encompass some of the
factors discussed earlier in this report, in an attempt to balance valuation issues
with other less obvious drivers of share price performance.
Acknowledgements
We would like to extend our heartfelt thanks and appreciation to Adam Arndt, Roger
Black, Christian Car, Karl Christensen, Lars Hedberg, Chris Hoar, Doug Ross, and
Mike Tsao, for their invaluable assistance and advice in the preparation of this report.
DIR EuroTelcorama 2
Executive summary
Growth and stability An examination of the state of the European telecom sector in 2003 shows revenue
reign growth for the industry above GDP trend, and growing signs of stabilizing expenditure
on PSTN services. Operator discipline since the sector collapse in 2000 has yielded
impressive improvements in productivity, asset efficiency, and cash flow generation.
The debt crisis facing the sector two years ago is now firmly behind us, and there is a
case emerging for ADSL as an effective revenue replacement tool. The rate of PSTN
line loss appears to be under control at present, which is a key issue.
Love it or hate it, it pay At an annual average ARPU of €360 for residential PSTN services alone, minimizing
the bills line loss is critical to stability of cash flows from domestic wireline businesses, which
we estimate will still account for at least 51% of 2010 group cash flow, on average, for
the companies under our coverage. The importance of this is underlined by what we
take to be a disproportionate reliance on termination revenues in mobile business units,
suggesting that underlying margins in many cases are significantly lower than headline
figures would suggest.
Usage and users are Traditional voice services (both fixed and mobile) are showing ongoing volume
changing, with sluggishness/decline, and it seems clear that a growing proportion of traffic is
unforeseen consequences migrating to other uses/behaviors, many of which do not generate revenue for the telco,
beyond perhaps the access revenues for the platform over which these uses take place.
Consumer surveys seem to strongly indicate a growing sophistication, eclecticism and
assertiveness among certain segments of the user base in finding and self-provisioning
services which may enable them to sidestep services provided by the access provider in
favor of a more tailored suite of services, some of which may deprive the telco of
traditional sources of revenues. In many cases these are younger, higher spending
customers, who we believe demonstrate a greater tendency to change suppliers or
abandon traditional services entirely. The poor long-term outlook for European
demographics may significantly complicate our ability to forecast future demand and
even likely market size.
The move toward a This growing separation of access from services presents the incumbents with a range
“dumb pipe” of complex challenges to be met if they are to avoid being gradually relegated to a
position of provider of a “dumb pipe,” to be filled with content and services which do
not contribute incremental revenue, or in fact drive outright substitution. Conversely,
we identify a number of challenges to the incumbents’ positions in the access
landscape as well. We believe both sides of this equation have the potential to
accelerate line loss and migration of traditional telco revenue streams towards low
revenue/no revenue applications.
The challenges in the access layer identified in this report are:
• Mobile substitution in access;

• Regulatory aggression in the areas of local loop unbundling and wholesale
line rental pricing and provisioning;
• Advances in alternative access technologies, namely Ethernet over cable,
viable Wi-Fi meshing solutions, and the growing synergies between VoIP
applications and Wi-Fi client devices;
• The rise of Big Broadband, initiatives designed to deliver vastly higher levels
of bandwidth to consumers than those currently on offer from telcos, at
similar or dramatically lower price points. The key instigators of the examples
we have seen to date include principally local and regional governments as
well as community activist groups, all of whom arguably have goals and
investment horizons which may inherently conflict with those of the
incumbents, who so far have abstained from participation.
DIR EuroTelcorama 3
Challenges in the service layer analyzed herein are:
• VoIP - in the enterprise, as a consumer product tied to broadband access, or as

an application independent of access provision. Whether delivered as a
commercial service or as a free self-contained community, such services are
feature-rich and allow huge opportunities for arbitrage on the part of the
consumer;
• New forms of web-based interaction, such as blogging, social networking
communities and invitation-only messaging systems;
• Darknets, or self-provisioned, self-configured communications networks
which provide secure and private communications between users on a peer-to-
peer basis.
Which way to jump? The telco has many weapons to deploy in responding to such developments, and we
have one valuable example from BT which highlights the benefits which may accrue to
the incumbents, as well as the difficult issues they face in deploying new services.
However, issues of threat awareness, culture and structure may inherently limit their
ability to respond effectively.
Uncertainties require a In light of the various uncertainties highlighted here, we conclude that even short-to-
new approach in stock medium term forecasts may be founded on poor visibility, and we are skeptical that
rating current market forecasts fully encompass all the sources of discontinuity facing the
incumbents. Arising from this complex set of issues, we propose and apply to the
companies under our coverage a new framework for evaluating investment stance. In it,
we seek to factor in issues beyond absolute and relative valuations, such as level of
technology threat, competitive and regulatory environment, corporate governance
standards, demographic exposure, and adaptability to change.
As a result of our review, we have made the following ratings actions:

On a capitalization–weighted basis, our current individual stock ratings leave us with a

[NEUTRAL] stance on the sector, and we would stress the caution with which we
believe the sector should be viewed.
DIR EuroTelcorama 4
Introduction
Fragmentation is the This time last year we published a note entitled European Wireline 2003 and beyond:
name of the game it’s a marathon, not a sprint, in which we looked at emerging issues which we believe
will serve as catalysts for telco dis-intermediation over the long term. In it we made
some statements which, at the time, probably seemed eccentric or heretical, but we
believe that newsflow in the sector over the past year has validated many of the ideas
we put forward twelve months ago. The unifying theme running throughout much of
this note is that of fragmentation – of the consumer’s pallet of services, of the patterns
of consumption across age and gender definitions, of the family unit, of the traditional
relationship between access and services, of the relationships which have heretofore
existed between telcos, as former state utilities, and units of the governments which
own/owned them.
Fixed businesses Our focus is overwhelmingly skewed towards the fixed line networks of the
generate 54% of group incumbents in our coverage universe, partly because of the breadth and depth of the
cash flow issues facing the wireline industry, but mainly because the fixed networks continue to
be such critical drivers of cash generation for the groups of which they are a part. For
the nine incumbents under our active coverage, we estimate that the domestic fixed
networks segment accounts for 54% of gross free cash flow (defined as EBITDA –
capex) on average. On our current forecasts, this number will not change dramatically
by 2010 (50.5%). Excluding the two more diversified southern PTTs (Telefonica and
PT) the average for the remaining seven players in 2003 is estimated to have been 64%,
moving to 57% by 2010. With the incumbents increasingly choosing to deploy cash in
dividend payments and share buybacks, such a critical source of cash is of central
importance, not only to attract and retain shareholders, but also potentially to fund
inorganic growth opportunities, where they exist. Any deterioration in the cash
generation profile of the fixed networks would ostensibly have a highly geared
negative effect on both group finances and investor sentiment.
Our focus is fixed Our ongoing research indicates that, on a long-term view the challenges to the stability
of the incumbents’ fixed businesses are multiplying and intensifying, rather than
stabilizing or receding. This report is not intended as a blow-by-blow, line-by-line
dissection of the sector (our team is only 1/8 the size of some of our competitors), and
is notable for its absence of discussion of issues such as the role of carrier pre-selection
and MVNOs, which are arguably better-flagged risk factors (though still interesting
and highly deserving of attention, particularly in light of the entry/impending entry of
brand extenders such as UK supermarkets, Virgin, Wal-Mart, Disney and others into
the telecoms arena).
This report is a draft Rather, as with last year’s note, we hope to use it to establish an agenda for future
work plan for this year analysis. we hope to explore less understood sources of pressure – generational, social,
demographic, technological and political – which we believe may materially accelerate
line loss (or perhaps just loss of control of the line?) and service substitution in the
PSTN network, depriving the PTTs of revenues and effectively narrowing incremental
sales opportunities. At the average annual expenditure for a residential PSTN customer
in the EU-15 of €360, the implications of an additional 2 - 3% line loss per annum for
the incumbents at an EBITDA margin of c.40% (on our estimates), across 224m access
lines, would present very serious challenges to stability, particularly for those with less
scope or desire to adapt.
Current industry/market We think that the forecasts currently circulating in the market, possibly including our
forecasts may not own fairly pessimistic assumptions, may not be anywhere near capturing the
encompass all the risks uncertainties and disruption facing the sector over the next decade and beyond. As an
example, forecasts produced in mid-2003 by Gartner Group show a CAGR decline
across Western Europe of only 0.5% in the period 2001 – 2007, with residential lines
declining at a CAGR of 0.6%, and business lines at CAGR of 0.3%, over the period.
Voice service revenue CAGR over the period was forecast at just –1.02% over the
period. We believe that the developments we highlight here may well accelerate this
process.
DIR EuroTelcorama 5
1. The state of the sector
In this section we briefly set the stage by looking at macro and micro trends in the
sector over the past three years.
Growth above GDP The EU 9th Report on Implementation of the New Regulatory Framework estimates
that the telecommunications market in the 15 EU member nations grew in value to
c.€251bn in 2003 from €242bn in 2002, for growth of 3.7 – 4.7% YoY. This compares
favorably with nominal EU-15 GDP growth of c.3.0% expected in 2003.
EU electronic communications market value (€bn)
260 251
250 242
240 231
230
220 211
210
200
190
2000 2001 2002 2003
Source: EU ninth report
While we don’t have the capacity or the space to drill down into each individual
country (even if we wanted to), we can see a similar trend implied at country level.
Looking at household final consumption figures published by UK National Statistics, it
is clear that expenditure on communications services has increased steadily since the
late 1980s, most sharply in the wake of the simultaneous internet and mobile booms of
1998 and 1999.
UK communications as a % of total household consumption 1981 – Q2 2003
2.40%
2.30%
2.20%
2.10%
2.00%
1.90% As % of household expenditure

3 yr. moving average
1.80%
1.70%
00
00
01
01
02
02
03
81
83
85
87
89
91
93
95
97
99
19
19
19
19
19
19
19
19
19
19
Q2
4
Q2
Q4
2
Q4
2
Q
Source: DIR, UK National Statistics
DIR EuroTelcorama 6
Growing share of According to this data, overall household final consumption grew at a CAGR of 5.71%
household consumption between 1990 and 2002, and 5.13% between 1999 and 2002. Communications
consumption grew at CAGR 7.21% between 1990 and 2002, and accelerated to 7.75%
between 1999 and 2002. During the 1997 – 2002 period, Communications grew its
share of consumption more than any other sector. The only segment of household
consumption to have grown faster over the 1990 – 2002 period is education, and in the
1999 – 2002 period, communications loses out to furnishings, perhaps as a reflection of
the housing boom prevailing in the UK over the period. It is important to note that
around 11% of the communications item is comprised of postal services and telephony
equipment, however, adjusting for this distortion, the underlying telephony services
component has grown by a CAGR of 7.7% in the period 1999 – 2002.
Household final consumption in the UK, shares and changes in shares

40.0%
Share of final household consumption 2002 Change in share since 1990 Change since 1997
30.0%
18.0%
20.0% 15.0%
12.5% 12.1% 11.7%
9.3%
10.0% 6.6% 5.8% 4.0% 2.3% 1.5% 1.3%
0.0%
-10.0%
-20.0%
-30.0%
ion
lth
es
ns
re
t
r
or
o
s
tel
ea
ce
ls
ge
ic
lt u
cc
ea
io
at
sp
tw
ue
rv
ho
an
ra
at
ba
uc
H
cu
an
se
rf
ic
en
ve
to
Ed
d
o
Tr
un
n
df
he
be
t
nd
an
nd
a
ain
m
ot
sa
an
s
lic
sa
n
nt
t/m
m
d
tio
od
ra
ng
ho
Co
ge
an
ea
en
au
go
co
hi
ra
as
pm
cr
st
ot
ve
al
isc
,g
Re
Re
Cl
n-
be
ui
ity
no
eq
ic
ic
ol
ld
ctr
an
oh
ho
ele
od
lc
se
A
r,
Fo
ou
te
,h
wa
gs
g,
in
sin
ish
ou
rn
H
Fu
Source: DIR, UK National Statistics
PSTN expenditure On the PSTN, long the source of concern for investors, the EU ninth report claims to
stabilizing see signs of stability in terms of average monthly expenditure in the EU-15,
particularly in the residential market, where stability seems to have crept in at the €29 –
30 per month level, but also in the business market, where the rate of decline has
slowed dramatically.
Estimated average monthly expenditure of a standard PSTN user (€)
90.0
78.5
80.0
70.3
70.0 64.0 62.7
61.0 60.7
60.0 Residential, ex-VAT
50.0 Business, incl. VAT
40.0 34.0 31.5 29.8 30.0 29.5 29.4

30.0
20.0
1998 1999 2000 2001 2002 2003
DIR EuroTelcorama 7
Is confidence growing? We might be tempted to say, based on these few shreds of evidence, that we see signs
on both a top-down and bottom-up basis that the industry as a whole may be growing
faster than the rest of the economy, and that some segments of the market which have
been cause for concern in the past may be stabilizing. This surely must be good news to
those looking for some sense of order or stability in the markets; however, we enter the
year with much of the market seemingly swept away on a wave of renewed optimism
about the prospects for the telecom sector. Judging from consensus EPS revision trends,
telecoms benefits from earnings momentum significantly better than that of the broader
market, though the PTT segment has not seen the overwhelming shift towards upgrades
seen in the mobile segment.
2004 EPS forecasts – ratio of upward to downward revisions
4.5
PT T s Mobile ST OXX 600
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Aug-03
Feb-03
Jun-03
Jul-03
Sep-03
Dec-02
Jan-03
Mar-03
Apr-03
May-03
Nov-03
Dec-03
Oct-03
Source: DIR, JCFQuant
Meanwhile, for the worst-offending companies, concerns over the long-term

implications of the financial excesses of the bull market seem to be inexorably fading,
and the focus on core businesses and productivity enhancement has delivered
demonstrable results.
Net debt/EBITDA evolution of Europe’s four “problem” PTTs
6.0
5.0
4.0 BT
DT
3.0
FT
2.0 KPN
1.0
0.0
Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 FY
2001 2002 2002 2002 2002 2003 2003 2003 2003
Source: DIR, company data
DIR EuroTelcorama 8
Turnover per employee (EUR)
6 0 0 ,0 0 0 2000 2001 2002

2 0 0 3 Q1 2 0 0 3 Q2 2 0 0 3 Q3
5 0 0 ,0 0 0
4 0 0 ,0 0 0
3 0 0 ,0 0 0
2 0 0 ,0 0 0
1 0 0 ,0 0 0
m
PN
ra
T
TI
FT
PT
ca
co
D
ne
ni
K
iss
So
fo
Sw
le
lia
Te
Te
Source: DIR, company data. N.B. data based on half-year results
EBITDA per employee (EUR)

2000 2 0 01 2002
1 8 0 ,0 0 0
2 0 0 3 Q1 2 0 03 Q2 2 0 0 3 Q3
1 6 0 ,0 0 0
1 4 0 ,0 0 0
1 2 0 ,0 0 0
1 0 0 ,0 0 0
8 0 ,0 0 0
6 0 ,0 0 0
4 0 ,0 0 0
2 0 ,0 0 0
0
m
ra
PN
TI
T
FT
PT
ca
co
D
ne
ni
K
iss
So
fo
Sw
le
lia
Te
Te
Source: DIR, company data. N.B. data based on half-year results
Reported personnel costs as % of sales

2000 2001 2002
30%
2003 Q1 2003 Q2 2003 Q3
25%
20%
15%
10%
5%
0%
om
PN
TI
up
FT
PT
a c
D
c
ro
ni
K
iss
G
fo
Sw
le
BT
Te
Source: DIR, company data.
DIR EuroTelcorama 9
Annualized EBITDA/average total assets
48%
43% BT
DT
38%
KPN
33%
SWISSCOM
28% TELIASONERA
FT
23%
PT
18%
TEF
13% TI
8%
Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003 Q3 2003
My broadband’s on fire! Broadband adoption continued its explosive growth trend in 2003, reaching an
estimated household penetration rate of 13.7% by Q3 (excluding Luxembourg, Greece
and Ireland). We believe final figures for 2003 will show household penetration of 15 –
16% when published. DSL claims a market share of net adds of c.82% across Western
Europe, with cumulative market share at c.75%. Cable (our numbers for “cable”
include fiber subscribers in Italy and Sweden) has steadily lost momentum versus DSL
since the second half of 2002. For many of the PTTs we track, ADSL uptake now
accounts for c. 7% of all PSTN lines, and late bloomers such as Swisscom and KPN
are now seeing their penetration rates converging on those of early leaders such as
Telia Sweden.
European broadband market development, 2000 – Q3 2003
90% 16%
DSL market share
80% 14%
70% Cable market share

12%
60%
10%
DSL share of net adds
50%
8%
40%
Cable share of net adds
6%
30%
4% T otal market growth
20%
rate (QoQ)
10% 2%
Broadband household
penetration (RHS)
0% 0%
2000 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2001 2001 2001 2001 2002 2002 2002 2002 2003 2003 2003
Source: DIR, from various
DIR EuroTelcorama 10
ADSL penetration of total access channels, Q1 2002 – Q3 2003
DT KPN Telia Sweden Telefonica
10.0%
Swisscom FT BT Group
9.0%
8.0%
7.0%
6.0%
5.0%
4.0%
3.0%
2.0%
1.0%
0.0%
Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003 Q3 2003
Source: DIR estimates, company data
Line loss is still under Despite the robust growth in ADSL, and ongoing concerns of mobile cannibalization in
control access, PSTN line loss has remained relatively modest compared to the carnage
witnessed in the US market in the past couple of years. This could be attributed to a
generally lower second line penetration rate in Europe compared with the US, and to
the absence of an aggressive UNE-P style wholesale access regime in Europe.
Whatever the driver, there is no visible sign of intensifying line loss at work in most of
the markets we track in Europe at present.
Sequential % PSTN line loss Q1 2002 – Q3 2003
0.00%
-0.50%
DT
-1.00%
KPN
-1.50% T elia Sweden
T elefonica
-2.00%
Swisscom
-2.50%
-3.00%
Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003 Q3 2003
Early signs of the ADSL Further offering some encouragement to observers is the growing body of evidence to
promise suggest that, if not a growth driver, ADSL can at least stem some of the revenue loss
from traditional sources. The graph below shows 2002 and nine-month 2003 changes
in PSTN calling revenues and PSTN subscription revenues, along with the total
contribution from ADSL, as reported by France Telecom. From this we calculate a net
change in the total revenues of these three sources, which shows a critical support role
being played by the ADSL component in the first nine months of 2003. To keep things
in perspective, however, it is important to remember that ADSL revenues account for
only c.4.3% of FT fixed line revenues in France, and c.1.6% of group revenues.
DSL revenues close to filling the gap – FT 2002 and 2003 comparison
600 530
395
400
200
0 2002 full year
-27
-200 2003 to Q3
-400 -246
-350 -311
-600
-800 -695 -650
PSTN calling PSTN access ADSL access Net change in
services revenue revenues revenues revenues
Voice MoU declining to Despite the encouraging signs coming from broadband, PSTN usage is another story.
sluggish We estimate that, in Q2 2003, the last period for which we have a reliable data set,
average outgoing voice minutes per channel per day on the PSTN in the eight countries
included in the chart below fell by 8% YoY, and was c.16.5% down on levels seen in
Q2 2001. While conventional wisdom in the press and much of the financial
community suggests that there is a direct transfer of these PSTN minutes to the mobile
networks, national regulator data on outgoing mobile minutes of use (though relatively
scant), is less conclusive. (We favor national regulator data on mobile usage, because
we believe the MoU figures reported by operators to be compromised by roaming and
terminating minutes, to the point where they are meaningless as a measure for
individual usage.) For the few countries where we have such data (Denmark, Sweden,
Portugal, France and the UK), we observed a c.8%YoY increase in outgoing mobile
minutes of use in H1 2003, and c.10.5% versus H1 2001. We estimate this equates to
an additional 0.43 minutes per channel per day for mobile, versus a decline of 0.76
minutes per channel per day on the PSTN for the eight countries where we track PSTN
minutes. On a more comparable basis, we present our estimates, based on national
regulator data, for the minutes per channel per day lost from the PSTN (YoY) and
minutes per channel per day gained on mobile networks in Portugal, the UK and
France quarter-by-quarter over the period Q2 2002 to Q2 2003.
Outgoing minutes per channel per day on the PSTN
18 UK
16 France
Germany
14
Sweden
12 Netherlands
10 Switzerland
Spain
8
Portugal
6
4
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3
2001 2001 2001 2001 2002 2002 2002 2002 2003 2003 2003
Source: DIR, from national regulator and operator data
Outgoing minutes per day per channel, mobile
5.0
4.5
UK
4.0
France
3.5
Denmark
3.0 Sweden
2.5 Portugal
2.0
1.5
1.0
Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2
2001 2001 2001 2001 2002 2002 2002 2002 2003 2003
Source: DIR, from national regulator and operator data
Outgoing PSTN minutes per channel lost YoY, vs. minutes gained on mobile
Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003
Fixed
UK -0.52 -0.69 -0.41 -0.43 -0.20
France -0.19 -0.33 -0.53 -0.37 -0.32
Portugal -2.45 -1.42 -1.08 -0.85 -1.18
Mobile
UK 0.24 0.17 0.16 0.11 0.20
France 0.01 0.19 0.39 0.54 0.64
Portugal -0.01 0.07 0.08 0.26 0.43
Source: DIR, from national regulator data
Takeaways – section 1
• Industry revenue appears to be growing ahead of GDP, and average consumer
and business expenditure on the PSTN seems to be stabilizing.
• Operator focus on debt reduction, efficiency and core businesses has resulted in
visibly better financial performance, and the debt bloat crisis of the post-bubble
era has passed.
• As ADSL penetration grows, there is a stronger case emerging for its role as a
revenue replacement tool.
• MoU per channel on the PSTN continue to trend downward, and mobile MoU
per channel looks sluggish to us. A significant proportion of minutes are simply
disappearing.
2. A tale of two markets
In this section we delve more deeply into the issue of telecom service consumption
patterns by means of a side-by-side comparison of two apparently very similar markets
– France and the UK.
France and the UK , so It might be instructive at this point to look at two superficially similar, yet disparate
close and yet so far markets in an attempt to gauge patterns in demand or consumption. Thanks to the
detailed market information published by the UK and French regulators (up to the H1
2003 period), we have been able to analyze some interesting elements of the respective
markets. Though virtually identical in terms of population size and GDP per capita, the
two markets share little in terms of penetration and usage trends. While fixed
teledensity levels (lines per 100 population) are very close between the two (H1 2003
UK c.59.5%, France c.57.6%), the French mobile market lags UK penetration
dramatically (the H1 2003 gap was 18.6 percentage points), probably a function of
competitive dynamics in the market (three operators in France versus five in the UK).
Outgoing minutes of use per fixed channel are significantly lower in France than in the
UK, averaging 8.6 minutes per day over the Q2 2002 – Q2 2003 period, versus 11.3
over the same period in the UK. Conversely, French mobile outgoing minutes of use
per channel per day are much higher (lower penetration overall leading to a higher
average usage?), averaging 3.99 minutes per day over this period, versus 3.02 per day
in the UK market. SMS usage remains stubbornly lower in France than in many
European countries (messaging as a proportion of French ARPU is only around 8% of
the total versus 16% in the UK), again perhaps as a function of lower penetration rates.
SMS per day per channel
1.85 UK
1.65 France
1.45
Germany
1.25
Denmark
1.05
0.85 Sweden
0.65 Switzerland
0.45 Portugal
0.25
0.05
01
01
01
02
02
02
03
01
02
03
00
20
20
20
20
20
20
20
20
20
20
12
1
Q4
2
3
3
Q
Q
Q
Source: DIR from national regulator and operator data
Different volumes, but What is interesting to note is that, in the fixed market, despite the different absolute
similar usage patterns levels of usage, the breakdown in types usage is very similar in terms of volume and
expenditure. The first of the side-by-side comparisons below shows a very high
correspondence between the various categories of ARPU, the second a similar picture
for outgoing traffic volumes. Where there are significant differences, as in national
calling ARPU, it is important to note that the UK numbers exclude the category
“other,” which includes mainly internet but also other types of non-geographic call
traffic, whereas the French data divides internet traffic into a separate category
(excluded from this analysis) and includes certain types of calls which would be
categorized as “other” in the UK data. It is surprising to find such a similarity in usage
split when usage levels and penetration rates themselves differ so drastically.
UK (left) and French (right) fixed-line ARPU breakdown
Fixed to mobile Fixed to mobile
22% 21%
Access
Access 42%
45% International
International
7%
8%
National calling ex-

mobile
Geographic calls
30%
25%
UK (left) and French (right) fixed-line call volume breakdown

Fixed to mobile Fixed to mobile
10%
11%
International
International
4%
5%
National Local
26% 58%
Geographic calls
86%
As a result of its more muted fixed line usage, French mobile outgoing traffic accounts
for a higher proportion of total outgoing fixed and mobile voice minutes than in the UK
(37% versus 29% as at Q2 2003), despite its dramatically lower mobile penetration rate,
as seen in the chart below.
France & UK mobile channel and mobile traffic shares

60% 38%
36%
58%
34%
56%
32%
France mobile channels/total
54% 30%
UK mobile channels/total
28% France mobile traffic/total (RHS)

52%
UK mobile traffic/total (RHS)
26%
50%
24%
48%
22%
46% 20%
Q4 2001 Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003
How expensive does one This may well have something to do with perceptions of relative pricing levels. Using
service appear versus the volume and channel data provided by OFTEL and ART, we backed out the
another? effective yield (revenue per minute) for fixed voice, mobile voice, and SMS, adjusting
the fixed figure for access charges to make it comparable with mobile revenue
reporting by the regulators. The two charts below present comparisons based on
multiples of one service over another. In the first case, it is apparent that the perceived
premium of mobile calling to fixed (adjusted for access costs) may be lower in France,
which may partly account for the lower fixed line usage relative to the UK. In the
second case, it appears that in the last two quarters under examination, the relative
price advantage of SMS over mobile voice became more pronounced, which may be
explained by the fact that in a more fragmented UK market, the likelihood of having to
choose between SMS and a more expensive off-net voice call is greater. Prior to the
last two quarters, however, the two markets track each other quite closely, which
makes the relatively lower French SMS usage levels somewhat harder to understand.
Perhaps it relates to lower mobile penetration in France and associated differences in
age profiles of users. However, we have no data to substantiate this.
France and UK voice yield mobile/fixed (x)
3.0
2.5
2.0
France mobile yield/fixed yield voice
1.5 UK mobile yield/fixed yield voice
1.0
0.5
0.0
Q4 2001 Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003
Source: DIR from national regulator data
France and UK mobile voice yield/SMS yield (x)
2.5
2.0
France voice yield/SM S yield
UK voice yield/SM S yield
1.5
1.0
0.5
0.0
Q4 2001 Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003
How much voice is In our view, perhaps the most interesting finding in this comparison was that, despite
enough? the vast differences in usage and penetration levels we have identified here, when fixed
and mobile are viewed as a single combined market, voice usage tracks very closely
between the two markets, and appears to be converging. In this exercise, we combined
total fixed and mobile channels in each market, and then looked at total outgoing
minutes against this number. What we found was that, while channels per capita are
rising steadily (5.9% in France and 6.5% in the UK, between Q4 2001 and Q2 2003),
the average minutes per day of outgoing use per channel is in a generally flat-to-
negative trend in France, and a profoundly negative trend in the UK.
Average channels per capita rising, but voice usage appears limited
1.50
6.90
1.45
6.70
1.40
6.50 Minutes per day per channel France
1.35
6.30 Minutes per day per channel UK

1.30
Average channels per population UK (RHS)
6.10
1.25
Average channels per population France (RHS)
5.90 1.20
5.70 1.15
5.50 1.10
Q4 Q1 Q2 Q3 Q4 Q1 Q2
2001 2002 2002 2002 2002 2003 2003
All talked out? This seems somewhat counterintuitive, as it seems logical (and consistent with
Metcalf) to assume that more people having access to more channels would drive
higher consumption overall, particularly when the additional usage opportunities
available occur in a new context which has heretofore been impossible (as with the
addition of mobility). This raises a couple of important questions to consider:
• Is consumption of voice inelastic (or is there a negative elasticity) with

increases in numbers of channels? In other words, is it possible that people
only have so much to say on average in a given day, no matter how many
opportunities they are given to talk?
• Does this mean that, as the proliferation of alternative access methods and
voice applications progresses, each will in turn be seeking to gain a share of
something which is inherently limited?
As for the first question, it is impossible to answer, but the foregoing analysis certainly
suggests that, in terms of the fixed/mobile duopoly, this may be the case. As for the
second question, if we accept that the first premise is true, then this is certainly a
possibility we have to entertain. If true, then we conclude that the only way to stimulate
meaningful growth is to generate a significant growth in the number of end users.
However, Europe has among the highest fixed and mobile penetration rates in the
world, and generally leads in the adoption of ICT globally (see table below). As we
will see later in some detail, the long term probability for a dramatic increase in
population in Europe would appear to be very low. In addition, the demographic shifts
confronting Europe may bring differences in user behavior or attitudes which are not
supportive of increasing levels of consumption.
World Economic Forum networked readiness rankings
Networked readiness Usage component Readiness component Environment
ranking index index component index
United States 1 1 3 1
Singapore 2 2 4 2
Finland 3 9 1 3
Sweden 4 5 2 6
Denmark 5 3 5 10
Canada 6 6 8 7
Switzerland 7 7 13 4
Norway 8 4 6 16
Austrialia 9 13 9 12
Iceland 10 14 16 5
Source: DIR from World Economic Forum Networked Readiness Index 2003 - 04
Who are you and what do Before we delve further into these issues, we think that we should try to gain a better
you want? understanding of end-user consumption patterns and needs. In the following section,
we examine a variety of data from the European and US markets (the latter has
produced some very interesting and informative studies relative to Europe) in an
attempt to identify the current key drivers of demand, and where possible, to speculate
on the long-term implications for European operators. Key to this, we believe, is an
examination, where possible, of variations in user profiles along demographic lines.
• An examination of French and UK ARPU reveals striking similarities in
composition despite differing usage patterns.
• Fewer people have mobile phones in France, but use them more on average,
though data usage is dramatically below UK levels.
• Viewing the mobile and fixed markets as one, it is clear that the number of
channels per capita is rising steadily in both countries, but voice usage seems to
have peaked and is trending down.
3. Understanding the customer
Here we attempt to go deeper into the various user segments, by means of a number of
recent research projects into consumer attitudes towards, and usage of, information
services and technology. We break this down by behavior-defined user groups, gender,
age and life cycle, and internet experience/broadband access.
If the data on average voice usage in European markets leaves us with a somewhat
ambivalent view of growth prospects, what further insights might we gain by drilling
down a bit deeper into the individual consumer segments by age, gender and income
levels? We explored a number of recent survey from both the US and European
markets, as well as the World Internet Project’s global study of internet usage patterns.
All yielded interesting information, which we attempt to synthesize below.
Defining user groups

We begin with some data from the US-based Pew Internet & American Life Project,
which has produced some very detailed and interesting studies on consumption. While
we acknowledge that the US is a different market to Europe, we nevertheless believe
the Pew data raises some interesting issues and questions, which are useful in this
discussion of Europe. Moreover, while the European user data available at present may
be somewhat less detailed, we nevertheless see some parallel patterns between the Pew
data and that published by EURESCOM, which we will examine later. The discussion
which follows draws on data from two studies released by the Pew project in
November and December 2003, which we mix as appropriate (occasionally including a
handful of other sources where relevant).
The November 2003 study Consumption of Information Goods and Services in the
United States, sought to break the user sample down by behavioral patterns, rather than
strict demographic criteria, though income and place in the life cycle seem to have
been important to determining the different user categories identified. The December
study focused more broadly on online activities and how they have changed over the
lifetime of the project. In the November study, the Pew researchers identified the
following eight consumer groups, as seen in the table below.
Summary of market segments identified by Pew research

Group name Description % of sample Mean age % internet Monthly spending on
users information services
Young tech elite Broad and deep users of technology, largely male 6% 22 100% $161
Older wired baby boomers Heavy tech users, mostly male 6% 52 100% $175
Wired GenXers Evenly split between men and women, broadly 18% 36 100% $169
engaged with technology
Wired senior man Small cluster of older ardent tech users 1% 70 100% $124
Young marrieds Opposites of young tech elites - modest tech use, 15% 24 66% $124
mostly women, lower income
Low tech older baby boomers Opposites of techie boomers, little tech use, mostly 21% 54 51% $124
women
Unwired young baby boomers Disposed toward technology, little free time to use 17% 39 45% $125
intensely
Low-tech elderly Mostly women, use "old media" 16% 73 12% $83
Average 100% 45 59% $122
Source: Pew Internet & American Life Project, Consumption of Information Goods and Services in the United States November, 2003
Just in this summary data, there are some interesting pieces of information:
• Three of the eight groups, c.30% of the sample, generate monthly
consumption of information services 32 – 43% higher than the average for the
sample. These three top-spending groups (young tech elite, older wired baby
boomers, and wired Generation X-ers) are largely comprised of well-educated,
affluent, mostly male users, with 100% internet penetration across all three
groups.
• The next four groups (wired senior men, young marrieds, low tech older baby
boomers and unwired young baby boomers), 54% of the total sample,
comprise a wide range of socio-economic and demographic groupings, yet
they each generate nearly identical levels of consumption, only slightly above
the average for the total sample. What is interesting here is the appearance of a
gender gap (low tech older baby boomers, young marrieds), as well as the time
impediments encountered by unwired baby boomers, who are otherwise able
and willing to use technology.
• The eighth group, low-tech elderly, demonstrates a clear gender gap in
consumption, with this predominately female segment generating a monthly
spend 32% below the average for the total sample.
Drilling down into each sub-group’s spending patterns, there is an interesting

divergence on telephony spend between the highest spending “older wired baby
boomers” and the next two highest spending groups (young tech elite and wired
Generation X-ers), which extends into cellular services as well. Conversely, these two
latter groups spend 8 – 17% more than the older wired baby boomers on internet
service, and about 7% more on cable/satellite.
Expenditure per month on information goods and services

$ per month for each group Total spent on information goods and services (wtd. avg.) Telephone Cellphone Cable/satellite Internet service
Young tech elite $161 $63 $61 $57 $28
Older wired baby boomers 175 69 66 53 24
Wired GenXers 169 60 59 56 26
Young marrieds 127 70 63 53 22
Low tech older baby boomers 124 55 59 51 22
Unwired young baby boomers 125 54 55 47 18
Low-tech elderly 82 44 37 43 na
Loyalty is hard to pin Viewed from a slightly different perspective, we see much the same pattern reflected in
down consumers’ statements about how essential certain services are to their lives. However,
it is interesting to note that, for the second-highest spending category, wired
Generation X-ers, the “stickiness” of fixed line telephony services is actually higher
than for any other group, whereas the commitment to the internet and cellular
telephone is notably weaker than among the older wired baby boomers, who spend
13% more per month on PSTN and cellular services, but less on internet. The young
tech elite expresses the least intense loyalty to the PSTN, but the highest to cellular,
and also demonstrate greater loyalty to cable TV and television generally, than do the
Generation X-ers. Thus, despite similar levels of expenditure and penetration rates of
technologies between certain user segments, the underlying preferences and
motivations appear to differ quite significantly from group to group, with some
apparent inconsistencies within individual segments.
Media preferences
% who say it would be "very hard" to give up Young tech elite Older wired baby boomers Wired GenXers The rest All
Computer^ 74% 64% 54% 25% 40%
Internet^ 68% 55% 51% 22% 39%
Cellphone^ 58% 50% 45% 31% 38%
Email^ 57% 49% 44% 23% 36%
Telephone 56% 57% 67% 63% 64%
Television 48% 50% 46% 48% 48%
CATV^ 40% 25% 34% 40% 37%
PDA^ 23% 32% 26% 15% 24%
Newspaper 12% 21% 14% 17% 19%
Magazine 11% 16% 11% 11% 13%
% of respondents 6% 6% 18% 70% 100%
^Only asked users of these services
Technology penetration rates by user segment

Group Internet Computer Cable Cellphone Satellite dish Premium channels DVD Digital video rec.
Young tech elite 100% 100% 69% 80% 25% 51% 82% 9%
Older wired baby boomers 100% 98% 68% 82% 27% 43% 67% 8%
Wired GenXers 100% 98% 66% 82% 22% 45% 64% 5%
Young marrieds 66% 78% 54% 56% 20% 40% 63% 8%
Low tech older baby boomers 51% 68% 61% 60% 25% 37% 37% 8%
Unwired young baby boomers 45% 41% 69% 62% 16% 21% 47% 6%
Low-tech elderly 12% 21% 68% 39% 14% 18% 17% 9%
Knowing what you want Whatever the explanations for differences in usage preferences and expenditure, what
and what you need is abundantly clear is that the three top-spending groups are more committed to the
internet, both economically and psychologically, than any of the other groups defined
here. This makes them a prime target market for advanced services and content, but
even here, there is little obvious consistency between the three groups as to preferences
and behavior in this environment. As the following table demonstrates, older users
(Generation X-ers and older baby boomers) who have a stated preference, prefer
independent ISPs to either the incumbent phone company or cable. In contrast, the
young tech elite is largely sure of its preferences (only 15% answered “don’t know”),
and has an overwhelming preference for cable ISPs. Whether this relates to the
American cable industry’s ability to satisfy this demanding segment’s bandwidth needs,
or rather, is a function of this segment’s greater commitment to cable TV generally (as
seen above), remains unclear. What is clear is that telco ISPs fare no better among the
elite than they do among the less-wired user groups, and the preference for them seems
to diminish the higher up the tech literacy curve we move. Furthermore, the study
indicates that among the young tech elite segment, 9% of those who used cellphones
had cancelled their wireline connection, a level three times higher than the average for
the entire sample, and well above the 7% of the sample who said they had seriously
considered canceling their wireline phone.
Types of ISP preferred by segment

Young tech elite Wired GenXers Older wired baby boomers The rest
Independent 19% 28% 31% 19%
Telephone 16% 16% 18% 18%
Cable 50% 27% 21% 18%
Don't know 15% 29% 31% 45%
Internet use and its The Pew study includes a couple of other interesting pieces of information regarding
impact on the telephone how tech-literate users of the internet modify, and are modified by, the internet
environment. Of principal interest to us here is the impact that the internet has on other
forms of behavior, particularly the use of the phone. In response to the proposition that
internet usage influences telephony usage, 31% of older wired baby boomers (who,
recall, spend 13% more than the other two high-spending groups on telephony)
responded that the internet decreases the number of calls made. This was a far higher
proportion than for any other segment, and this group also posted the lowest proportion
of responses that the internet would either have no effect on calling, or increase the
number of calls. To put this into perspective, this highest-spending group spends less
on internet services and more on telephony than its two heavy-user counterpart groups,
yet it perceives the impact of the internet on its calling habits to be much more
significant.
Perceived impact of internet usage on calling patterns

% who say the internet increases number of calls 16% 11% 10% 9%
% who say the internet decreases number of calls 24% 20% 31% 19%
% who say the internet has no impact on number of calls 60% 68% 58% 69%
Similarly, the older wired baby boomer segment appears to be the most likely to vote
with its feet, as indicated by the higher proportion who have changed ISPs at least once.
Interestingly, a higher proportion of this group either knows how to change the ISP’s
homepage, or has already done so, than that seen among the Generation X-ers. Perhaps
unsurprisingly, a significant majority of the young tech elite has had experience
changing the ISP’s homepage. Even the relatively less savvy Generation X-ers still
demonstrate a level of independence well above the average for the entire sample.
Tinkering with the interface and switching ISPs

Homepage is the same as what ISP provided 44% 54% 51% 62%
Has switched from ISP-provided homepage 64% 51% 55% 30%
Knows how to switch homepage on browser 59% 45% 48% 34%
% who have changed ISPs 51% 51% 52% 29%
% who have changed ISP more than once 60% 61% 63% 52%
Source: DIR, truncated from Pew Source: Pew Internet & American Life Project, Consumption of Information Goods and Services in the
United States November, 2003
You can give a man a fish Why should such behavior be viewed as significant? If we start with the assumption
and he’ll eat for a day, that the key for the telcos in gaining a significant piece of the broadband value chain,
but teach a man to fish beyond mere access provision, lies in being able to channel, mediate or otherwise
and he’ll eat for a lifetime influence where a broadband user goes and what he does, then the picture we see
reflected in these figures is somewhat ambivalent over the longer term. Consider the
following:
• The highest spending proportions of the user base demonstrate a much higher
than average confidence to self-configure and self-provision services, which
would appear to largely remove them from the influence of the access provider.
• Besides demonstrating a greater willingness to cancel the PSTN service, or to
embrace ISPs other than those operated by incumbent telcos, the tech elite
examined here has a greater tendency to try out new applications such as voice
on the net (16% for the young tech elite versus 8% overall).
Takeaways – section 3, so far
• The Pew study segments user groups by distinctive characteristics of age,
gender, life cycle stage and familiarity with/attitudes toward information services
and technologies.
• While the three most technologically savvy user groups spend considerably more
on info services than the average for the sample, the other groups in the study
are largely undifferentiated by expenditure, despite differing demographic
profiles and life cycle stages.
• The highest spending groups are far more likely to modify or tailor their usage
environment/interface or change suppliers, are more likely to choose a non-telco
ISP, and in some cases, far more likely to ditch the fixed line entirely.
• Judging from this data, we conclude that perhaps 70% of users will not deviate
much from the average expenditure level, for a variety of reasons. For those who
can and do spend more, they are more selective and demanding, i.e., better able
to evade the influence of the access provider.
Gender
Women are from Venus, Through all the various sources we examined in compiling this report, one of the
men are from Planet recurring themes which we encountered was the significant differences in uptake and
Geek usage patterns between men and women. It might be a sweeping generalization to say
that women are indifferent to more advanced forms of communications services, but
the available evidence suggests that they are, at least, visibly less enthusiastic users of
many. Even in terms of “basic” service consumption, there are some clear differences.
Fixed line gets the female The data for fixed line usage per day in the EURESCOM e-living survey (Wave One
vote data, covering the UK, Italy, Norway, Germany, Israel and Bulgaria) shows a
consistently heavier usage pattern among women than men across the age spectrum.
Conversely, men were more frequent users of mobile phones, making almost twice as
many calls per day across much of the age spectrum. Men under the age of 40 are also
much heavier users of the internet than women, though usage converges from that age
group onward. Also, it is interesting to note that while fixed line and internet usage
appears to be heaviest in younger age groups for both sexes, mobile usage is generally
flattish for women across the life cycle, while male usage is much more volatile
generally, and heavier in the earlier years. There is no information available on call
duration for mobile, so we have no way of knowing if men make more, but shorter,
calls, or whether women make fewer, but longer, calls. Accordingly, it is difficult to
draw much from this data, particularly in a sector such as mobile, where change since
data collection in 2001 may have been dramatic.
Minutes per day using fixed line phone, gender split, by age
50
45
40
35
30
25
20
15
10 Men
Women
5
0
85+
16 - 19
20 - 24
25 - 29
30 - 34
35 - 39
40 - 44
45 - 49
50 - 54
55 - 59
60 - 64
65 - 69
70 - 74
75 - 79
80 - 84
Source: DIR, from EURESCOM e-living project, 2002
Mean number of mobile calls per day, gender split, by age
7
Men
6
Women
5
80+
16 - 19
20 - 24
25 - 29
30 - 34
35 - 39
40 - 44
45 - 49
50 - 54
55 - 59
60 - 64
65 - 69
70 - 74
75 - 79
Mean minutes per day using the internet at home, gender split, by age
100
90
Men
80
Women
70
60
50
40
30
20
10
0
65+
16 - 19
20 - 24
25 - 29
30 - 34
35 - 39
40 - 44
45 - 49
50 - 54
55 - 59
60 - 64
Even high penetration The recently released World Internet Project report showed some surprising disparities
countries have a gender between the sexes in internet uptake and usage, extending even to highly penetrated
gap countries such as Korea and Singapore, as well as relatively highly penetrated countries
such as Spain and Italy. It appears to us that the more “mature” internet adopting
countries in the study’s sample (Sweden and the US) are more egalitarian in this regard,
though whether this is a function of maturity itself, or other factors, remains unclear.
The EURESCOM study found higher levels of usage in Italy and Germany generally,
but still exposed the same disparity in usage along gender lines (less in Italy, greater in
Germany).
Percentage of males and females who use the internet
100%
90%
80%
70%
60%
Male
50%
Female
40%
30%
20%
10%
0%
en
SA
in
e
ry
K
a
y
an
ly
an
ao
or
e
an
a
U
ed
ga
Ita
or
p
iw
ac
U
Sp
ap
m
Ja
Sw
K
un
Ta
ng
er
H
G
Si
Source: World Internet Project, 2004
Internet use by gender, 2002
90%
80%
70%
60%
50% Men
40% W omen
30%
20%
10%
0%
UK Italy Germany Norway
Source: EURESCOM 2004
Male internet usage more Not only is absolute penetration markedly different across gender lines, but so is
intense intensity and nature of use. The World Internet Project survey determined that men
spend 16 – 70% more time online than women, and are generally much more likely
than women to conduct a purchase online, including stock-trading. The Wave Two data
from EURESCOM (collected in 2002, partially released in 2004) shows that the heavy
internet user is more likely to be young, male and unmarried.
Average hours spent online per week (all locations), by gender
18
16
14
12
10 Male
8 Female
6
4
2
SA
en
ain
e
ry
ea
y
n
a
ly
an
ao
or
in
an
pa
ed
ga
Ita
or
w
ac
U
Sp
ap
Ch
rm
Ja
Sw
K
un
i
M
Ta
ng
Ge
Si
Percentage of male and female users who purchase on the internet
60%
50%
40%
Male
30%
Female
20%
10%
0%
SA
en
n
e
K
ry
ea
y
n
a
ly
an
ao
or
ai
in
an
pa
U
ed
ga
Ita
or
iw
ac
U
Sp
ap
Ch
Ja
Sw
K
un
Ta
ng
er
H
G
Si
Characteristics of light and heavy internet user

Light users Heavy users
Average age 40.6 34.0
Percentage with high school degree 44.2% 50.5%
Percentage with higher education 46.1% 39.5%
Percentage males 38.3% 60.0%
Percentage working 72.8% 60.6%
Percentage married 70.8% 46.6%
File-sharing, chat, IM – The same male bias largely holds true in terms of advanced applications, including
boys will be boys communications platforms. Judging from data from the Pew project, file sharing is
predominately a male pastime, as it has been since its inception. The same is true for
chat services.
File sharing dominated by men
Source: Pew Source: Pew Internet & American Life Project, America’s Online Pursuits, December 2003
Proportion of internet users using chat, by gender
35%
30%
25% M en Women
20%
15%
10%
5%
0%
Mar-00 Apr-00 M ay-June-00 Oct-Nov-01 Nov-Dec-01 Jun-Jul-02
Instant messaging usage shows the same pattern, though the gender gap is narrower in
younger age groups. Despite the many differences in audience size, preferred IM
platform from country to country (ICQ is very popular in Germany, for example),
levels of broadband penetration, and other factors, the gender gap in IM usage is
remarkably similar across all age ranges in the three countries covered by the charts
which follow.
UK, France and Germany IM users, proportion by age group and gender
40%
M ales
Females
20%
0%
Persons: 18+ Persons: 35+ Persons: 50+ Persons: 55+
Source: DIR, comScore MediaMetrix, June 2003
UK, France and Germany male IM user audience, proportion by age group
100%
90%
80%
70%
60% UK
50% France
40% Germany
30%
20%
10%
0%
UK, France and Germany female IM user audience, proportion by age group
100%
90%
80%
70%
60% UK
50% France
40% Germany
30%
20%
10%
0%
Workforce representation There may be very good and obvious reasons for such differences. One may be that the
may be behind the bias figures, many of which are based on usage in all locations (including work), may be
biased against women because of a generally lower representation in the workforce in
many of the countries covered. Related to this may be issues of life cycle stage (which
we discuss below), and therefore skewed by the relatively lower levels of free time
available to women engaged in caring for small children. However, the IM user data
seems to partly conflict with this idea, as the gender gap widens as age increases.
Perhaps another factor is that women are somehow fundamentally different in terms of
their communication and information consumption requirements and styles. The Pew
survey cited above noted that women showed a markedly higher level of interest in
obtaining health-related information online, relative to other forms of online activities.
Perhaps the internet is a more practical and utilitarian pursuit for women.
Sisters are doing it for Alternatively, perhaps advanced services are marketed in such a way as to make
themselves women less interested in them. Research consultancy Horowitz Associates recently
released a study, Women's Worlds III: Women as Media Consumers and Household
Decision Makers, in which it highlighted drastically different awareness levels and
attitudes towards advanced cable services on the part of women and men in the US.
The authors of the report cite a tendency for operators marketing such services to stress
the technology aspect of such services, which they believe may skew awareness and
uptake towards a male audience. (It is our own observation that, on our local cable
service, most advertisements for pay-per-view events seem to be promoting sports
events or action movies, i.e., content targeted at a largely male audience.) However,
with 10m out of 37m family households in the US headed by single mothers, their role
as decision makers in the consumption of information services is undoubtedly of
increasing importance. As we will discuss in a later section of this report, we think
much of Europe faces a similar scenario.
Awareness of advanced digital and interactive features
52%
HDTV
68%
36%
Video on demand
50%
Personal video recording 25%

services 41% Men Women
Subscription video on 14%

demand 20%
0% 10% 20% 30% 40% 50% 60% 70% 80%
Source: Horowitz Associates. N.B. Sample of people living in areas with cable access
(As a final aside, we wonder how much of the public profile of the industry is
influenced by issues of gender. Of the nine companies under our coverage in the PTT
space, only three had women visible in senior management roles at divisional level,
and of the lead analysts who cover these companies, fewer than 10% are women.)
Age and life cycle
You’re not getting older, Some of the information in the previous two sections has already touched on this, but
you’re getting more tech age seems to be a key determinant of uptake and usage of information services, in
literate virtually every market for which we have data. In virtually every category of service,
adoption and usage rates decrease across the age spectrum. The fixed and mobile usage
data in the section above showed this trend clearly, and mobile phone ownership
broadly follows it as well, as seen in the chart below. The notable exception is Norway,
where the country’s relatively earlier adoption of wireless technology has led to higher
penetration rates than in the other European countries shown.
EURESCOM e-living sample mobile phone ownership by age
100%
90%
80%
70%
UK
60% Italy
50% Germany
Norway
40%
30%
20%
10%
16 - 24 25 - 34 35 - 44 45 - 54 55 - 64 65 - 74 75+
EURESCOM e-living sample mobile calls and SMS per day, by age
Voice
3
SMS
0
16 - 24 25 - 34 35 - 44 45 - 54 55 - 64 65+
A second childhood for Internet usage shows the same pattern generally, though the age gap in usage varies
empty nesters? tremendously from country to country included in the World Internet Project report,
and also in the data from the EURESCOM survey (Wave One). It is interesting to see
the rebound in usage among Germans aged 45 – 54 (and somewhat later for the
Italians), and we can only ponder if this group may share some of the characteristics of
the Pew grouping known as “older wired baby boomers”. Perhaps most striking is the
wild divergence between the UK and Germany in the 65+ category. In the UK there
appears to be a strong suggestion of the existence of a “wired senior man” identified by
the Pew research, which we think may be related to the “cult of equity” in the UK. In
contrast, the drop in German usage is indeed striking. The IM data which follows is
consistent with a general trend of lower penetration as we move up the age range.
Percentage of users online more than 5 hours per week (all locations), by age
100%
90%
80%
70%
60% 16 - 24
50% 35 - 44
40% 55 - 64
30%
20%
10%
0%
Germany Hungary It aly Japan Korea Spain USA
EURESCOM e-living sample online minutes per day
80
70
60
50 UK
Italy
40
Germany
30 Norway
20
10
0
16 - 24 25 - 34 35 - 44 45 - 54 55 - 64 65+
UK, France and Germany IM user audience, proportion by age group
100%
90%
80%
70%
60% UK
50% France
40% Germany
30%
20%
10%
0%
You’ve got to grow up It is apparent from the preceding information that usage of all categories of electronic
sometime, or do you? communication and information services tends to decline as we move across the age
spectrum. Penetration and intensity of usage is generally greater within the younger
segments of the market. For newer services such as the mobile phone, the internet, and
particularly non-voice messaging systems (SMS, IM) this is particularly evident. One
key question for planners of telecom strategy departments is whether this is merely a
function of the adoption curve or if it relates to changes in needs and time usage
depending on the stage of the life cycle in which the user finds himself. In other words,
are young people more avid users of the new technology because they have grown up
with it, or is it also related to the fact that they may have more free time, and that the
social networks within which they operate incorporate these technologies freely? Are
habits established early in life sustained into adulthood, or do they give way to other
considerations, new interests, or time constraints?
Married to the Net The Pew data shown below gives some vague indication, if only on a small scale. It
shows the proportion two age groups who are IM users, over a period of nearly two-
and-a-half years. While penetration in the 18 – 29 bracket rose, the overall proportion
of IM users in the general internet user population appears to be in gradual decline.
What would be interesting to know is if the usage of IM by users in the younger
segment declines as they get older. We suspect it may, if the EURESCOM internet
usage stats on pages 24 and 31 are indicative of trends over the life cycle. In both cases,
there is a precipitous decline in usage with age up to the mid-40s, after which we see
an inflection point. We think this is suggestive of the experience seen in the Pew
grouping identified as “unwired young baby boomers” – shortage of time related to
career or family commitments. We think this may also be reflected in the EURESCOM
Wave Two data (page 26) which showed that the heavy internet user is likely to be of
average age 34, and that only 46.6% of this user group may be married. In contrast,
light users’ average age is 41, and 70.8% are married.
IM usage over time
80%
70%
60%
50%
18 - 29
40% All internet users
30%
20%
10%
0%
M ar-00 Apr-00 M ay-June-00 Oct-Nov-01 Nov-Dec-01 Jun-Jul-02
Internet maturity and changes in behavior
Phone usage and TV Related to the idea explored in the previous section is the issue of how usage patterns
viewing decreases as change according to internet maturity level (the period over which one has been
internet maturity “wired”) and how behavior may change with the adoption of broadband. It is clear that
increases internet usage has an impact on other forms of media consumption, and this pattern
seems to intensify over time. As the two charts below suggest, the greatest casualty of
the online usage is TV viewing. In the EURESCOM data, daily internet usage is more
or less flat across all levels of internet maturity, suggesting direct substitution of leisure
activities. Phone usage also seems to suffer significant erosion in direct relationship to
internet maturity, and for those who have used the internet for seven years or more,
phone usage appears to be c.25% below than that of newcomers. This is not terribly
surprising, and is consistent with some data previously produced in the US market,
though it is interesting to see the pattern repeated in Europe.
Hours per week watching television: users vs. non-users
30
25
20
Users
15
Non-users
10
0
en
SA
e
K
ry
a
y
an
)
ao
or
go
e
an
pa
U
ed
ga
or
iw
ac
U
ap
ia
Ja
Sw
K
un
Ta
ng
nt
er
H
a
Si
(S
ile
Ch
Technology use, minutes per day, by years of online experience
160
140
120
100 TV
PC
80
Internet
60 Phone
40
20
0
0.5 1 2 3 4 5 6 7+
% of respondents using phone & writing letters less, by internet experience
80%
70%
60%
50%
Write fewer letters
40%
Use phone less
30%
20%
10%
0%
0.5 1 2 3 4 5 6 7+
And web usage gets more Perhaps more interesting is the change in patterns of online usage itself, as the user
serious, but maybe not matures. Here we see that in the case of more experienced internet users, there is a
heavier marked increase over time in “practical” usage of the internet such as banking, travel
planning/booking and shopping, versus a more-or-less unchanged likelihood of using
the internet for educational pursuits or to download music. Recall that, for this survey
sample, daily internet usage is essentially flat irrespective of level of experience. The
suggestion is that, as the novelty of the internet for the new user wears off, time online,
rather than increasing steadily with experience, is increasingly rationed in favor of
more productive or practical pursuits.
Home online activities, last three months, by years of internet experience
80%
70%
60%
Travel
50% Banking
Shopping
40% Education
Music
30%
20%
10%
0.5 1 2 3 4 5 6 7
Broadband users do Lastly, it is clear that broadband adoption drives changes in consumption patterns.
different things Consumption of bandwidth intensive media shows a clear increase versus narrowband
users, and the range of content available seems to further drive a wedge between the
user and other forms of media. However, the Wave Two data just released from
EURESCOM’s e-living project poses questions as to how sustainable this shift is. In
comparing the minutes online per day of users in 2001 and 2002, there is a clear
increase in time online by those who migrated to broadband over the previous year
(except for the Norwegians, who seem indifferent). However, for those who had
broadband the previous year, there was a general decrease, and this is consistent with
the issue of internet fatigue which we explored in the PTT Pulse in 2003.
Narrowband and broadband users’ consumption of internet broadcasting
80%
70%
60%
50%
Narrowband
40%
Broadband
30%
20%
10%
0%
Ever Last month Last week
Source: Arbitron, 2003
Time online rivals TV and radio among broadband users (% leisure time)
Newspaper
6%
Internet
27%
Radio
31%
TV
36%
Source: Arbitron, 2003
Changes in daily time on internet 2001 to 2002 (min.)

UK Italy Germany Norway
Narrowband to narrowband 1.4 -2.9 -0.7 -2.2
Narrowband to broadband 14.6 7.0 11.5 0.3
Broadband to broadband -9.1 13 -0.9 -2.6
OFTEL Broadband Legacy UK regulator OFTEL (now subsumed within super-regulator OFCOM)
survey released a study of consumer internet usage in the UK in October 2003, which also
contained a number of interesting observations regarding broadband users’ behavior
online, and their management of the broadband environment, which are relevant to our
conclusions.
• Broadband users’ usage of applications is clearly different, enhanced and

shaped by the availability of additional bandwidth - Asked which applications
they use currently in comparison with those used prior to their migration to
broadband, former narrowband users cited an enhanced quality of usage for
applications such as file downloading, gaming, video clips/webcasts and
digital radio. As the chart below suggests, the availability of these services,
which broadband users found so frustrating in a narrowband environment,
seems to be driving higher usage.
• Broadband users use multiple simultaneous services to a great extent – A full
69% of broadband users claimed to have made some simultaneous use of the
phone while online, and 47% claim to do so frequently. For those former
narrowband users tied to just one phone line, this must be a revelation, and we
think this probably gives rise to a closer association between computer and
phone (stretching to computer as phone?)
• Broadband users are active at self-provisioning of services – Of the broadband
subscribers using email, personalized web pages and virus protection, only
75% reported using the ISP’s email service, 63% used a webpage provided by
the ISP and just over one quarter used the ISP’s virus protection service.
Comparison between previous and current use of broadband
Email 14%
General surfing 20%
Making purchases 13%
Downloading 51%
Gaming 23%
Instant messaging 11%
Video clips/webcasts 26%
Digital radio 19%
Chatrooms 3% Poor quality or unavailable with previous

connection
SMS messaging 4% Currently use
Real-time gambling 5%
Other 2%
0% 20% 40% 60% 80% 100%
Base: Broadband decision makers that previously used another connection, Aug. 03 Source:
Consumers' use of Internet, OFTEL residential survey, October 2003
% using the fixed line phone and the internet simultaneously
Don't know
Never 2%
16%
Frequently
Rarely
47%
13%
Sometimes
22%
Source: Consumers' use of Internet, OFTEL residential survey, October 2003
% using additional services and whether supplied by main ISP
65%
Email 48%
59%
Virus Use
protection 16%
Provided by main supplier
19%
Personal
webpage 12%
0% 20% 40% 60% 80%
Source: Consumers' use of Internet, OFTEL residential survey, October 2003
The opportunities to self- Recent data from Nielsen//NetRatings on the US market shows that, in the case of
provision applications is email applications, the unique audience for the access independent, web-based services
separating access from (Yahoo! Mail and Hotmail) is nearly twice that for AOL. We think this is yet another
services figure in support of the notion that, once users are in the “cloud,” the opportunities to
find and self-provision services result in some profound changes in the way they make
use of the internet. We suspect that the more tech-literate probably have multiple email
accounts (one author of this report has seven, if we count work and Bloomberg, and
finds the account supplied by the ISP used at home to be by far the most technically
unreliable and prone to mass spamming).
Top 10 Online Telecom/Internet Services Destinations week of Dec. 21, 2003
Brand or Channel Unique Audience (000) Active Reach (%)
AOL Instant Messenger 32,755 32.03
AOL Email 19,107 18.68
Yahoo! Mail 17,744 17.35
Microsoft Passport 16,833 16.46
MSN Hotmail 16,053 15.70
MSN Messenger Service 12,628 12.35
Yahoo! Messenger 9,233 9.03
EarthLink 4,711 4.61
Yahoo! Address Book 4,533 4.43
AOL Screenname 4,174 4.08
Source: Nielsen//NetRatings, January 2004
Fighting for mindshare This highlights, in our view, the increasingly tenuous relationship between access and
services, which we discuss later in this note. The implication of this for the telcos is
that, in order to avoid being relegated to the status of a “dumb pipe,” they will have to
fight harder for “mindshare” in the applications space, and that increasingly includes
voice. Such concerns for the telcos would seem to be borne out by information released
at the end of December 2003 by Nielsen//NetRatings. This showed that in the US
market, 76% of active Web users access the internet using a non-browser application,
the five most popular being instant messaging and media player platforms. This seems
to indicate a growing tendency to bypass the intermediary and get straight to the
content/services.
Nielsen//Netratings Top 5 internet applications ranked by active reach

Internet application Unique audience (000) Nov 03 Active user reach
Windows Media Player 48,159 34.43%
AOL Instant Messenger 28,363 20.27%
RealOne/RealPlayer 27,648 19.76%
MSN Messenger Service 27,018 19.31%
Yahoo! Messenger Service 17,148 12.26%
Source: Nielsen//NetRatings, December 2003
Viral services Recent history demonstrates that some self-provisioning of services and applications
can become epidemics. KaZaA, the de facto successor to Napster, has been
downloaded an estimated 300m times as at this writing, having broken the 200m mark
in early March 2003. It continues to be a hugely popular application, particularly in
Europe. The following table of top search terms from Terra Lycos, though containing
an insanely eclectic list of terms (including Paris Hilton – are there really that many
Americans interested in staying there?), is topped by KaZaA, and KaZaA ranks very
highly in the active reach tables from both Spain and Germany (KaZaA was also the
top search term on Yahoo! and Google). Across Europe, it is estimated that KaZaA
drew 9.35m home users, or 9.6% of all European home internet users, to its site in
October 2003, the most recent data point from Nielsen//Netratings. This is significantly
higher than the US audience, which accounted for 6.5% of home users. (This may
either reflect the impact of the RIAA lawsuits, a shift in preference to other sharing
platforms in America, a move to “darknets,” or all of the above.) More generally, the
four major music destinations on the web (KaZaA, RealOne/RealPlayer, WinAmp and
Windows Media Player) were visited by 43% of European internet users.
Top 10 search terms on Lycos, 2003
1 KaZaA
2 Britney Spears
3 Dragonball
4 Paris Hilton
5 The IRS
6 Kobe Bryant
7 Christmas
8 NFL
9 Pamela Anderson
10 Brooke Burke
Source: Terra Lycos, December 2003
Top 5 internet applications Spain and Germany (home panels)

Unique audience Active reach Time spent per person
Spain
MSN Messenger 4,245 51.94% 3:56:57
Windows Media Player 3,365 41.17% 1:13:14
WinAmp 1,599 19.56% 1:01:08
KaZaA 1,231 15.06% 1:16:20
RealOne/RealPlayer 928 11.35% 0:20:48
Germany
Windows Media Player 8,886 33.49% 0:44:52
T-Online Startcenter 8,692 32.76% 0:23:31
RealOne/RealPlayer 2,908 10.96% 0:16:54
WinAmp 2,358 8.89% 1:15:34
Apple Quicktime 2,271 8.56% 0:05:23
Source: DIR, from Nielsen//NetRatings data, December 2003
Combined reach of KaZaA, RealOne/RealPlayer, WinAmp and Windows

Media Player in select European countries
Country Active Reach
Spain 53.1%
Netherlands 48.1%
France 47.6%
Italy 44.2%
Switzerland 43.4%
Germany 41.6%
UK 37.7%
Sweden 37.2%
Source: Nielsen//NetRatings data, December 2003
Bandwidth utilization by P2P protocol

90%
80%
70%
60%
FastT rack
50%
eDonkey
WinMX
40%
Gnutella
30%
20%
10%
0%
Germany Israel Norway UK North America
Source: Sandvine Incorporated

Takeaways – section 3, part 2
• Disparity in adoption and usage by age group may be a function of the adoption
curve and generational differences, or may be indicative of changing behavior
through the life cycle of users.
• Women generally appear to be less enthusiastic users of information services
than men, though whether this is down to differences in taste or poor marketing
is unclear.
• Internet usage apparently tends to become more practical with greater
experience, though average usage does not appear to increase. Rather, TV
viewing and telephone usage suffer erosion which increases with maturity.
• More mature users and broadband users show a heightened tendency to self-
provision services. Many of the most popular applications are wholly
independent of the access provider, and generate no incremental revenue for
telcos.
General points before we move on

In the next section of this report, we are about to make an apparent hyperlink to another
topic, but there are direct connections between where we’re going and where we’ve
been. Key messages to retain as we make the transition are:
• So far most of the user data we have looked at strongly suggest that a younger,
predominantly male audience is the most active user base for a range of services
which broadband service providers – incumbent telcos foremost among them –
will rely on to drive future growth.
• There are serious questions in our minds as to whether the rampant use of such
services (some revenue generating and highly profitable, such as SMS) by the
youth segment will continue through life, or whether this is a temporary phase.
Internet usage to date suggests that usage becomes more practical and
productivity oriented as users gain experience/mature.
• Conversely, it is by no means clear whether the generally poor rates of uptake and
usage of advanced communications services by older user segments are reflective
of the adoption curve, or of changes in user needs and behavior across the life
cycle.
• Broadband users and more experienced internet users are apparently more likely
to change their user interface to cut out the ISP/portal/browser, change service
providers, use communications platforms not associated with their access provider,
choose an ISP other than the incumbent telco, and jettison the PSTN completely.
• The heaviest internet users, and the most aggressive in asserting their
independence (the point raised above), tend to be single.
4. Europe: The demographic timebomb
In the following section we examine the long-term demographic forecasts and recent
demographic trends seen in Europe, and consider the implications for the
sustainability of consumer demand for telecom services over the longer term, in light of
the user data we looked at in the previous section.
In our examination of consumer demographics and behavior in the previous section, we

posed a number of questions relating to whether major observable differences in
consumption of information services across the age groups were a function of the
adoption curve (i.e., do fewer older people use the advanced services because they are
less comfortable with them, or is there something inherent in the life cycle which
makes older people less likely to use them?). We also noted differences in levels of
awareness and usage of certain types of services based on gender, to which we posed
similar questions (i.e., are women generally less interested in advanced technology
features because they are marketed with a bias toward a male audience, or is there a
fundamental difference in women’s lives, psychologically or in practical terms, which
makes such services appeal less to them?).
Shuffleboard, here we One reason we think these questions may be important to the longer term outlook for
come the sector is that, according to a number of recent projections, Europe possesses the
weakest demographic trend outlook of any region globally. The chart below, replicated
from the United Nations’ The State of World Population 2003 report, shows that
Europe is expected to see the single largest percentage decline in population of 10 – 19
years of age in the period 2003 – 2050.
Total population 10 – 19 years of age by region, 2003 and 2050
800
700
600 North America
500 Latin America/Caribbean

Sub-Saharan Africa
400
Europe
300 Arab States region
200 Asia and the Pacific
100
0
2003 2050
Source: DIR, from United Nations The State of World Population 2003
Europe to see a 13% According to the UN’s projections, Europe as a whole should have a fertility rate of
decline in population to only 1.38 in the period 2000 – 2005, significantly below the 1.56 expected in the
2050 world’s “more developed” regions, and just more than half the 2.69 for the world as a
whole. Eastern and Southern Europe fare the worst, at 1.18 (suggesting depopulation of
15 – 30% for the region’s countries) and 1.32, respectively. Northern Europe, and the
Nordic countries in particular, have much better outlooks, and Western Europe is in
line with the developed world total. However, within this, there are pockets of
weakness, such as Austria, Switzerland and Germany. It is striking that of the big five
countries constituting the bulk of EU population (Germany, France, Italy, Spain and
the UK), only the UK and France are facing an increase in population over the period
2003 – 2050, while Europe as a whole sees a contraction of 13%. Meanwhile, the total
global population is poised to increase by 41.4%, and the more developed nations by
1.4%. We have never been great fans of overpopulation by any means, but in a sector
where the conventional earnings formula runs “channels x volume x price = revenue”,
the idea of a major contraction in population is naturally worrisome.
UN projections of European population and fertility rates
Fertility rate 2000 GNI per capita
2003 2050 Change -2005 PPP$ (2001)
Austria 8.1 7.4 -8.6% 1.28 27,080
Belgium 10.3 10.2 -1.0% 1.66 28,210
France 60.1 64.2 6.8% 1.89 25,280
Germany 82.5 79.1 -4.1% 1.35 25,530
Netherlands 16.1 17.0 5.6% 1.72 26,440
Switzerland 7.2 5.8 -19.4% 1.41 31,320
Western Europe 184.9 184.5 -0.2% 1.58
Bulgaria 7.9 5.3 -32.9% 1.10 5,950
Czech Republic 10.2 8.6 -15.7% 1.16 14,550
Hungary 9.9 7.6 -23.2% 1.20 12,570
Poland 38.6 33.0 -14.5% 1.26 9,280
Romania 22.3 18.1 -18.8% 1.32 6,980
Slovakia 5.4 4.9 -9.3% 1.28 11,610
Eastern Europe 300.3 221.7 -26.2% 1.18
Denmark 5.4 5.3 -1.9% 1.77 27,950
Estonia 1.3 0.7 -46.2% 1.22 10,020
Finland 5.2 4.9 -5.8% 1.73 25,180
Ireland 4.0 5.0 25.0% 1.90 27,460
Latvia 2.3 1.3 -43.5% 1.10 7,870
Lithuania 3.4 2.5 -26.5% 1.25 7,610
Norway 4.5 4.9 8.9% 1.80 30,440
Sweden 8.9 8.7 -2.2% 1.64 24,670
UK 59.3 66.2 11.6% 1.60 24,460
Northern Europe 94.8 100.1 5.6% 1.61
Albania 3.2 3.7 15.6% 2.28 3,880
Bosnia & Herzegovina 4.2 3.6 -14.3% 1.30
Croatia 4.4 3.6 -18.2% 1.65 8,440
Greece 11.0 9.8 -10.9% 1.27 17,860
Italy 57.4 44.9 -21.8% 1.23 24,340
Macedonia 2.1 2.2 4.8% 1.90 4,860
Portugal 10.1 9.0 -10.9% 1.45 17,720
Serbia and Montenegro 10.5 9.4 -10.5% 1.65
Slovenia 2.0 1.6 -20.0% 1.14 18,160
Spain 41.1 37.3 -9.2% 1.15 20,150
Southern Europe 146.4 125.6 -14.2% 1.32
Europe 726.3 631.9 -13.0% 1.38
More developed regions 1,203.3 1,219.7 1.4% 1.56
World total 6,301.5 8,918.7 41.5% 2.69
What does an older This is not just a game of numbers, however, but also one of customer quality. The
population mean for info other issue which concerns us is the age profile, as advancements in medical care and
service usage? lifestyle improvements mean longer average life expectancy. Obviously, our intent is
not to be ageist in this exercise, but if there is any validity to the view that older users
may inherently consume less of the kinds of services telcos expect to be offering in the
future, then we must again take this as a negative. We are no more eager to be
perceived as sexist in our views either, but given the fact that the projected average life
expectancy for girls born in the EU in 2001 is 8% longer than that of boys, it is also
clear that the market for information services in Europe will be a very different (and
perhaps less enthusiastic?) one in 2050, if indeed there is some inherent gender-based
difference in consumption patterns as suggested by the data we have examined so far.
Some countries much Clearly, not all European countries will suffer equally, and such long range forecasts
more exposed than others may be of limited value in the short term, if not entirely speculative. The UN forecast
for Sweden, for example, shows a 2% decline in population to 2050, whereas Sweden’s
own forecasts show a 19% increase in population over the same period. Conversely,
the authors of Population of Germany Today and Tomorrow (published before the
UN’s 2003 update) remark that they view the previous forecast by the UN to be
optimistic. Indeed, the UN’s updated projection of population for Germany seems to
fall between Germany’s central and optimistic cases, though nearer the optimistic end
of the range. Germany’s own pessimistic case forecasts an 18.8% decline in population
to 2050. The central case assumes a decline of 8.9%, with the proportion of persons
over 65 growing from 17% in 2001 to 29.6% in 2050. The proportion of people over
the age of 80 is forecast to grow from 5% of population in 2010 to 12.1% by 2050. On
its own forecasts, Sweden will grow older, but maintain a stable pool of persons under
24 over the next half-century.
Historical fertility rates and life expectancy in Europe

Trends in fertility rate Life expectancy of newborns 2001
1990 1995 2000 Boys Girls
EU 15 1.57 1.42 1.48 75.3 81.4
Belgium 1.62 1.55 1.66 74.6 80.8
Denmark 1.67 1.75 1.77 74.3 79.0
Germany 1.45 1.32 1.38 74.8 80.8
Finland 1.78 1.76 1.73 74.6 81.5
France 1.78 1.72 1.88 75.5 83.0
Greece 1.39 1.30 1.29 75.4 80.7
Ireland 2.11 1.89 1.89 73.0 78.5
Italy 1.33 1.20 1.24 76.7 82.9
Luxembourg 1.61 1.76 1.80 74.9 81.3
Netherlands 1.62 1.53 1.72 75.7 80.6
Austria 1.45 1.40 1.34 75.4 81.2
Portugal 1.57 1.40 1.52 73.5 80.3
Sweden 2.13 1.73 1.54 77.5 82.1
Spain 1.36 1.18 1.23 75.6 82.9
United Kingdom 1.83 1.71 1.64 75.7 80.4
Iceland 2.30 2.08 2.10 78.0 81.4
Japan 1.54 1.42 1.41 77.6 84.2
Norway 1.93 1.87 1.85 76.0 81.4
Switzerland 1.59 1.48 1.50 77.2 82.8
United States 2.08 2.02 2.06 74.4 80.0
Source: DIR, truncated from Population of Germany Today and Tomorrow, Federal Statistical Office of
Germany, July 2003
Three scenarios for German population decline 2002 - 2050
85,000
83,000
81,000 82,522 81,253
79,000
77,000
75,000 75,117
73,000
minimum
71,000
medium
69,000
maximum
67,000 67,046
65,000
2002 2010 2015 2020 2025 2030 2035 2040 2045 2050
Source: DIR, trom Population of Germany Today and Tomorrow, Federal Statistical Office of Germany,
July 2003
Germany population forecast 2001 – 2050 (central case)
84,000 45%
82,000 43.15% 40%

41.54%
80,000
36.89% 35%
36.10%
78,000 34.64% 34.16%
Total (central case) 30%
76,000 Under 20's as % (RHS)
20 - 50's as % (RHS) 25%
74,000 20.94%
18.69%
17.57% 17.15% 20%
72,000 16.39% 16.10%
70,000 15%
2001 2010 2020 2030 2040 2050
Source: DIR, from Population of Germany Today and Tomorrow, Federal Statistical Office of Germany,
July 2003
Sweden population forecast 2002 - 2050
11000 45%
10500 40%
40.56%
39.75% 38.95% 39.01%
10000 38.57% 35%
36.30% 36.82% 36.17%
29.76% 29.73% 29.75% 28.97% 28.99%
28.31% 28.76% 28.50%
9500 30%
9000 Total 25%

Under 24 as % (RHS)
8500 20%
25 - 54 as % (RHS)
8000 15%
2002 2005 2010 2015 2020 2030 2040 2050
Source: DIR, from SCB, December 2003
Which countries offer the Forecasting, as with history (and some would say accounting), is largely a matter of
least risk from opinion, we believe. Nevertheless, taking the various data sources together, we can
aging/depopulation? identify two groups of countries which are relatively better and worse bets from the
standpoint of demographic development. It is not intended as a guide for stock-picking,
but it will figure in our overall assessment of market dynamics in the multifactor rating
system which we introduce later in this note.
Ranking of European countries by demographic trend

Change in population 2003 - 2050 Total fertility rate 2000 -2005
Positive trend
Ireland 25.0% 1.90
UK 11.6% 1.60
Norway 8.9% 1.80
France 6.8% 1.89
Netherlands 5.6% 1.72
Stable trend
Belgium -1.0% 1.66
Denmark -1.9% 1.77
Sweden -2.2% 1.64
Negative trend
Germany -4.1% 1.35
Finland -5.8% 1.73
Austria -8.6% 1.28
Spain -9.2% 1.15
Portugal -10.9% 1.45
Greece -10.9% 1.27
Switzerland -19.4% 1.41
Italy -21.8% 1.23
Household creation may While Europe as a whole grows older and some countries grow smaller, it is likely in
outstrip population our view that the number of households may grow dramatically in a number of
growth countries, based on falling marriage rates and rising divorce rates. Since 1960, the EU
nations’ marriage rate has dropped by an estimated 54%, and the divorce rate has risen
by c.250%. The table below shows EU marriage and divorce rates as of 2001, while the
following table shows trends from 1980, 1990 and 2001. The latter shows a fairly
dramatic shift in the marriage rate, which has fallen by 20 – 35% for many of the key
countries in Europe over the period, and comparable levels of increase in the divorce
rate. Though still low in relative terms, divorce rates in countries such as Italy and
Portugal have shown the most startling increase, while marriage rates in the former
Eastern Bloc nations have fallen most dramatically.
EU marriage and divorce rates, 2001 (per 1,000 population)

Marriages1 Divorces2
Denmark 6.6 2.7
Portugal 5.7 1.8
Greece 5.4 0.9
Spain 5.2 1.0
United Kingdom 5.1 2.6
France 5.1 2.0
Irish Republic 5.1 0.7
Netherlands 5.1 2.3
Italy 4.9 0.7
Finland 4.8 2.6
Germany 4.7 2.4
Luxembourg 4.5 2.3
Austria 4.2 2.5
Belgium 4.2 2.9
Sweden 4.0 2.4
EU average 5.1 1.9
1 2000 data for UK, Italy and EU average. 2 2000 data for UK, Austria, Germany, Spain, Italy, Irish
Republic and EU average. 1999 data for France.
Source: Eurostat
Marriage and divorce trends 1980 - 2001
Marriages per 1 000 Divorces per 1 000
Notes population population Divorces (% of marriages)
Europe 1980 1990 2001 1980 1990 2001 1980 1990 2001
Austria 6.2 5.8 4.2 1.8 2.1 2.5 29 36 60
Belgium a 6.7 6.5 4.4 1.5 2.0 2.6 22 31 59
Denmark 5.2 6.1 6.8 2.7 2.7 2.7 52 44 40
Finland 6.1 5.0 4.8 2.0 2.6 2.6 33 53 54
France b,c 6.2 5.1 5.1 1.5 1.9 1.9 24 37 37
Germany c 6.3 6.5 4.7 1.8 1.9 2.4 29 29 47
Greece a,d 6.5 5.8 4.3 0.7 0.6 0.9 11 10 21
Ireland 6.4 5.1 5.0 … … 0.7 … … 14
Italy c 5.7 5.6 4.5 0.2 0.5 0.7 4 9 14
Luxembourg 5.9 6.1 4.5 1.6 2.0 2.3 27 33 51
Netherlands 6.4 6.4 5.0 1.8 1.9 2.3 28 30 46
Portugal 7.4 7.2 5.7 0.6 0.9 1.8 8 13 32
Spain c,d 5.9 5.7 5.1 … 0.6 1.0 … 11 19
Sweden 4.5 4.7 4.0 2.4 2.3 2.4 53 48 60
United Kingdom a 7.4 6.5 5.1 2.8 2.9 2.6 38 44 51
Norway a,f 5.4 5.2 5.6 1.6 2.4 2.2 30 46 39
Switzerland f 5.7 6.9 5.0 1.7 2.0 2.2 30 29 44
Czech Republic f 7.6 8.8 5.1 2.6 3.1 3.1 34 35 61
Estonia 8.8 7.5 4.1 4.1 3.7 3.2 47 49 78
Hungary 7.5 6.4 4.3 2.6 2.4 2.4 35 37 56
Latvia f 9.8 8.9 3.9 5.0 4.0 2.4 51 45 62
Lithuania 9.2 9.8 4.5 3.2 3.4 3.2 35 35 71
Poland 8.6 6.7 5.0 1.1 1.1 1.2 13 17 24
Slovakia f 7.9 7.6 4.4 1.3 1.7 1.8 16 22 41
Slovenia 6.5 4.3 3.5 1.2 0.9 1.1 18 22 31
North America:
Canada e 7.9 6.8 5.0 2.6 2.9 2.3 33 43 46
United States a 10.5 9.8 8.4 5.1 4.7 … 49 48 …
Source: DIR, UNECE
a/ Marriages and divorces: data for 2000 instead of 2001.
b/ Mean age at first marriage: data for 1998.
c/ Divorces: data for 2000 instead of 2001.
d/ Mean age at first marriage: data for 1999.
e/ Marriages and divorces: data for 1999 instead of 2001.
f Mean age at first marriage: data for 2001.
g/ Divorces per 1000 population: data for 2000 instead of 2001; divorces (in % of marriages): data
for 1999 instead of 2001.
More single parent It is also interesting to note that, in many cases, countries with lower than average
households headed by marriage rates and higher than average divorce rates also have higher than average
women fertility rates. This suggests to us that the positive population growth trends expected in
some of these counties will result in a larger proportion of single parent households,
many of which will be headed by women.
High divorce, low marriage, high fertility countries

Divorce rate 2001 Marriage rate 2001 Fertility rate 2001 – 05
Finland 2.60 4.80 1.73
Sweden 2.40 4.00 1.64
United Kingdom 2.60 5.10 1.60
Netherlands 2.30 5.00 1.72
Denmark 2.70 6.80 1.77
Belgium 2.60 4.40 1.66
Europe average 1.99 4.93 1.55
Source: DIR, from UN data
Single person households The repercussions of this trend so far are that many European countries have attained
already at startling levels significantly higher levels of single person households than that found in the US, which
has a significantly higher divorce rate (and apparently more people getting remarried).
This is most evident in the markets of Northern Europe where average household size
has fallen to the 2.1 – 2.3 band in a number of countries, and single person households
are approaching 40% of the total. We do not have detailed longitudinal data sets at our
disposal to assess how fast things are changing, but looking at data from the
Netherlands, which has neither excessively low marriage nor high divorce rates for
Europe, we can see that between 1995 and 2003 household growth (8.2%) outstripped
population growth (5%) by a significant margin, and the average household size fell
from 2.35 to 2.28. Extending this CAGR (-0.4%) out to 2050, average household size
falls below 2.0 persons in the mid-2030’s.
European household data, single person households, average size

Year Total Households (‘000) 1 person households (% of total) Average household size
Europe:
Austria 2001 3 310.6 31.6 2.4
Belgium 1998 4 178.7 30.6 2.4
Denmark 2002 2 456.1 37.2 2.2
Finland 2002 2 295.4 37.3 2.2
France 1999 23 810.0 31.0 2.4
Germany 1998 37 532.0 35.4 2.2
Greece 1991 3 203.8 16.2 3.0
Ireland 2001 1 290.6 21.9 3.0
Italy 1991 19 909.0 20.6 …
Luxembourg 2000 163.1 … …
Netherlands 1998 6 758.0 32.5 2.3
Portugal 1991 3 145.7 13.8 3.0
Spain 1998 12 626.0 14.1 3.2
Sweden 1990 3 830.0 39.6 2.1
United Kingdom 2001 24 818.0 28.7 2.4
Norway 1998 2 049.0 40.5 2.2
Switzerland 1990 2 841.9 32.4 2.3
Czech Republic 1998 3 731.1 19.2 2.7
Estonia 2001 569 30.8 2.4
Hungary 2001 3 837.0 25.6 2.6
Latvia 2000 802.8 25.0 2.7
Lithuania 2001 1 330.0 24.1 2.6
Poland 1995 12 501.0 19.7 3.1
Slovakia 1991 1 778.0 20.9 2.9
Slovenia 1991 640 18.0 3.1
North America:
Canada 1996 10 820.0 24.2 2.6
United States 2000 105 480.1 25.8 2.6
Source: DIR, UNECE
Single life isn’t all Quite apart from the glamorous depictions of single life promulgated in television and
cocktails and innuendo popular culture, EURESCOM data suggests that, in fact, single persons are generally
less sociable across the entire age spectrum than their married counterparts (Metcalf’s
law strikes again?). The data seems to suggest a greater tendency to use the telephone
(or other communications media, we assume) among singles over 60 years of age,
though single parents are only marginally more likely to use the phone rather than
make a personal visit, and in either case, the sociability index of single parents appears
to be relatively low.
Sociability orientation by stage in life cycle
Source: EURESCOM, 2001
Conclusion
The foregoing analysis is admittedly not particularly scientific, but we are not
demographers by trade. Our point is that household formation in many European
countries may exceed population growth significantly over the long term thanks to
lower birth rates, lower marriage rates, higher divorce rates and longer life expectancy.
Thus, in theory, there may be more household connections for telcos to serve in some
countries. However, we think this scenario is complicated by a number of conclusions
and questions, which we must consider at this point.
• The nations of Europe are to see their populations getting significantly older,
and in some cases, smaller, over the next 50 years. With the longer lifespans
attributed to women (8% on average for girls born in 2001, compared to boys
born the same year), Europe is also likely to become more female.
• Rising levels of divorce and flagging marriage rates, coupled with higher than
average fertility rates in some countries, are creating more households of smaller
average size, and the proportion of single person households is approaching
40% in some countries. Research suggests that, across the life cycle, single
people are less sociable than their married counterparts.
• Thinking back to the Pew data, which showed a high-spending, young, single
tech elite as being the most likely to prefer a non-telco ISP and to drop a fixed
line connection in favor of mobile-only, are single person households actually
more likely to drop domestic telecom services due to their circumstances? (Why
maintain a PSTN phone if you have a broadband cable connection, there’s no
one home to call, and you carry a mobile with you at all times?)
• Do the research findings on the relative lack of interest in some advanced
services on the part of women have long-term negative implications in a Europe
increasingly populated by women, who are more likely to be heads of households
and decision-makers in the consumption of information services?
• What proportion of the enlarged group of elderly Europeans will be housed in
institutional settings in future, and what implications will this have for total
number of channels and usage of consumer information services?
General points before we move on
So far we have covered a wide range of issues, but where this is all leading ultimately
is towards a consideration of some factors which we think may drive access line loss
for the telcos, and/or increasingly drive a wedge between access provision from the
pallet of services available to consumers. The revenue loss to telcos may be significant.
In some cases, such as the VoIP services market, incumbents may be well placed to
play, and margins may be higher in the long run, though market share and revenues
may be smaller, and the transition may be painful and costly in the medium term. In
other cases, such as the rise of closed communications platforms or the development of
metropolitan fiber initiatives, telcos may expect to only generate revenue from access,
or perhaps not at all. Key concepts to retain:
• Demand for voice communications may be intrinsically limited regardless of the

number of channels or platforms available to the user. Substitution of revenue-
generating voice with alternative, non-revenue-generating usages appears to
increase as internet experience increases.
• Younger and more experienced users of info services are eclectic and
promiscuous. Loyalty to the access provider may be tenuous, and the tendency
toward self-provision of services intense.
• Whether higher consumption trends among the young and early adopter segments
continue though life or not, projections suggest that the number of channels open
to telcos may contract with demographic change in certain countries. Conversely,
the number of channels in some countries may have the potential to increase as
average household size decreases, but we have concerns as to demand levels in
smaller or single person households. These may be compounded by gender issues
or differences in sociability levels in such households.
• Conversely, it is by no means clear whether the generally poor rates of uptake and
usage of advanced communications services by older user segments are reflective
of the adoption curve, or of changes in user needs and behavior across the life
cycle.
• Nevertheless, the relatively weak consumption pattern of older user segments is
an obvious concern in light of long-term demographic trends.
5. Challenges for the telcos - the access layer
We believe there is a growing tension between the provision of access and the growing
range of services available to consumers. While telcos arguably have a very strong
position in the former, and are principally under attack in the latter, we can identify
some significant sources of pressure which may further fragment the access market.
Some of these are explored below.
Mobile substitution
Based on national regulator data, independent surveys and comments from operators,
we think the proportion of mobile-only households in Europe is currently around 8 –
9% on average, though surveys from the Italian statistics office in 2002 showed a
figure c.13%. In our note from February 2003, we went in to some depth as to the
impacts of mobile substitution in access and usage (see European wireline 2003 and
beyond: it’s a marathon not a sprint, 7th February, 2003). In addition to finding that the
revenue and EBITDA trade-off for the integrated operator is negative (despite a higher
revenue yield on mobile calls versus fixed), we also determined that reported mobile
margins are highly geared to termination revenues, and thus at risk from further
regulatory intervention.
A bad revenue and In summary, we determined that the risk of fixed line loss for the integrated operators
margin trade was a significant one. While the yield per minute for mobile services is 10x greater
than for fixed, the European integrated operator will typically have a mobile market
share in the mid-to-high 40% range, rather than the average 94% share they enjoy in
fixed access. Thus the line loss on the fixed side may only result in a transfer of
minutes to the mobile network roughly half the time. On top of that, excluding
termination revenues, the underlying margin on services looks to be significantly lower
than that for fixed services, according to the analysis we did last February, and this
view is supported by recent events.
In December 2003 we saw the results of the Dutch regulatory review into mobile
termination rates, which resulted in mandated cut of 41.8% over three years for the two
GSM900 operators, phased as seen in the table below.
Termination rate cut schedule for KPN Mobile

Termination rate (€ cents) Percentage change from previous
Jan 1 '04 15.5 -18.0%
Dec 1 '04 13.0 -16.1%
Dec 1 '05 11.0 -15.4%
Total decline 7.9 -41.8%
Source: DIR, from company data
What can the Dutch KPN’s estimated impact on the results of KPN Mobile NL in 2004 were revenue -
teach us? €110m and EBITDA -€65m. As we did with T-Mobile UK and O2 UK last year, we
have attempted to use KPN’s guidance to work down to the underlying margin of the
mobile business as implied by the impact of termination cuts. There are several steps to
this process, some of which involve some assumptions on our part.
• Based on our estimates of €2.5bn in revenue for NL in 2004, we estimate that

for a termination cut of 18% to have an impact of -€110m at the revenue level,
termination must account for c.25% of total revenues. (2.5 x 0.25 x 0.18 =
112.5m )
• Based on our estimate of EBITDA of €1bn for 2004, KPN Mobile NL claims
an EBITDA margin of 40.2%. Taking the difference between revenue and
EBITDA we derive “cash costs,” of which we attempt to break out
interconnection and specifically mobile-to-mobile interconnection costs, based
on our estimates. This is to allow us to adjust the net EBITDA decline to
compensate for the reduction in payments to other operators which KPN
Mobile will enjoy.
• This, in turn, gives us an estimate of the EBITDA impact on the lost
termination revenue, before reductions in termination payments are made.
Based on this analysis, the gross EBITDA loss may be €92m on revenue loss
of €112m, implying a margin of 81.7% on termination.
• In other words, with a company margin of 40.2%, if 25% of the revenues
generate an 81.7% margin, then the underlying margin on other services is
probably c.26.4%.
Our dissection of impact of termination cuts on KPN Mobile NL

2004 Comments
Revenue 2,499 DIR estimate

Percentage termination 25% DIR estimate
Cash costs 1,493 Difference between DIR revenue and EBITDA estimates
of which interconnection 373 DIR estimate
of which to other mobile 149 DIR estimate
EBITDA 1,006 Company guidance prior to termination cut

margin 40.2%
Termination revenue 625 DIR estimate

Reduction in pricing 18%
Revised termination revenue 512 Termination revenue post mandated cuts, our estimate
Revised revenue 2,387

Revenue drop -4.5%
Revenue lost 112 This is in line with company estimates of revenue impact
Reduction in payments to other operators -27
Revised EBITDA 941 Our original estimate minus company estimate of -€65m impact
new margin 39.4%
Net EBITDA loss 65 Company comment on impact of termination cuts was EUR65 full-year 2004
Headline EBITDA loss 92 Net EBITDA loss adding back savings on payments to other operators
Implied EBITDA margin on termination 81.7% Headline EBITDA loss/Revenue loss
Implied termination EBITDA 510 Applying implied margin to termination revenues
Implied underlying EBITDA margin 26.4%
Source: DIR estimates
Technology
We believe there are a plethora of technology issues which confront telcos in the
access environment. The ongoing debate over the use of new wireless technologies
such as Wi-Max and Ultrawideband opens up a whole range of new issues for the
telcos, but we believe these are of longer-term consequence. Elsewhere, a number of
powerline deployments are ongoing in the UK, Spain, Finland and elsewhere, and we
believe these may deserve some separate attention at a later date. However, a couple of
developments are more immediate, and capable, in our view, of delivering some
negative surprises for the incumbents: the apparent development of a low-cost
Ethernet-over-cable solution, and the ongoing rapid development of Wi-Fi as an access
to free/cheap voice.
Ethernet over cable – This innovation is being led by Essent Kabelcom, the second-
largest Dutch cable operator, based mainly in the eastern portion of the Netherlands,
which serves 1.8m households, c.300,000 of which are broadband subs. Essent was
looking for a way to eliminate the costly and complex problem of node-splitting
necessary with EuroDocsis cable modem deployments (the practice of progressively
subdividing the number of homes served within one node, in order to offset the effects
of bandwidth contention brought about by increasing penetration of broadband
services).
Fiber speeds on a coax The solution developed by Teleste, in response to an idea from Essent, was a very
pipe? small Ethernet switch which resides in the end amplifier (familiar to pedestrians in
many countries as the mysterious green metal cabinet found at intervals along
sidewalks and on street corners - in the case of Essent, this box is normally only 50cm
high x 30cm wide). In the case of Essent's network topology, these end amplifiers are
normally found within 100m of the households they serve. Connected through
multitaps into the individual coaxial feeds, the ethernet switches are capable of
delivering 10Mbps symmetrical Ethernet to the home. This compares to the 1 to 2
Mbps on offer in Essent's existing premium cable modem packages, and is comparable
to the speeds delivered in metro fiber Ethernet services like B2 Bredbandsbolaget and
Fastweb, but notably without the installation of fiber. The other advantage is that the
customer premises only require a standard passive RJ45 connection in the wall.
Essent is planning a trial of 3,000 homes in the Southern Netherlands in Q2 2004, for
around three months. We therefore expect a commercial launch in H2 2004, perhaps
more towards the end of the year. Network-wide roll-out should not be expected to be
as fast as EuroDocsis, which was completed in three months, but should be completed
within a year of start. Essent also claims to be seeing intense interest in the technology
from other cable operators, and it is our understanding that Teleste is working on
variations of the technology to suit different network topologies.
Could change the balance In addition to the trials and deployment of Ethernet, Essent has completed its VoIP trial
of power for some cable and is proceeding to commercial launch in 2004 (along with Casema and UPC we
players believe), initially targeting its installed base of cable modem subs, which we think may
number roughly 335,000. For the same unit cost as a EuroDocsis deployment, Essent
(or others) may be able to deliver a broadband service 10 to 20x faster than typical
consumer DSL services in Europe, with greater bandwidth efficiency and scalability.
This could allow cablecos to deliver a fiber-grade service without the investment in
fiber, potentially at a price very competitive with far slower DSL offerings.
Additionally, the frequency of the networks is immaterial in this environment, so it
would work on older, lower capacity systems, as long as they have a back channel
(return path - in other words two-way data flow). It is Essent’s chief engineer’s view
that the solution would be particularly suited to high-rise MDUs, suggesting that
markets such as Germany, long "immune" to competition from cable, may be eager
adaptors of this innovation. In short, networks which have so far struggled with the
costs of capacity upgrades and EuroDocsis deployments may in time be able to close
the gap in the broadband race, or indeed, to lure DSL subs away with a far superior
triple play product.
Wi-Fi as a backdoor – As we began tracking over a year back, 2003 ended up being a
huge year for the Wi-Fi phenomenon in Europe, with the inevitable land-grab from
PTTs in a rush not to miss out. Besides the commercial deployments seen across
Europe, the consumer has also played a significant role. Data from In-Stat/MDR shows
that global shipments of network interface cards (NICs) and access points (APs) grew
from 7.2m units in 2002, to 22.7m in 2003. Europe led the growth, moving up from
9% of total shipments in 2002 to 15% in 2003, implying 3.4m units shipped to
European users last year versus 650,000 in the previous year.
Wi-FI network interface card and AP shipment trends 2002 - 2007
50
40
30
20
10
0
2002 2003 2004 2005 2006 2007
Source: In-Stat/MDR
The Wi-Fi vision is Early last year we went out on a limb and wrote about the potential of Wi-Fi to erode
coming true, sort of incumbent revenues via a combination of broadband line sharing and VoIP deployment
in the residential market, as well as the impact that VoIP in a public Wi-Fi hotspot
(commercial or otherwise) could have on the mobile industry, particularly in terms of
roaming. We have spent a lot of time observing and talking to people within the
community networking area, and to be honest, we have not yet seen anything in Europe
to rival the success of US initiatives such as Personal Telco in Portland (which still
claims more active hotspots than all other commercial operators combined) or
NYCWireless (which has driven the deployment of free hotspots in a number of city
parks in New York City). Nor have we seen the arrival of any large-scale WISPs,
though there are a number of smaller ones in operation.
In the zone What we have seen, on the other hand, is the VoIP over Wi-Fi concept move from
speculation on our part to more-or-less a mainstream concept. SIP softphones are now
available for use on a number of commercial and free SIP services in the US, and we
know people who use Skype and Free World Dialup over Wi-Fi (in airports and
elsewhere) and are perfectly happy with it. We have also seen some early examples of
Wi-Fi handsets, which admittedly require more technical finesse than the mass market
is capable of mustering – for now. As suitable devices and public hotspots proliferate,
we think leakage from the PSTN and mobile networks into the cloud of free/cheap
VoIP over Wi-Fi is a potentially huge phenomenon. Wi-Fi/multimode handset
initiatives from BT Group and other more forward-looking incumbents and mobile
players will seek to address this issue, and BT has a relatively free hand, in having no
cellular business to cannibalize. For the others, the risk of such a strategy is significant
collateral damage, and no doubt presents the organization with raw material for some
very sensitive discussions.
Xten X-lite SIP softphone
Source: Xten
How about video IP A huge amount has been written recently about the transformation of the PC within the
telephony on the TV home into a multimedia server, and this whole subject area is probably deserving of a
screen during the separate note. However, we think that Wi-Fi also provides the potential to transform
commercial ? other devices within the home, with some potentially significant implications for telcos.
We were fascinated by the release in mid-December 2003, of the D-Link i2eye Wi-Fi
(802.11b/g compliant) videophone unit (DVC-1100), which sits atop the television,
converting it into a display terminal for IP-based video and audio communication. The
unit is a standalone device connecting to the Wi-Fi access point in the home, and
employs H.323 (making it suitable for businesses, we expect), though we also believe
that we will see similar products using the SIP protocol. The D-Link solution contains
a directory service which supplies registered users with a “phone number” for direct
dialing from other users of the D-Link videophone. In so doing, D-Link is following
the approach of the SIP-based “closed” systems which we highlight later in this report,
crossing over from equipment manufacturer to service provider/directory manager. We
are also interested in the fact that, as a standalone device, the DVC-1100 cuts the PC
out of the transaction, while allowing the use of “PC-to-PC-style” communication (for
example, the video/audio IM applications discussed later in this note) in a more natural
and comfortable environment such as the living room.
D-Link i2eye DVC-1100 Wi-Fi videophone
Source: D-Link
Steady progress in There are also some examples in Europe of fairly substantial wide area links between
meshing technology nodes (see map of Berlin community overleaf), and a few good examples of
community networks using Wi-Fi in the last mile (see next section), but this becomes a
more compelling possibility if Wi-Fi’s cousin Wi-MAX (802.16) actually sees
adoption into the mainstream as a backhaul technology for connecting/aggregating
smaller Wi-Fi communities. Additionally, we think the Wi-Fi industry, both at a
multinational corporate level, and at a much lower level, is pushing hard to iron out the
issues surrounding meshing technologies. We have documented a number of such
developments over the past year, most recently the open source Mesh Cube developed
by German start-up 4G Systems. We saw this device at a wireless networking event in
London in November 2003, and were very impressed with its size, 5.5cm wide x 5.5cm
tall x 5.5cm deep (excluding antennae). At the time pricing was unclear, but we were
told to expect something around €200 – 300 per unit at the time. The company is
actively engaged in partnering for testing, development and distribution, and has
already agreed with a German company to port accounting, authorization and
authentication software into the cube networks. Such a plug-and-play meshing solution,
we believe, could significantly strengthen the hand of those attempting to create local
ad hoc open networks across Europe (see the section on community Wi-Fi and its
possible implications from European Wireline 2003 and beyond: it’s a marathon, not a
sprint). This theme is ripe for revisiting, and we intend to do so later this year in a
separate note.
Map of Berlin Wi-Fi backbone (nodes in blue, links in red)
Source: Freifunk.net
Mesh Cube from 4G Systems
Source: 4G Systems
Regulatory pressure on access obligations and pricing
A sense of impatience In the past year we have seen growing pressure from the alternative operator lobby on
the issue of wholesale PSTN access products, as well as some selective aggression on
unbundling pricing and procedures at national regulator level. We believe there is a
growing risk that, sensing stagnation in the market liberalization process, the EU may
exert pressure on national regulators to open up PSTN access on a wholesale basis to
alternative providers, and take a more aggressive stand on unbundling.
Only 19% of alternative The EU ninth report expresses satisfaction with the results of the Commission’s
ADSL lines are intervention in the French ADSL market in 2003, which resulted in a 30% drop in
unbundled unbundling charges from France Telecom. The report also stresses that its
determination of a margin squeeze in Deutsche Telekom’s pricing strategy was a clear
example of how the Commission will make use of anti-trust laws where appropriate in
furthering competition. It notes that, while the rate of unbundling has improved, the
process has been concentrated in six countries (Germany, Italy, Denmark, Finland, the
Netherlands and Sweden), and that of all retail ADSL lines offered by non-incumbent
operators, only 19% are fully unbundled. We sense from the tone of the report that the
Commission may believe that mandated price declines and market intervention has
driven the broadening of the ADSL market, and that there may be more efforts made in
this area in future. We believe there may be a similar stance taken with regard to
wholesale line rental.
Evolution of unbundled lines and pricing
WLR’s leveling effect Alternative operator Tele2, in a late 2003 position paper on European regulation, cites
the c.20% average increase in line rental charges as a stop-gap measure employed by
the incumbents to halt revenue erosion. Tele2’s proposal, and indeed that of a number
of proponents of competition in the European markets, is to require cost-based
wholesale access to PSTN subscription services. Tele2 cites the case of Denmark,
where since 2000 the incumbent is obliged to offer wholesale PSTN subscription to
competitors at a maximum price of retail minus 21%. It would appear that this strategy
has so far proven relatively effective at eroding the incumbent’s grip on access. Danish
regulator data from mid-2003 showed TDC to have an 84% share of access. In contrast
the EU estimates that just under 6% of EU-15 consumers are using an alternative
service provider via direct access. Indeed, this is comparable to BT’s share of access in
the UK residential market, where it faces facilities-based competition from two
relatively powerful cable operators.
Average market shares of EU-15 incumbents
100%
90%
80%
70%
60% 2000
50% 2001
40% 2002
30%
20%
10%
0%
Local calls Calls to mobile Long distance calls International calls
Changes in average prices for consumer fixed access EU-15 (1998 = 100)
30%
20% 20%
Consumer line rental
10%
0% Local call charge, 10

-2%
minut es
-10% -12% Average consumer
-20% expenditure
Int ernational basket
-30%
-40% -38% Nat ional call charge, 10

minut es
-50% -49%
-60%
1998 1999 2000 2001 2002
Source: EU eighth report
Different strokes, but Whereas Denmark implemented draconian pricing for the WLR regime, the UK is to
something bigger may be implement WLR at a mark-up to retail (£28 per quarter), though the background to the
coming product relates to a regulatory trade, rather than a mandated pricing structure. However,
in OFTEL’s conclusion statement from early 2003, it states that, as BT agreed to the
price in good faith in 2002, it would not be appropriate to attempt renegotiate,
suggesting a level of dissatisfaction with pricing. It is interesting to note that OFTEL
also set as its measure of progress a monthly level of at least 25,000 new WLR
connections and a further 25,000 monthly migrations from CPS to WLR (i.e., a
minimum of 600,000 lines transferred per annum). TeliaSonera in Sweden has
proactively approached the PTS with a plan to offer wholesale PSTN subscription at
cost, and anticipates starting in H2 2004. In contrast Italy last year negated the
feasibility of introducing such a product in the Italian market.
This entire issue has complex implications worthy of more in-depth analysis as the
situation develops in 2004, and we will return to it. However, we think the momentum
is building for a more harmonized approach to WLR adoption in the EU, and we think
the tone of the ninth report and also of the Commission’s recent intervention against
the Finnish NRA Ficora suggest it may adopt a more activist approach.
Big broadband: public policy and people power
“All new media, taught Marshall McLuhan, are destined to subsume and extend all old
media, and to use the old media as their content, much like large fish filling their
stomachs with small fish… The big fish of today is Big Broadband – access to the Web
at 10 to 100 megabits per second for homes and 1 to 10 gigabits per second for
businesses. The small fish are broadcast, DSL, cable modem, and voice.”
--- Reed Hundt, December 2003
“It is important to note here that the current generation of broadband technologies
(cable and DSL) may prove woefully insufficient to carry many of the advanced
applications driving future demand. Today’s broadband will be tomorrow’s traffic jam,
and the need for speed will persist as new applications and services gobble up existing
bandwidth.”
U.S. Department of Commerce, September 2002
“Telecom companies have said it is too expensive to run fiber to every home and
business. What they really mean is that it would not provide satisfactory returns to
their shareholders. We understand that. No single company could justify such an
investment. But cities are well suited to install this infrastructure, and because the
capacity is so great it can be shared among many firms and the investment repaid over
20 years, a far longer repayment period than a private company would accept…Let the
cities build the community-owned digital pipeline. Let the private sector offer
competing services over it. Let the users pay for it.”
--- Mayors of Brigham City and West Valley City, Utah, December 2003
Like sucking peanut We believe there is a growing proportion of public opinion in many countries which
butter through a straw views current broadband deployment as inadequate in the long term, both in scope and
depth, to meet the evolving demands of the market. As the P2P file-sharing
phenomenon (and associated bandwidth crunch for many ISPs) clearly demonstrates,
the provision of dramatically higher levels of bandwidth to mass market consumers
over a relatively short period of time can result in some decidedly negative, and largely
unforeseen, outcomes for both service providers and subscribers.
Bandwidth hogs run As an example, we would cite the chart below (best viewed in color), showing the
amok downstream traffic for an unnamed European ISP over a five day period, segmented by
protocol. The striking thing to notice is that the thick red band through the lower third
of the chart, the FastTrack protocol used in P2P file-sharing applications such as
KaZaA, either matches or swamps all HTTP protocol (in other words standard web
usage) throughout a typical 24-hour cycle. This is a prime example of what happens
when technologically savvy users gain access to unmediated, self-provisioned
applications in a broadband environment, but it is only one of a number of such
protocols to be found on this chart, and collectively these add up at certain times to
account for as much as 70% of all traffic on broadband networks. Beyond the
broadband access charges paid by customers to be able to use such applications, their
use generates no incremental revenue for the telcos. To the contrary, they have often
been the source of capacity constraints and traffic management problems for the
service providers, necessitating incremental investment to offset. In short, silicon may
have gotten ahead of copper.
Downstream traffic from a European ISP, by protocol
Source: Sandvine, PointTopic
Despite the recent high-profile RIAA lawsuits, file sharing continues to be a huge
phenomenon worldwide, and particularly strong in Europe. With a number of European
telcos making very bullish noises about the potential for DSL as a viable television
platform (the forecast below shows DSL accounting for 7.4% of all digital TV
households by 2010), there are voices within the industry who increasingly fear that
some networks may be reaching a critical stress point, as ADSL household penetration
moves towards 20%, and ADSL access speeds move towards 4Mbps. (We thought it
was interesting that just yesterday Wanadoo announced a new 512kbps ADSL product
with a 5GB download cap.)
Digital TV households in Western Europe 2010

Cable DTH DSL Digital terrestrial Total
France 18% 21% 4% 8% 51%
Germany 32% 10% 5% 4% 51%
Italy 0% 19% 4% 10% 33%
Netherlands 55% 1% 3% 1% 60%
Spain 9% 24% 4% 15% 52%
Sweden 52% 20% 5% 5% 82%
Switzerland 64% 1% 3% 2% 70%
UK 17% 32% 2% 15% 66%
Europe 19% 12% 3% 7% 41%
Source: Informa Media Group, 2003
Building something This, to some extent, fulfills the bull market era view that increases in bandwidth will
future proof inevitably be filled by new applications. If we embrace the idea that technological
advances will feed an ever-increasing requirement for higher bandwidth to the end-user,
then there emerges an argument in favor of overbuilding capacity, or providing
infrastructure which is “future-proof”.
Forecasts of residential consumer needs

Requirements Bandwidth
Minimum communication needs
Voice (one line) 64kbps
Data 1Mbps
Video (two TV sets) 10Mbps
Total 11Mbps
Reasonable near future needs
Voice (two lines) 128kbps
Data 5Mbps
Video (two TV sets) 54Mbps
Total 60Mbps
Source: UTOPIA, 2003
Big Broadband is being In short, what we are looking at is what has been termed “Big Broadband,” an
born – and it may be a infrastructure robust enough to handle all foreseeable bandwidth demand from the
painful delivery consumer, and scalable enough to adapt to many unforeseen changes in demand. There
is only one technology in existence to match this description, and it is fiber to the home
(FTTH). The key question surrounding such infrastructure, we believe, is “what will it
cost and who should build it?”
It can be done, and All the benchmarks we have seen in research, and all the conversations we have had
relatively cheaply, but the with people involved in FTTH projects suggests that, subject to the state of existing
investment horizon is a infrastructure, topology/topography, building density, technology choice, and a number
long one of other variables, the cost per home connected for an FTTH project may range from
€1,100 to €1,400, and in cases where the deployment is done in an area of densely
packed multiple dwelling units (MDUs), it can be as low as €800. A 2003 Gartner
research paper in support of a California initiative known as One Gigabit or Bust
(which seeks to create a ubiquitous fiber network in the state, using the public/private
partnership model) estimated the cost to deploy fiber during new construction at well
under $500 per home, and pointed out that this may provide an added value element for
the developer (we will look at the potential role of housing associations later). For
existing homes, the research pointed to a cost of $1,500 per home connected, and
assuming a 20-year usable investment life, the implied monthly cost to amortize the
project cost was estimated at only $6.50 per month per household. This is far below the
monthly cost of DSL to the consumer ($45 - 50). Furthermore, a 2002 report by the
California Public Utilities Commission estimated that applying a Basic Service
(Universal Service) classification to DSL, necessitating a DSL connection to every
PSTN connection in the state, would quadruple the cost of telecommunications
services for all California customers.
Some things are worth Paraphrasing a key figure within one of the next generation Big Broadband network
waiting for projects we spoke to, the private sector may not be willing to accept the low returns
and long durations of such projects, and the benefits accruing to those who support
such projects may not necessarily be visible on a balance sheet. In other words, this
may be a job better suited to government rather than telcos, because the benefits of
such projects are likely to be more general, in the areas of economic stimulus and
overall improvement in opportunities for the population covered by them over the long
term.
Bandwidth to determine The UTOPIA project (see below) white paper from November 2003 offers a historical
the future? analogy in support of its business case. It cites the case of St. Louis, Missouri, which
was a boom-town and major urban center in the early to mid-1800s, but which short-
circuited its own long-term growth prospects by legislating against next-generation
infrastructure investment (bridges and trains) which might have bolstered its position
as a transport hub. Subsequently, economic activity gravitated to the more forward-
looking Chicago, which grew prodigiously thereafter. We have visited St. Louis and
think it is a fine city, but it is admittedly a different world, even today, when compared
to Chicago.
A historical example
Source: Utopia
GDP uplift could be great The One Gigabit or Bust Gartner study forecast a 17% increase in gross state product
(GSP) per capita arising from a “universal” fiber infrastructure over a ten-year period,
and also determined that without such an initiative, the outlook for GSP per capita was
only a 2% rise over ten years. Just to provide a bit of perspective, the Gartner report
provides a ranking of 2001 GDP levels, which reveals that if California had been an
independent country in 2001, it would have sat between the UK and France in terms of
output, in fifth place globally. Clearly, the stakes are very high if so much economic
development potential rests on the state of the infrastructure.
Correlation with Such local and regional issues may be peripheral to telcos, but are crucial to
networked readiness is governments, and this is where the key tension arises. The costs involved in such
high projects may prove unpalatable to any individual carrier, as we think investor pressure
over short-term cashflow generation and shareholder returns may inherently conflict
with the longer-term development agendas of national/regional/municipal governments.
In contrast, for the public sector bodies concerned, the prospective benefits of such
investments may be an obvious “no brainer,” and certainly the apparent correlation
between IT infrastructure, social/institutional infrastructure and GDP per capita would
seem to be quite a clear validation of such a view. Below is a partial logarithmic
regression analysis produced as part of the World Economic Forum’s “Networked
Readiness” report from 2003, which seems to support this view.
Partial log regression, networked readiness vs. GDP per capita
Source: World Economic Forum, 2003
Telcos living in a The development of a dynamic in which investors, telcos, government, business, and
Dystopian world? constituents increasingly disagree over goals, strategy and implementation is one in
which, in our opinion, the position of the telco becomes more tenuous. We think the
best example of how these tensions may be manifested in practice is UTOPIA, the
Utah Telecommunication Open Infrastructure Agency, a partnership formed three
years ago by a group of 18 municipal governments in Utah (including the largest, Salt
Lake City). The UTOPIA municipalities cover a total of 724,000 population, 249,000
households and 35,500 businesses, all of which would be eligible for “future proof”
broadband services under the UTOPIA business plan. UTOPIA aims to construct a
fiber network passing all homes and businesses in its 18 member municipalities,
delivering symmetrical 100Mbps Ethernet services to all customer premises. The
network itself would be opened up on a wholesale basis to service providers, which
would market their own portfolio of services as appropriate, for a fee payable to the
UTOPIA network administration, which will be a freestanding entity mutually owned
by the 18 member governments. So far, UTOPIA is in discussions with a number of
potential service providers, and AT&T has formally committed to the project.
Fingers crossed The UTOPIA infrastructure is estimated to require $450 – 530m in funding for
completion within 3.5 years. The estimated cost per home passed is $1,200, with $500
- 700 estimated for connecting a single home. Roughly 52% of the fiber to be dropped
into homes is achievable via an aerial drop, while 48% will require digging. This
relatively high proportion of aerial drops reduces the overall cost significantly.
Currently, the 18 UTOPIA member cities are in the process of deliberations and voting
on the project, which is expected to conclude in the next six to eight weeks. Assuming
a positive outcome (the smaller and separate iProvo project recently received approval
at local level by a 6:1 margin), the initiative will take its case to Wall Street to seek
funding for the project in the municipal bond market. The issue will effectively be a
securitization of the project’s revenues, backstopped by the municipal governments,
and we think, if successful, it could prove an important impetus to the development
similar plans elsewhere.
I don’t wanna play in Local incumbent carrier Qwest has repeatedly declined to participate in the project as a
your yard service provider, and has been publicly hostile to the project. Qwest’s position may be
understandable for obvious reasons on a short-term view; however, if we accept that
this may be the future direction of the telecoms market, then the prospect of
participating in a cutting edge project for zero capex outlay is probably a good learning
opportunity in our view. If UTOPIA is the new reality for telcos, there probably needs
to be a major change in management mentalities if the telcos are to make the most of a
fluid situation.
FTTH? Let’s go Dutch! One sign of some relatively forward-looking thinking from a European incumbent
came late last year in the Netherlands, in the form of the Delta Plan Glas proposal by
KPN. In it, KPN proposed that, to avoid duplication of effort and infrastructure in
deploying FTTH, a new entity should be created, which would house the legacy
infrastructures of KPN and the Dutch cable players, and jointly pursue the development
of a single national FTTH architecture. KPN estimated the total build cost at €7.5bn for
80% population coverage, of which a significant amount could be funded via the
internal cash flows of the combined legacy businesses (the draft version of the
presentation in Dutch stated this number as €5bn, though the version currently
available online says only “a significant part of the required €7.5bn investment”). This
€7.5bn estimate amounts to c.€1,100 per home passed, which matches reasonably well
with other benchmarks we have heard quoted.
What role should central KPN advocated that the Dutch government contribute €200 – 400 per line in order to
government play? ensure that services could be delivered at price points comparable to current services,
and suggested that this could be accomplished via an expansion of the existing “PC
from Gross Salary” ICT subsidy facility. KPN estimated that, if 50% of households
were to opt for fiber-based broadband, the government’s likely outlay would be
c.€100m per annum over 10 years. A later section of the document suggests that,
without a coordinated effort between KPN and its cable competitors, the two camps
would continue expanding their respective platforms (DSL and EuroDOCSIS 2.0), and,
having spent a combined €2bn, would still eventually have to migrate to a FTTH model.
Other KPN proposals included the use of European Structural Funds and the
incorporation of municipal FTTH projects into the Delta Plan.
EU Structural Funds a On the subject of Structural Funds, we think this is one option which similarly-inclined
realistic option operators and governments may turn to in future, and the terms for their employment
seem to fit quite well within the framework defined by KPN. The Commission Staff
Working Paper of July 2003 stresses some important criteria which projects qualifying
for Structural Funds should fulfill:
• “To promote the development and structural adjustment of regions which are
lagging behind, or the economic and social restructuring of regions”;
• “To ensure critical mass of users in public administrations whilst avoiding
dependence on one single operator”;
• “Should be limited, in principle, to infrastructure, i.e. installations (dark fibre,
ducts, masts,…) and equipment which is open to all operators and service
providers”;
• “Can remain owned by a public authority, a private entity which provides co-
funding, or by a public-private entity. In all cases, access for all operators to the
infrastructure at non-discriminatory conditions must be granted. In principle,
Community support should not strengthen a dominant position by any operator, or
distort competition rules.”
• “Modalities of renting of the infrastructure to private undertakings will have to be
accurately defined on a case by case basis. In some countries, national regulatory
frameworks are being modified, e.g., local authorities having the right under
certain conditions to become operators.”
Vision not going to plan Thus, subject to the vagaries of EU legalese, it would appear that there is a clearly
defined framework for partial Community funding of precisely the kind of structure
which KPN proposed in the Delta Plan Glas. In the event, the cable players in the
Netherlands have declined the offer, and have opted to pursue independent
VoIP/Ethernet strategies and a coordinated marketing strategy. However, the concept
embodied in the final point cited above, i.e., the role of local authorities as “operators,”
is gaining momentum in the Netherlands, and we believe developments there may
serve as an important example for the development of Big Broadband in Europe.
During the preparation of this report, we were treated to a visit by a consultant

involved in many of the municipally backed FTTH projects currently ongoing in the
Netherlands, who gave us some valuable, ground-level insights into developments
there. By way of general background, it is important to note that the municipalities
involved have had a couple of common interests in pursuing their FTTH strategies:
• Economic development – Whether it be the remote regional centers of the

Northern Netherlands or the Rotterdam waterfront area, one key driver of FTTH
deployment is to create an infrastructure capable of attracting and retaining
investment from new businesses (in the case of the Rotterdam waterfront, there is a
focus on attracting the motion picture industry, in the Northern Netherlands it is to
some extent focused on attracting new business to the area to replace the jobs
eliminated when KPN decided to re-centralize its corporate functions to Den
Haag/Amsterdam).
• Fatigue with redundant carrier infrastructure projects – KPN makes reference in its
Delta Plan whitepaper to the annoyance of the municipalities over repeated
digging-up of city streets for the provision of new duct space, and this indeed
would seem to be a key motivator for policy formation. In fact, in extreme cases,
sections of streets in major Dutch cities may contain as many as 95 separate ducts,
a level of concentration which has actually led to a material degradation in the
structural integrity of these streets – which are city assets, after all. In response to
the glut of infrastructure, a draft telecom law before the Dutch parliament
reportedly imposes a disincentive to new digging - a provision that new duct space
either be filled within four years, withdrawn, or a penalty fee paid to the local
government.
At present there are the following projects in various stages of development in the
Netherlands, each of which sheds some light on the issues which may drive similar
projects elsewhere in Europe. The Dutch central government has invested €8m Euro in
various broadband stimulation projects, under which eight cities have received money
to speed up development and possible deployment (www.breedbandproeven.nl).
• Rotterdam – After the failure of Bredband Benelux (a unit of B2) and its plans for
FTTH in Rotterdam, the City of Rotterdam decided to invest €4.5m in two areas
for FTTH covering 7,000 homes and 70.000m2 of office space. The project creates
an “open infrastructure model” of the access network, a choice which has
reportedly allowed the creation of an infrastructure wherein bandwidth allocation
is carried out dynamically in real time, and network maintenance for 7,000 homes
can be handled by only a few full-time employees. The Rotterdam project is
entirely funded by municipal funds. Rotterdam is also considering expanding the
existing projects to 300,000 households (all of greater Rotterdam).
• Amsterdam – We understand that local authorities in Amsterdam have given the
green light to FTTH deployment throughout the city (400,000 households) under
the control of a public-private partnership (PPP) structure comprised of the City,
local housing associations and banks. The city has initially committed to a
deployment in 40,000 homes before going for full coverage.
• Den Haag – The city has developed a plan to cover all households, via creation of
a new joint venture between the PTT and local cable company. Interestingly, we
understand that certain proportion of project funding is earmarked to fund the
removal/decommissioning of existing PSTN/coaxial infrastructure.
• Eindhoven – This is the center of a central government funded experiment known
as Kenniswijk (“smart community”), under which it was originally envisioned that
48,000 homes would be connected to broadband at speeds of at least 10Mbps. For
service providers meeting the access speed criterion, there is an annual subsidy of
€800 per activated connection (€500 for the infrastructure and €300 for the ISP
service in the first year) available under a €50m centrally funded facility. At this
point, however, the conflicting agendas of the various partners in the project have
led to a low penetration of under 2,000 homes. This implies that there is c.€45m in
potential subsidies available to service providers through 2005, after which it will
be withdrawn.
• Leeuwarden – This northern regional center had no fiber infrastructure, but pooled
interest from local non-profit organizations and government institutions to amass a
22.5km fiber network connecting 36 organizations.
• Tilburg – This project is very similar to Leeuwarden, though schools were given a
choice of fiber, DSL or coax connections.
Bringing home the The last, but potentially very important development, which we have highlighted in a
bandwidth late 2003 e-mail, is the plan by regional housing association Portaal to deploy FTTH to
its properties. Though the project is short on detail at present, we nevertheless believe
the company to be in an important position in the Big Broadband model. Of the
estimated 7m homes in the Netherlands, 3m are estimated to be owned by non-profit
housing associations such as Portaal, which covers 55,000 homes in five major cities in
the Netherlands. We think there are both commercial (incremental revenue streams,
enhanced property values) and social (bridging the Digital Divide, potential access to
national or Community project funds) motivations which work in tandem to make the
deployment of FTTH an attractive prospect for housing companies, and not only in the
Netherlands.
Sweden provides an Developments in the Netherlands are in an early stage, though if both the Rotterdam
example of success and Amsterdam projects are fully built out, the result will be a 10% household
penetration of FTTH in the Netherlands, with limited, or possibly no involvement on
the part of the incumbent. We are still some way from such a situation, but to get an
idea of where the various Dutch projects may be leading, we should look at what the
Swedish market has achieved since the founding of STOKAB in 1995. STOKAB was
formed by the Stockholm regional governmental bodies to create a wide area fiber-
based Ethernet network, which could sell capacity on a wholesale basis to independent
service providers. Following an estimated SEK2.5bn (c.€xxx) investment in
infrastructure, STOKAB today comprises 5,000km of ducting, containing 1m km of
fiber. STOKAB has developed beyond its initial mission, to connect to various regional
city networks. The Swedish Urban Network Association (SUNA) estimates that over
200 of Sweden’s 290 municipalities have some sort of municipal network
infrastructure, many of which are linked into a national fiber network of one
description or another. SUNA believes that, through the combined efforts of the
municipal networks and ongoing initiatives by the national railroad and power grid, an
alternative national fiber infrastructure will take shape over the next five years.
STOKAB network map
Source: STOKAB
MDU paradise Sweden’s urban topology is positive for the development of such a model. Of
Sweden’s estimated 4.2m households, 47% are housed in multiple dwelling units
(MDUs), and roughly 8% (300 – 400,000) of total households have been connected
with fiber, either to the building or to the home itself. Again, as in the Netherlands, the
potential role of the housing associations in playing a part in this process must not be
underestimated. Approximately 20% of all housing stock in Sweden, and 1/3 of MDU
dwellings, are managed by SABO, the Swedish Association of Municipal Housing
Companies. There is arguably a strong incentive for such associations to get involved
in deploying Big Broadband, either as a strategy for differentiating and increasing the
value of the property, or for gaining incremental revenues via partnerships with
infrastructure and service provider partnerships, or both.
A vibrant market place In the Stockholm area, the existence of the STOKAB initiative has led to a market
served by 65 operators making use of its infrastructure. Most of these are involved in
the carrier or business markets, but the infrastructure has also fostered the creation of a
handful of consumer service providers, of which the best known is B2 (see page 79).
The head of SUNA provided us, as an example, with the pricing typically available to
consumers under such a model, using his own household as an example:
• 10Mbps symmetrical internet access for €30 per month

• IP telephony for €8 per month, with 30 hours of inclusive national calls, and fixed
to mobile calls at half the price level of PSTN players
Beyond the benefits being enjoyed by individual consumers, we enquired as to the

broader economic benefits of the municipal network to the Stockholm region. As far as
we could determine, there have been no formal studies done on this subject, but SUNA
estimates that, for every SEK1 invested in STOKAB, other parties (service providers,
businesses) have invested SEK5 – 6 in the area, and we believe that the choice of
Stockholm as the site for banking giant Nordea’s consolidated group IT center was
influenced by the competitive infrastructure and pricing to be had in the city.
What if there’s no momentum in the municipal government?
In the absence of municipal policy formulation to drive deployment of Big Broadband
as a social and economic development tool, others have chosen to employ a more
grassroots approach. Though operating, in some cases, without any governmental
funding at all, and using volunteers during the early phases, at least, the achievements
of such initiatives have been no less impressive, and are worthy of some attention.
People Power in Denmark provides a couple of good cases of do-it-yourself networking, one urban and
Denmark one rural, both of which highlight the potential for properly-funded projects to supplant
the incumbent, and provide affordable broadband services, in some cases at a
dramatically higher performance level.
Bryggenet (www.bryggenet.dk) is an initiative of the residents of the Islands Brygge

district of Copenhagen. The area is a traditional working class waterfront district,
which has undergone a renaissance over the past twenty years, and is now viewed as
one of the up-and-coming districts of Central Copenhagen for young professionals to
locate. The area is dominated by apartment buildings of six floors, housing 200 – 300
apartments each.
Location map of Islands Brygge (circled)
Source: ebutik.com
Telco attitude The project was begun in early 2001, when a number of buildings housing local
residents were put up for sale, and subsequently bought by the residents’ associations.
Building on a long tradition of community activism, a small group of residents came
together to pursue the idea of collective purchasing power in the provision of telephony,
CATV and broadband internet services to the area. Starting out with 13 buildings of
200 – 300 apartments each (for a total of 2,955), the group approached incumbent TDC,
which had coaxial cable installed throughout many of the buildings, to investigate
leasing/purchasing the plant. Bryggenet was flatly refused access to the cable plant
either on a buy or lease basis. Roughly 10% of residents have TDC cable boxes
currently, and Bryggenet put forth a couple of proposals for distribution of TDC’s
programming, which were either refused outright, or accepted at price levels which
were financially infeasible for an organization which is essentially run by volunteers.
Cost of DIY in line with Left with no options, Bryggenet developed a business plan which involved pulling
other estimates entirely new coaxial cable through the participating buildings, and fitting each
apartment with two RJ45 connections and two coax ports. The group sought and
received bank funding for the project at an attractive rate, given the attractive credit
profile of MDUs in the eyes of Danish lenders (particularly, presumably, if they are in
an area undergoing a revival). The group has its own satellite downlink center, leases
an inbound fiber from Telia, and has laid its own outbound fiber. Since launch in 2001,
the project has seen new buildings join, and Bryggenet’s infrastructure now reaches
3,300 apartments in the area, offering telephony, CATV/radio, and fast internet. The
total cost of the project to date is estimated at DKK25m (€3.4m), inclusive of tax, split
roughly equally between the core investment (head end, IRU, fiber build, etc.) and the
in-building cabling and connections. This seems to imply a cost per home connected of
just over c. €1,000 per home, though we think the incremental cost will be much lower
as the network eventually encompasses the c.1,500 new apartments being built in the
area.
Paying up for the public Of the 3,300 homes connected, c.2,200 take telephony service, c.2,300 take TV, and
good c.1,900 take fast internet, with around 2,000 triple play subscribers. This equates to a
triple play penetration of 61% (NTL and Telewest in the UK, by contrast, have
achieved triple play penetration of c.20% and c.15%, respectively). Pricing is
extremely attractive, though the charging structure requires some explanation, as it may
be unfamiliar to many. Every resident of the participating buildings pays a DKK75
(€10) monthly fee to enable the service, irrespective of whether they use it or not
(apparently only eight or nine out of 3,300 residents resisted - this may say something
about the state of the social contract in Denmark!). The DKK75 fee splits into a
DKK30 payment relating to the installation and maintenance of the connection into the
home (two RJ45s, two coax ports), which goes to the building owners (the residents)
and a DKK45 payment related to the creation and maintenance of other elements of the
project (which goes to the Bryggenet project).
On top of the DKK75 monthly maintenance fee, services are charged as follows:
• Telephony monthly line rental – DKK39 (€5.23), versus TDC’s standard DKK
monthly fee of DKK95.30. Provisions in Bryggenet’s wholesale supplier
contract ensure that it’s international rates will always be 22% below TDC’s,
and national rates are 24% lower for both peak and off-peak minutes;
• Basic CATV – eight channels/20 radio stations, DKK39, versus c.DKK119 at
TDC;
• Premium cable - DKK142 (€19.06) versus c.DKK260 on TDC for a
comparable package;
• Broadband internet – DKK65 (€8.72) for lower bandwidth, DKK135 (€18.12)
for higher bandwidth.
More bang for your DKK Currently Bryggenet has 1,200 subs on the slower broadband service (10Mbps shared
connection between 1,000 users) and 700 on the faster (45Mbps shared connection
between 1,000 users). Our conversation with one of the Bryggenet project volunteers,
himself a user of the service, revealed that lower speed users average 4 – 8Mbps on
average, though at peak traffic times this occasionally drops below 2Mbps per second.
Even when we take into account the monthly maintenance fee of DKK75, the cost of
bandwidth for the Bryggenet user is still staggeringly cheaper than for ADSL users,
and even at a downstream speed of 2Mbps rather than the 6Mbps we assume here, the
price per Mbps would still be 78% below that of the TDC 2Mbps product. The
comparison with the bundled ADSL/PSTN offerings from TDC is also highly
competitive.
Pricing comparison – Bryggenet vs. TDC ADSL and bundled ADSL/PSTN
TDC Bryggenet
Cost per downstream Mbps Broadband "lite" at 6 Bryggenet discount
ADSL DKK (DKK) Mbps + monthly fee to TDC per Mbps
256/128 kbps 280 1120 23.3 -97.9%
256/256 kbps 316 1264 23.3 -98.2%
512/128 kbps 348 696 23.3 -96.6%
512/512 kbps 468 936 23.3 -97.5%
1024/256 kbps 428 428 23.3 -94.5%
1024/512 kbps 548 548 23.3 -95.7%
2048/512 kbps 628 314 23.3 -92.6%
Broadband "lite" + Bryggenet discount to

ADSL/PSTN bundled* DKK phone + monthly fee TDC
256/128 kbps 383 179 -53.3%
256/256 kbps 415 179 -56.9%
512/128 kbps 447 179 -60.0%
512/512 kbps 567 179 -68.4%
1024/256 kbps 527 179 -66.0%
1024/512 kbps 647 179 -72.3%
2048/512 kbps 727 179 -75.4%
Source: DIR, Bryggenet, Tarifica * Includes 15% discount on PSTN calling
We think that Bryggenet is far from an isolated example of Big Broadband being
delivered being delivered by someone other than a telco. Bryggenet told us there are at
least two other projects live in urban Denmark currently, one of which is on a larger
scale (c.5,000 apartments). Away in the sparsely populated hinterlands lies
DjurslandS.net, another example of DIY networking.
Green Acres is not the DjurslandS.net (www.DjurslandS.net) - This network covers parts of the Djursland
place to be – if you want region of Denmark, which has a population density of only 57.6 per km2, less than half
broadband the average for Denmark. In the past five years, it has seen the closure of one ferry
harbor, the local newspaper, one regional hospital, and plans hatched to close another
ferry harbor and a regional airport. ADSL population coverage in the region is only 75
– 80% versus 95 – 98% in Denmark as a whole. DjurslandS.net was formed in mid-
2001 as an umbrella organization to coordinate the construction and management of a
broadband access network for the eight municipalities constituting Djursland. Its
ongoing task is in the area of supervision, service, support, product development and
also to deliver benefits of scale in purchasing.
Wi-Fi in the local loop The closure of the regional hospital meant that a fiber ring previously installed to
delivering 1 – 2Mbps connect the facility to local municipalities (and to the national backbone as well)
became redundant, and this now forms the backbone of the DjurslandS.net project, via
a 15-year exclusive IRU. The current fiber connection is 10Mbps, burstable to 20Mbps
(and very scalable – this is fiber after all), and each local municipality is connected to
the DjurslandS.net coordination center at 10Mbps. Currently the network guarantees
256kbps, but typically users achieve a range of 1 – 2Mbps. The last mile connection is
served by an omnidirectional Wi-Fi antenna (radius >1.5km), beaming to Wi-Fi access
points in the users’ premises set to client mode. These are equipped with a €30 antenna
(manufactured by Terma) with gain of >8dBi, expandable to >14dBi with some home-
brew modification.
Terma Wi-Fi antenna, normal and modified
Source: DjurslandS.net
An 84% discount to the The most recent information we have shows five of the eight regional municipalities on
telco line, with over 600 users connected in 30 locations. Users pay an initial connection fee
of DKK2,000 (€270), or a monthly fee of DKK 100 (€13.50) over two years, to cover
installation and infrastructure costs. After that, the monthly charge is €13.50 for
unlimited access. For a guaranteed access speed of 256kbps downstream, this is a 64%
discount to the slowest TDC ADSL product on the market, and at the commonly
attainable speed of 2Mbps, the discount to TDC is 84%. All this, for people who
despaired at ever having a broadband connection of any type!
DjurslandS.net coverage map, September 2003
Source: DjusrlandS.net
Sanity check
In the preceding discussion, we would not like to give the impression that we embrace
some naïve utopian vision of the world. It is clear that, from Utah to the urban
Netherlands, and on to rural Denmark, the creation of, or success of, these kinds of
projects is dependent on ingredients not readily found everywhere. We think that a
certain vision is required of regional or local government, even if it is a self-serving
one (economic and social development). Beyond that, residents and their elected
representatives must possess the conviction and bravery to incur the necessary costs,
whether it be the monthly maintenance fee of Bryggenet building residents, or the
fiscal risk associated with a municipal bond issuance of the kind which UOTPIA seeks
to pursue. However, we suspect that only a handful of successful precedents would be
required for less progressive or adventurous governments to move in the same
direction, particularly if there is compelling evidence of a material change in the
economic fortunes of municipalities/regions adopting such strategies. For the telcos,
the issue is a complex one, because we believe that while participation in such
collaborative projects may be more efficient and beneficial from the cash outlay
standpoint, there is an intrinsic risk in the short term of lower revenue share and
margins. Alternatively, there is a potentially larger risk in not being involved, and it is
often referred to by a bubble era term – dis-intermediation.
• We believe sources of pressure are mounting on the incumbents’ control over
access, and will generate significant financial risk.
• Mobile substitution in access is already at a significant level, and if it
accelerates, the incumbents will face market share loss and a negative margin
trade on recaptured revenues. We think evidence suggests that mobile
profitability is still disproportionately reliant upon termination revenues, and
therefore vulnerable to regulatory risk.
• Alternative access technologies are growing more robust, and may strengthen
the hand of cable players and informal ad hoc networking initiatives.
• We think the regulatory risk in access is to the downside as consensus builds
around the need for WLR products and the EU seems to see progress in the
broadband market as being driven by ULL price cuts.
• We think the economic interests of local and regional governments may
increasingly be in conflict with PTT “ownership” of a broadband market based
on ADSL. Developments in pockets of the US and Europe demonstrate a growing
tendency to seek solutions via carrier-independent FTTH deployments funded in
public/private partnerships, access to municipal debt markets, direct government
funding, or community activism. The role of the telcos is increasingly uncertain
in such an environment.
6. Challenges for the telcos – the service layer
As broadband proliferates, the range of choices for consumers grows, and many of
these supplant the role of traditional telcos in the service layer.
VoIP in the European enterprise

We have written an awful lot over the past year about the many shades of VoIP, and
we are not alone. The press coverage of the issue, particularly in the US, has been
nothing short of staggering in the past six months, and we think the IPO of Iliad in
France will help to heighten awareness of the issue in the consumer space in Europe,
along with the arrival of Vonage and other SIP service providers on the scene in the
first part of this year. On the enterprise front, it is clear that the technology is already
firmly on the European map, and set for dramatic growth over the next decade.
VoIP is in the office Figures from IDC in 2003 showed a total of 4.7bn outgoing IP voice minutes from
enterprises in 16 European countries in 2002, of which 46% were external calls carried
on/terminating on the PSTN, and 42% were internal calls. According to IDC’s
estimates, outgoing IP voice minutes from the enterprise are set to grow at a CAGR of
79% between 2002 and 2007, with the mix shifting significantly away from calls
carried on/terminating on the PSTN (only 26% in 2007), in favor of internal calls
(50%) and inter-site WAN/IP service provider calls (at 11% respectively). The latter
two groups are expected to see the strongest growth (CAGR of 91% for inter-site
WAN, and 121% for IP service provider calls).
Estimated outgoing IP voice minutes from Euro enterprises, 2002 & 2007
45000
40000
35000
30000
Ext ernal via t he P ST N
25000
Int ernal only
Int er sit e over a W AN
20000
Ext ernal via an IP service provider
15000
10000
5000
0
2002 2007
Source: IDC, 2003
Proportion of outgoing IP voice minutes in European enterprises ‘02 & ‘07

Internal only
Inter site over a Inter site over a
42%
WAN WAN
8% Internal only 11%
External via an IP 50%
service provider
4%
External via an IP
service provider
11%
External via the External via the

PST N PST N
46% 28%
Source: IDC, 2003
Interest and potential Such a mix shift seems broadly consistent with data on adoption and interest levels, as
adoption rates look seen in the British Chamber of Commerce Business Broadband report from September,
promising 2003. It found that IP telephony was already in use in 42% of UK businesses with more
than 250 employees (and thus more likely to need to migrate traffic between office
sites away from the PSTN where possible). An additional 23% of these large
businesses said they intend to use VoIP in future, suggesting an adoption rate in larger
UK businesses approaching 80% in the next three to five years, in our view. Further
insight comes from an IDC 2003 survey of enterprise WAN managers in Europe,
which revealed that 50% believed the main attraction of VoIP is in reducing long
distance costs. However, c.15% of managers stated that a primary appeal was the
unique feature set offered by VoIP (conferencing, unified messaging, “follow me,”
etc.), and another 10% said that easier use of PABX facilities (we presume this means
both line provisioning and general PABX management) was a key driver. Thus we
believe that, as adoption increases driven by short-term cost savings, enterprises will
realize incremental benefits through enhanced productivity and lower running costs,
further accelerating the migration to VoIP. Returning to the BCC survey, we found the
levels of penetration and interest among firms of 50 – 249 employees to be surprisingly
high, suggesting that this segment of the market may be moving toward 50% uptake in
the medium term.
Adoption/plans for adoption of VoIP in UK companies of different sizes
45%
40%
35%
30%
25% Currently use
20% Plan to use
15%
10%
5%
0%
Self- 1 - 19 20 - 49 50 - 249 250+
employed persons
Source: British Chamber of Commerce, Business Broadband, September 2003
What this means for the market, and over what time period, are complex questions, but
we think there are already some early signs of the implications in UK market data. The
UK is the only market we cover where the regulator routinely splits the business and
residential markets in detail. This allows us to see some fairly alarming trends in the
business market, as shown in the chart below. Though in a clear downward trend, the
overall UK market in calls and access (as measured by OFTEL) generates £2.9 – 3.1bn
in revenues every quarter, split roughly 2:1 in favor of call revenues. While volatile,
the business segment’s share of access revenues has been basically flat over the eight
quarters shown below (at around 44%, growing in line with access revenues generally),
the share of business call revenues has declined sharply, with revenues declining at
roughly twice the rate of decline for the overall market. This issue itself is probably
deserving of a separate research project, but our preliminary assessment is that 42%
adoption of VoIP in larger enterprises is probably one key driver of this decline, and at
some point we expect to see this feeding through to the access segment of the market,
particularly as smaller enterprises begin to accelerate deployment. We question how
many industry analysts currently have this sort of scenario explicitly in their forecasts.
UK market total revenues and business segment’s share of calls and access
3,200 45.0%
3,150
44.0%
3,100
43.0%
3,050
3,000 42.0%
2,950
41.0%
T ot al UK market revenues - calls and access
2,900
UK business share of access revenues (RHS)
40.0%
2,850
UK business share of call revenues (RHS)
2,800 39.0%
Q3 2001 Q4 2001 Q1 2002 Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003
Source: DIR, from OFTEL data
Access-tied consumer VoIP services

In the consumer market, we have seen a handful of examples of VoIP services tied to
provision of access. The earliest case was Cablecom’s Swiss market launch, still in a
controlled rollout phase. We understand that Dutch MSOs UPC, Casema and Essent
are in various states of readiness for VoIP deployments, and we expect to see more
such developments over the next year in cablecos independent from the incumbent
telco (UPC Austria, for example). We have also seen the success of Free in the French
market, based on an unbundled DSL strategy and a pricing approach very similar to
that employed by Yahoo! BB in the Japanese market. Concurrent with the launch of
Free’s Freebox service, Swedish metro-Ethernet service provider B2 launched its own
IP telephony product, with similarly aggressive pricing. Based on the examples we
have seen to date, we rank countries with high broadband penetration, a high
proportion of broadband cable subs, and a relatively competitive DSL market as being
at high risk from VoIP attack.
VoIP risk assessment in Europe – from the perspective of the incumbent’s position
B’band Penetration Cable/other share 3rd party ADSL share Penetration Cable ADSL Regulation Risk rating
Germany 5.12% 1.65% 7.18% 3 1 1 2 1.7
Netherlands 9.14% 58.91% 31.20% 5 4 3 5 4.1
UK 3.67% 50.50% 49.23% 3 4 4 5 3.8
France 4.07% 14.12% 32.40% 3 2 3 4 2.8
Spain 4.29% 25.29% 24.38% 3 2 2 2 2.3
Italy 3.02% 16.61% 32.28% 2 2 3 3 2.4
Portugal 3.52% 70.02% 19.37% 2 2 2 2 2.0
Switzerland 5.66% 49.35% 43.83% 4 4 4 5 4.1
Austria 5.94% 58.61% 20.79% 4 4 2 3 3.3
Norway 5.44% 24.83% 41.37% 3 3 4 4 3.4
Sweden 9.04% 40.27% 25.76% 5 3 2 4 3.4
Denmark 10.35% 28.71% 20.67% 5 2 2 4 3.1
Finland 6.63% 18.37% 21.79% 4 2 2 4 2.8
Belgium 10.62% 38.37% 15.69% 5 3 2 3 3.3
Source: Company and regulator data, DIR estimates
Cablecom seeing interest Cablecom now has c.22,000 VoIP subscribers out of a total broadband user base of
well above rate of net over 200,000, and is seeing 3 – 4,000 monthly net adds on the VoIP product.
additions Management state that interest is far greater than the net addition rate, reflecting the
prudent deployment strategy. The churn rate seen in the VoIP product is below
company expectations to date, and the company sees a high correlation between the
availability of the voice product and broadband takeup (this echoes similar comments
in the past from Vonage). Commensurate with the VoIP initiative, Cablecom last
November increased access speeds on its cable modem services, delivering a 2Mbps
and 3Mbps product at price points comparable to its previous 1Mbps offering on a
price per megabit basis. (Subsequently, Swisscom announced in December that, from
February 2004, it would double access speeds on existing ADSL products at no extra
cost to existing customers, and move to introduce a 2.4Mbps product.)
Head-to-head price comparison of Swisscom and Cablecom pricing (CHF, all including VAT)
Cablecom Swisscom Savings for Cablecom user Comments
PSTN line rental 20 25.25 20.8% Marketed as a primary line, but in fact is VoIP using the cable
modem.
Second line rental 10 25.25 60.4%
National call peak 0.03 0.08 62.5% All Cablecom calls to the PSTN are charged at, or below,
Swisscom’s off-peak rate. Cablecom calls carry a CHF0.08 set-
up charge, but are billed per second. Swisscom bills in CHF0.10
increments, with rounding.
National call off-peak 0.03 0.04 25.0%
Call to mobile peak 0.45 0.55 18.2% All calls to mobile are at Swisscom’s off-peak rate.
Call to mobile off-peak 0.45 0.45 0.0%
International weekdays 10 12 16.7% All international calls are at Swisscom’s weekend rate.
International weekends 10 10 0.0%
Source: Company data, DIR
Free: Yahoo! BB with a Iliad SA (which floated today on the Paris Bourse) subsidiary Free’s ADSL strategy is
French accent based entirely on unbundled access, which accounts for 34% of its c.500,000
subscribers. The company has developed its own IP DSLAMs and modems based on
Linux, for use with its Freebox service. The Freebox modem is available for free for
customers within 2.5 kilometers of a co-lo or dedicated Free facility, while those
outside of this range are supplied with a standard Sagem ADSL modem. Either option
carries a monthly ADSL subscription fee of €29.99, 25% below the FT 1Mbps offering.
Freebox subscribers are guaranteed a theoretical download rate of 8Mbps, and the box
contains a USB port, Ethernet socket, two telephone sockets for voice over DSL
services and a scart connection for TV signals. The Freebox voice over DSL service
launched in August 2003 and is offered to all 100,000+ Freebox subscribers. Pricing is
extremely attractive, with no minimum or set-up charges, per-second billing, and rates
significantly lower than competitors in the French market. The one area where
Freebox’s headline pricing is more expensive is in off-peak mobile, but we think the
company is counting on the transparency of its pricing (no minimum charges), as well
as the free domestic calling element, to draw users.
Freebox modem schematic
Source: Iliad SA
Free.fr Freebox pricing comparison

Freebox FT Tele2 One.Tel Comments
Local calls free (up to 10 hours per 0.033 0.014 0.010 minimum charges 0.091 to
month, then 0.01 per 0.122 by competitors
minute)
National calls free (up to 10 hours per 0.091 0.034 0.01 minimum charges 0.110 to
month, then 0.01 per 0.122 by competitors
minute)
International calls From 0.03 (UK 0.22 0.08 0.059 minimum charges 0.110 to
example), per-second 0.130 by competitors
billing
Calls to other free cost of local cost of local local call
Freebox call call
subscribers
Mobile (all
networks)
peak 0.19 per minute, per- 0.23 - 0.29 0.22 - 0.334 0.22 - 0.32 minimum charges 0.220 to
second billing 0.340 by competitors
off-peak 0.19 per minute, per- 0.11 - 0.14 0.08 - 0.174 0.08 - 0.16 minimum charges 0.225 to
second billing 0.340 by competitors
Pay channels and from 0.49 per month na na na
packages
General TV free na na na
channels
Source: Iliad SA
B2 leveraging the Stockholm-based metro Ethernet operator B2 Bredbandsbolaget launched a residential
STOKAB promise IP telephony offering in August 2003. The service now has 37,000 telephony
subscribers, out of a total customer base of 135,000, for total penetration of 27% just
five months after launch. Line rental of SEK99 is 21% below that of Telia, and
discounts on calls range from 8 – 26% on average versus Telia and Tele2. As with the
Freebox service, on-net calls are free. Subscribers lease equipment, and the contract is
on a one-month rolling basis. The service comes as a pure incremental revenue
opportunity on top of B2’s core broadband access service, which typically offers
10Mbps Ethernet for SEK320 per month (Telia 2Mbps ADSL carries monthly fee of
SEK380). The company is currently compiling customer interest for a planned launch
of 100Mbps Ethernet services in April.
B2 IP Telephone service pricing comparison (SEK)

B2 Telia Tele2 Average savings
Monthly line rental 99 125 125 21%
Call charges
On-net calls free n.a. n.a.
Fixed national - peak 0.19 0.23 0.20 12%
Fixed national - off-peak 0.10 0.12 0.11 11%
To Telia Mobile - peak 2.25 2.50 2.45 9%
To Telia Mobile - off-peak 1.35 1.50 1.45 8%
To Vodafone - peak 2.66 2.95 2.90 9%
To Vodafone - off-peak 2.03 2.25 1.75 -1%
To Comviq/other - peak 2.35 2.95 2.40 12%
To Comviq/other - off-peak 1.35 2.25 1.40 26%
Source: B2 Bredbandsbolaget
Access-independent VoIP
We expect the European market in 2004 to witness the arrival of next-generation voice
services in the guise of SIP-based application service providers who are independent of
the access market. Readers of our research will be familiar with our work on Vonage
(Eurotelcorama No. 3, 3rd November, 2003), and thus we will not go into detail here as
to the mechanics of the SIP ASP business model. However, we should stress the
challenges that such operations pose to incumbent telcos by citing a few of their key
characteristics in contrast to the incumbents:
• Asset base - The current flock of SIP-based service providers operating in the
US (and those gearing up for launch in Europe) operate on very thin asset
bases. As they are not involved in access provision, have no legacy switches,
and largely use software and hardware elements available off the shelf, the
cost of building a network is exceptionally low. In the case of Vonage, the cost
of network deployment to enable coverage of the entire US and allow
termination in most countries of the world is estimated at $xxm. By contrast,
AT&T had net PP&E of $24.7bn on the balance sheet in Q3 2003. The
comparison is an unfair one, as the companies come from different places and
do different things. However, the characteristics of the SIP-based service
(portability, multiple virtual phone numbers, free on-net calling) have the most
obviously serious implications for a long-distance provider such as AT&T,
which in Q3 2003 derived 60% of operating income from its consumer
division. If there are implications for the core business, then there must also be
implications for the value of legacy equipment on the balance sheet. We think
the incumbents face significant levels of write-offs related to redundant
equipment and facilities in future. Implied asset lives are trending down
generally, but we question whether this is enough.
Implied asset lives (Q1 2002 = 1)
1.30
1.20
BT
1.10 DT
KPN
1.00 SWISSCOM
PT
0.90 TEF
TI
0.80
0.70
Q2 2002 Q3 2002 Q4 2002 Q1 2003 Q2 2003 Q3 2003
• Operating costs – We have done some speculative modeling based on our

discussions with various SIP service providers, and this exercise points to a
gross margin on services (before CPE costs) of 70% and higher, depending on
the scale of the business. The margin on value-added services (real-time self-
provision of a new phone number, privacy features, etc.) is above 95% in most
cases, we believe, and feature enhancements to the entire user base can be
added in a matter of minutes. Most account management issues can be dealt
with by the subscriber online, so no physical bill is generated, which we think
is another key source of cost savings. The result is a very small headcount – in
the case of Vonage, around 300 currently, and the company indicates that at a
level of 1m US subscribers it could operate comfortably with total headcount
of 500. Our preliminary work leads us to the conclusion that a similarly-
situated operator of moderate size (2 – 3m customers) would generate revenue
per employee three to four times higher than the average for our incumbent
coverage universe (€350,000 per annum). The corollary of this, in our view, is
that the retail divisions of the incumbent telcos may be three to four times too
large, either in facing a serious onslaught from SIP players, or in the event that
they adopt the technology wholeheartedly themselves.
You can only fire so We don’t expect the operators to roll over and die, but we do think that the challenges
many so fast of adapting to the changing dynamics brought about by VoIP will entail some hefty
cash outflows. Just for the sake of argument, a back-of-the-envelope redundancy cost
estimate for 75% of a 100,000 strong retail division would amount to c.€3.8bn. Even
phased over five years, we’re talking about an average annual charge of €750m to
achieve it. Such a level of payout would be equivalent to the total annual dividend
disbursements of some companies in the sector, not to mention the level of upheaval
such a move might cause in certain less-flexible labor markets.
BT’s response Nevertheless, the VoIP genie is out of the bottle, and the telcos’ response to it will be
interesting to watch. Already we have one example of a pre-emptive move into the SIP
space by an incumbent, but with important differences which highlight the “change
management” issues which the telcos face. BT Group has launched "Broadband
Voice," a Vonage-like service initially targeting subscribers of NTL and Telewest who
have broadband connections and subscribe to the Talk Unlimited (NTL) and Talk
Evenings & Weekends (Telewest) discounted calling plans.
Broadband Voice conceptual diagram
Source: BT Group
Available exclusively online, the service costs £7.50 per month, and includes an
analogue telephony adaptor (from Cisco or Fujitsu) free of charge for those who sign
up before 31st, March 2004. This monthly price compares with Telewest's £7.50 and
NTL's £9.00 per month for comparable discounted evening/weekend calling packages.
It also offers "unlimited" weekend and evening calls on similar terms to the cablecos'
plans and its own Together packages (free for up to 60 minutes, after which a 1p per
minute charge kicks in unless the number is redialed). The initial press release
indicated aggressively lower pricing on calls to UK mobiles, but the stated discount
currently is 5% off BT standard rates.
BT Broadband Voice pricing

Daytime Evening and Weekend Notes
Local & National UK only 3p a minute free Up to one hour per call
Broadband 05 numbers 3p a minute free Up to one hour per call
International to 17 countries 4p a minute 4p a minute Only to Australia, Belgium, Canada, Denmark, France, Germany, Italy, Luxembourg,
The Netherlands, New Zealand, Norway, Portugal, Republic of Ireland, Spain, Sweden,
Switzerland, USA. 4p per minute price is introductory offer until 31st March, 2004
International mobile 29p a minute 29p a minute Only to same countries as above. 29p per minute price is introductory offer until 31st
March, 2004. USA and Canada mobile calls are 5p per minute.
Source: BT Group
Messages and issues What does the BT product tell us about the likely strategy of the incumbents in
addressing the technology?
• They may selectively embrace it early to win back some business and
thwart competitors. BT is employing SIP as a way to extend the reach of its
BT Together pricing plans to access customers it does not own – cable
customers. Cable operators have been a partner of choice for SIP players in the
US so far, particularly in the case of Vonage. With Vonage clearly intending
to launch in the UK, this would hardly seem to be a coincidence, and the
branding is very similar to what Vonage uses in the US.
• They may tread lightly on pricing. All the SIP service providers we have
come across in the US offer free on-net calls, as do most mobile operators in
the UK, and some access-tied VoIP operators like Yahoo!BB Phone and Free.
However, in this scenario calls to 05 (VoIP) numbers are charged at the same
rate as to the PSTN. We may expect SIP newcomers will make much of the
fact that on-net calls (even international ones) will be free on their services.
Also, the pricing looks geared to preventing cannibalization of BT Together
revenues. Clearly, to an NTL customer paying £9.00 per month for a similar
plan, the BT price represents a savings, but to a BT Together subscriber, the
rates are not compelling enough to shell out an additional £7.50, especially if
on-net calls are classified no differently from PSTN calls.
• The scope of the service may be limited. It is interesting to note that
international calling (at the promotional period 20% discount to NTL) under
this plan only extends to 17 countries, all of which are in the developed world.
SIP players in the US, by contrast, terminate traffic in virtually every country
in the world (and make a reasonable margin on it, we believe), and our
conclusion is that BT has intentionally limited the scope of international
calling to avoid being seen as yet another international price arbitrage service,
and risking additional pressure on its international calling business.
• They may not stress some features. Notably absent from BT’s product
description is the fact that ATA adaptors are portable devices, and that the
number assigned to them is non-geographic. Obviously, BT might not be
thrilled with the prospect of users taking the box to Ibiza on holiday and
making unlimited free calls to the UK, but someone is bound to make the
connection.
Treading a fine line, and In summary, we think the BT product is both aggressive and conservative, illustrating
carrying a big stick the line the incumbent must tread between fighting back and cannibalizing parts of the
core. Broadband Voice offers a new angle of attack on major competitors and positions
the company in precisely the segment where SIP newcomers might focus, but it also
offers less flexibility than some of the offers those newcomers might unleash, and it is
not in BT’s interest to highlight some features of the service which other competitors
will seek to exploit as a unique selling point.
Does PSTN connectivity One of the key reservations which investors have expressed around the VoIP theme is
really matter? in the area of interconnectivity with the PSTN. Most assume it is essential to the mass
adoption, and moreover, the financial viability of such services, and we agree, at least
as it applies to a service such as Vonage, which is marketed more or less as a fully-
fledged telephony product. However, we have also long assumed that, for most
residential users of telephony, the vast majority of calling involves persons falling
within the caller’s social or family network. We searched high and low for some
definitive empirical evidence of this, to no avail, so we undertook the monumental task
of surveying our entire research department in London. The result was that our sample
estimated that 93% of home-based and mobile phone calls were made to persons
falling within their circle.
DIR London user survey – estimated proportion of calls to social circle
13% 7%
13% <90% of calls

90% of calls
40% 95% of calls
>95% of calls
100% of calls
27%
Source: DIR
How wide is the calling Obviously, everyone will, at one time or another, need to order a pizza, call our dentist,
circle? or phone the gas company to complain. However, if most of us are paying hard-earned
money to mostly speak to the same people over and over again, then there is arguably a
huge opportunity to develop and promote closed, free systems for communicating with
our own social network. In practice, we concede, the proportion of on-net calls in the
Vonage service is only 4 – 5% of total traffic on the network currently, despite being
free, though we think this relates principally to the low penetration of the Vonage
service itself. However, as broadband penetration grows, and adoption of VoIP grows,
this becomes less trivial a threat.
Go on, take a bite As previously, we focus on the UK as an example, by virtue of the detailed market data
available. The UK residential calling market generates revenues of £1.0 – 1.1bn per
quarter consistently, with domestic geographic calls accounting for c.30% of this total.
UK household penetration of broadband is around 12% at present and adding one to
1.5 percentage points of penetration per quarter. At present, then, we can say that 12%
of the households in the UK could theoretically move the 93% of their calling
accounted for by the social circle to a closed system (or to free on-net calls from a
VoIP newcomer), which would theoretically remove 3.5% of the revenues from the
residential calling market. At 30% household broadband penetration, we see 8.7% of
current revenues disappearing from geographical calls alone.
Over the past year, we have seen some interesting examples of “closed” VoIP systems
(voice applications which do not interconnect with the PSTN, nor in many cases, with
other similar services) generating significant levels of interest, attention and use.
Recently we have also begun to examine other non-VoIP developments which we think
also have the potential to channel communication between individuals and groups in
such a way as to take it largely beyond the realm of the telcos. We discuss each of
these developments below, and we believe that each of the categories of services we
examine has the potential to work (probably in tandem with the others) to either
displace or obviate some voice usage for the telcos.
Voice/video IM
Regular readers of Eurotelcorama may remember the piece we produced on instant
messaging back in July 2003, in which we highlighted enhancements such as a voice
and video IM packages being rolled out by MSN, Yahoo!, AOL and Apple Computer.
We continue to believe that, given the huge installed user base of IM worldwide
(c.400m), such applications should over time displace a significant amount of voice
traffic on the PSTN. Interoperability remains an issue to be overcome if the individual
IM platforms are to solidify their common positions as alternative voice platforms.
However, we have seen one case already wherein a third party company has achieved
interoperability in establishing an own-label IM platform for a Clear Channel
Communications affiliate (Z100 FM) in New York City. We also believe that, despite
interoperability issues, a single enhanced IM package probably fulfills many of the
communications requirements of users, and that the value-added element of being able
to do video telephony (effectively for free) may serve as a powerful peer pressure
marketing incentive. Finally, we believe that, in the absence of interoperability,
dedicated IM users are likely to use multiple platforms in order to maximize their
network.
Apple iChat AV screen shot
Source: Apple Computer
“Closed” VoIP services
A dark horse of a Running in parallel to the Vonage/Voiceglo/Voicepulse/iConnectHere school of
different color access-independent, PSTN-interconnecting VoIP application service providers, there
exists another strand of the VoIP scene – systems which are either largely, or
completely closed. We have written about each of them individually over the past year,
but just by way of summary, the principal examples so far are:
• Free World Dialup – Free SIP-based community with c.150,000 members

globally (45k in the US, 30k in Israel, Canada, Brazil and the UK round out the top
five countries with ≤ 5% of total subs in each). Currently free, though we believe
that chargeable premium services will emerge in due course. Supports a variety of
interfaces, including hardware SIP phones, softphone clients and Wi-Fi handsets.
Officially no termination on the PSTN, though independent parties have
established gateways into Europe, and free calls to a number of countries were
offered as a special promotion during the Christmas holidays. Sister company
markets a device for forwarding SIP calls via cellular phones. Has been successful
in forming peering arrangements with other services, both commercial and
free/closed.
• IPtel – Free SIP-based community with several thousand users globally, though
exact number is undisclosed. No interconnection with the PSTN. Interconnects
with other SIP communities such as SIPPhone, FWD and Iaxtel.
• SIPPhone – SIP-based community with 50,000 members (combined hardphones
and softphones). Membership is free but users must purchase SIP phones from the
company. Limited interconnection with the PSTN (toll-free numbers, collect calls
and calling card calls accessible). Interconnects with other SIP communities such
as FWD, IPtel and Iaxtel.
• Iaxtel – Free SIP-based community, unknown number of users. Associated with,
and primarily used by, the Asterisk developer community. Some interconnection
with other SIP communities and commercial SIP providers, as well as
interconnection with toll-free numbers on the PSTN.
• Skype – Free P2P voice application using FastTrack protocol common to file-
sharing services. Over 6m downloads since launch in August 2003, with 2.4m
registered users worldwide. No interconnection with PSTN or other VoIP
communities or services. Basic service will remain free, though chargeable
premium services should be introduced soon.
All the services we highlight above have been significant developments in the march of
VoIP, though as we have written previously, we think Skype represents a quantum
change in both its scale and nature. By virtue of the founders’ association with KaZaA,
the service has grown far faster than any before it, and faster than KaZaA did at a
comparable stage in its life. The heavy encryption of the service and its distributed
nature make it a different animal indeed, though we do not believe it will be the last of
its kind. We think the potential exists in other file-sharing platforms to implement a
similar service, and fully expect to see this in due course.
Blogging
Highly personal, highly The weblogging or “blogging” phenomenon has received a tremendous amount of
social press coverage over the past year, and while this is perhaps the least obvious source of
displacement of activity on the PSTN, we think it nevertheless deserves attention as a
new form of network. We would argue that blogging is a dramatically different form of
communication than has occurred in the past, which typically fit into either the one-to-
one, one-to-a-few, or one-to-many categories. Blogging, in contrast, is both a highly
personal, as well as highly social, activity. Responses are invited to postings by the
blog author and exchanges ensue between people not directly involved in the blog itself.
Also, it is common for bloggers to cite other blogs, and increasingly to provide links to
them, within their own blogs. The logging and tracking functions enabled by the RSS
system used in blogs also provides the ability to discern who is viewing and linking to
what. We have no accurate estimate for the current worldwide number of blogs, or
their growth rate, but Blogstreet.com, which tracks popularity and diffusion of blogs,
currently tracks 145,000 blogs. The potential power of the blog is to decentralize
information flows and broaden potential audiences for alternative points of view on
issues, especially among those with shared interests. (One recent industry-specific
example we saw came in a the Wi-Fi Networking News blog [ rating], in
which one author linked in an article from the Seattle Times, reviewing a recent
wireless panel discussion, where a representative from T-Mobile USA reportedly
admitted that the company has the issue of Wi-Fi cannibalization of cellular revenues
“on our mind.” Thus, a throwaway remark in a local newspaper in the US finds its way
to analysts and investors interested in this issue all over the world. We think both the
industry and brokers’ research departments would be wise not to ignore the power that
such a decentralized information flow may ultimately possess.)
Upside in the moblog? While blogging on the fixed line may do nothing to generate revenue for the fixed line
businesses, and in fact may add further impetus to substitution usage patterns (blogging
/responding rather than talking), it is not necessarily all bad news. As clients who had
presentations from us last spring/summer may remember, one of the more interesting
phenomena we have observed in the past year is the rise of the mobile weblog, or
“moblog,” which we estimate now has c.150,000 users in the United States. The best-
known example of this phenomenon is TextAmerica, which was the first site to launch
and estimates it has 100,000 active moblogs currently. We recently spoke with founder
Chris Hoar about the site in particular and the phenomenon more generally.
A clear driver of usage, Mr. Hoar estimates that moblogs hosted by TextAmerica are currently growing at a
for some compound monthly growth rate of 30 – 40%, and that the typical moblogger sends over
100 MMS photos a month (versus their estimate of two per month per non-bloggers),
often at considerable personal expense. Our investigation of the site revealed a
significant number of postings of short videos as well. As such, it is precisely the sort
of service which could bolster the mobile data business case for carriers, should it be
marketed to subscribers more aggressively. We understand that TextAmerica has
recently closed a deal with an unnamed carrier to license its platform and create a
“white label” version to be targeted at its customers. We assume that operators seeking
to move in this direction may tailor data packages specifically for mobloggers, thereby
driving higher usage and greater loyalty. TextAmerica sees as its mission the creation
of an alternative forum for information creation and exchange, and views much of the
content on its site as journalistic in nature.
Screen shot of TextAmerica moblog page
Source: TextAmerica
Social networking, common interest and invitation only services
We think one of the most interesting developments of the past year is the emergence of
a range of services and applications built around the concept of communication within
a defined “social” group. This takes a number of different forms, but the most
interesting to us were the following:
Social networking sites – the two best examples here are Friendster and Tribe, two
well-known sites in the US. Primarily used for dating, they are not exclusively
available to single people (other, more targeted sites, such as LinkedIn, accommodate
professional/industry networkers). Unlike conventional personals and dating sites,
which are anonymous, Friendster and Tribe make use of the user’s existing circle of
friends, as well as those met through the site itself, to establish contact with others. In
broad terms, the sites are practical applications of the “six degrees of separation”
concept, and are structured in such a way that direct contact between users can only be
established between the first two degrees of separation – friends and friends-of-friends.
Controlling your world A new user of Friendster, for example, will create a profile which can only be viewed
by those users of the service whom he has defined as friends, or their friends. He may
also invite other friends to join the service, thus expanding his potential range of
contacts. Friendster allows users to sort contacts by relationship, interests and marital
status. Users can also request introductions to friends-of-friends, post messages on
bulletin boards related to various topics, post testimonials to the profiles of their friends,
and publish personalized reviews of books, movies and other interests. Friendster is
currently free in beta version, though the site says that in future certain features will be
chargeable, though basic membership will remain free.
Example of Friendster user profile
Source: www.friendster.com
Wanna be in my tribe? Users of Tribe also create a personal profile, but access to this is not restricted to
friends or members of the friends’ extended networks. However, access to information
is limited on a hierarchical basis, depending on degrees of separation, and the user
profile displays whether or not there is any connection between two users. The Tribe
user’s friends will see the user’s full name, friends-of-friends will see only first name
and the first letter of last name, and people outside the network will see only the first
name. Unlike Friendster, Tribe users are able to contact all other users on the service,
though there is a strict anti-abuse policy in play. Users may partially screen themselves
from contact from outside their social circle by using a message prioritization system,
which distinguishes messages from those designated as friends and those outside this
group. Unlike Friendster, Tribe members may invite unknown persons to become their
friends via a contact option on the other user’s profile. Also different from Friendster is
the notion of “tribes,” or defined user groups connected by a common interest or theme.
They can either be open to all, or entirely private, as defined by the founder/moderator.
Users may also create listings (“roommate needed,” “car for sale in central San
Francisco”) which may also be configured so as to include or exclude any number of
users or user types.
Though quite different in a number of ways, both Friendster and Tribe share a couple
of common characteristics:
• They are designed to give the user a greater degree of control over his or her social
interactions and communication than a conventional personals/dating site.
• They are designed so as to be largely governed by principles of etiquette. As social
networks, which either leverage or create friendships and networking contacts,
abuse of the system (spam, stalking, antisocial behavior) is theoretically subject to
a backlash from the offender’s personal network, with the result being ostracism.
You can laugh, but this We think that part of the reason for the level of interest in Friendster and Tribe is an
stuff generates traffic and underlying sense of frustration with the internet generally, and a desire for some ability
absorbs time to control/limit information flows and filter communication with others. Ostensibly,
this should enable a more fruitful experience for the user, and perhaps as a validation
of the concept, we note with interest that Nielsen//NetRatings released data in
November which showed that Friendster users spend nearly three times as much time
in the site per month as users of the next most popular personals site, Match.com. This
is in spite of a user group (927,000 at the time) only one-quarter the size. Intuitively, it
could be expected that a larger group of users would generate more potential contacts,
and therefore, more usage. However, the fact that Friendster allows users the kind of
targeted access to people (not just people, but people who are known to the people the
user knows) not offered in a conventional personals/dating site, seems to produce
results.
Friendster vs. top 5 online personals destinations, October 2003

Unique audience Unique audience Monthly CAGR
Brand Jun-03 Oct-03 growth Time per person, Oct-03
Friendster 532 927 14.9% 1:51:54
Yahoo! Personals 4,110 4,923 4.6% 0:35:07
Match 4,327 3,916 -2.5% 0:55:40
American Singles 3,950 3,725 -1.5% 0:17:54
MSN Dating & Personals 3,381 1,946 -12.9% 0:01:46
Netscape Love & Personals 1,221 1,470 4.7% 0:04:29
Source: Nielsen//NetRatings data, November 2003
Everybody’s getting As an aside, we noted recently that music file-sharing colossus KaZaA is in partnership
social with Match.net, Germany-based owner of personals and dating sites in several
countries, to provide a personals service for KaZaA users. We think the fit between this
kind of service and a shared-interest community such as KaZaA is a natural one, and
we wonder how long it may be before the social networking sites begin bundling in
voice/video chat and other features to enhance the experience of their members. Even
within such seemingly well-defined segments of the internet world, it appears that we
will see an increasing blurring of definitions (file-sharing site becomes matchmaker,
matchmaker becomes communications tool complete with VoIP/IM, and so on).
Further evidence of this trend may be seen in the arrival of Orkut, a new social
networking site developed in affiliation with Google (the developer of the site is a
Google engineer - Google has a policy of urging engineers to work on personal
projects one day per week). While not formally part of Google’s service portfolio,
Orkut gives us some ideas about how the relationship between search and social
networking may be developed.
Glad I found you An early example of this concept comes in the form of Eurekster, a search engine
which employs the principle of social networking in processing and ranking search
results. By building up a personal profile of the registered user’s responses to the
search results it produces, Eurekster can not only present personally-tailored results to
a user re-visiting the site in future, but it can also present these as preferred choices for
friends of the user, if they happen to make the same search query. The idea is that the
social relationship between the connected users will lend a greater sense of relevance
and trustworthiness to the results it produces. Though the site does not directly identify
the individual who made a particular selection, another member of the social network
viewing the search results can contact the individual, and he in turn can also elect not
to be contacted. Users can also elect to conduct “private” searches, i.e., searches which
do not register in the social network’s list of preferences. Furthermore, users may
delete their search history in part or in full. As with Friendster and Tribe, it appears to
us that Eurekster is driven by a desire to maximize the benefits of a social network,
while retaining a high degree of control over one’s preferences and contacts within the
confines of the network. Again, we think this springs from a growing sense of
frustration with the state of the internet.
Consensus building on Another twist on the power of consensus is Public Mind, which (in our description) is a
the web web-based consumer advocacy/lobbying community which seeks to pressure
companies to produce/improve/revive products by amassing the support of persons
united by a common interest. A glance at the site reveals, for example, that one
category of interest is Skype, the P2P VoIP application we highlighted earlier in this
discussion. One subgroup, with 61 members so far, is lobbying for a webcam
capability within Skype. Another subgroup, with 71 members, is pushing for PSTN
interconnection on Skype. As with the Friendster and Tribe sites, the motivation here
seems to be to create a space where users’ efforts and energy have a reasonable focus
and sense of control.
Screen shot of Public Mind
Source: Public Mind
Welcome to my world Lastly, and most simply, we recently stumbled across, SafeMessaging.com, which
offers a free email service which is available by invitation only. The new user must
invite his desired contacts into the system, and the personal network is linked by a
public key infrastructure. Messages can only be exchanged between members of the
same network. Interestingly, the user interface contains a standard inbox and also a
“Winbox” which contains, for want of a better word, spam. Within one of the many
“promotional messages” in the Winbox may be a $300 cash prize payable if the user
takes the time to search for, and find, it. In this way, the email provider is able to sell
addresses to “marketing partners” without these marketing messages intruding upon
the world of the user. The user must willingly go into the segregated messages in order
to see them. For the user, everything is consensual and above board, and he has a
secure, private communication tool with a self-defined user group.
Screen shot from SafeMessaging invitation e-mail
Source: DIR, SafeMessaging
Darknets and small world networks
“The music industry's attempts to force digital data to behave like physical objects has
had two profound effects, neither of them about music. The first is the progressive
development of decentralized network models, loosely bundled together under the
rubric of peer-to-peer. Though there were several versions of such architectures as
early as the mid-90s such as ICQ and SETI@Home, it took Napster to ignite general
interest in this class of solutions. And the second effect, of course, is the long-predicted
and oft-delayed spread of encryption. The RIAA is succeeding where the Cypherpunks
failed, convincing users to trade a broad but penetrable privacy for unbreakable
anonymity under their personal control. In contrast to the Cypherpunks "eat your
peas" approach, touting encryption as a first-order service users should work to
embrace, encryption is now becoming a background feature of collaborative
workspaces. Because encryption is becoming something that must run in the
background, there is now an incentive to make its adoption as easy and transparent to
the user as possible. It's too early to say how widely casual encryption use will spread,
but it isn't too early to see that the shift is both profound and irreversible.”
---Clayton Shirky, 2003
A walk on the wild side A more-or-less natural extension of the social networking/invitation only network
development, in our view, is the phenomenon of the “darknet,” which, under some
definitions would include such file sharing applications as KaZaA and eDonkey, where
the network is formed dynamically between the active nodes on it, in a decentralized,
peer-to-peer fashion. Indeed, we think that under this broad definition, Skype already
qualifies as a large voice darknet, so we could argue that darknets are already in the
mainstream of the telecom world. Beyond what Skype is doing in the voice space, we
think the other P2P networks contain interesting possibilities as alternative
communications platforms beyond content sharing, especially in light of their huge
installed user bases (KaZaA has been downloaded an estimated 300m times). As we
saw earlier in our discussion of social networking groups, KaZaA has entered the
dating service scene, and we expect further diversifications of these services to evolve.
Freenet and free speech There are other examples of darknets in existence, some of which share and store
information in interesting ways. Freenet is essentially a grid computing project, which
has the aim of allowing users to publish and obtain information on the internet in an
anonymous capacity. Users agree to allow a portion of their bandwidth and disk space
to be used in the storage and transmission of information across the network, but no
user is allowed to control what his machine is storing or transmitting. Storage of
content is dynamic and dictated by popularity of such content, thus less popular content
is eventually replaced by more popular content in response to user preferences. The
commitment of the Freenet founders to absolute freedom of speech means that users of
the service cannot guarantee that they will not be unconsciously aiding in the
distribution of offensive content or the planning of terrorist acts, yet the application has
been downloaded over 2m times and is apparently in active use in countries governed
by some of the world’s less libertarian regimes. (The commitment to promoting free
speech is nonetheless a noble cause [one which Skype recently tapped into as a “brand
message”], and we wonder how many broking research departments may be
considering the darknet concept for research distribution, in the wake of certain recent
court actions resulting in damage judgments for alleged bias in research!;-)
Overnet and rewarding Overnet is a relatively new P2P file sharing application developed by MetaMachine,
netiquette the developers of the eDonkey2000 system. It is fully distributed and makes use of
multifile source transmission protocol (MFTP), which allows a user seeking to
download a file to do so simultaneously from multiple sources, by dividing the target
file into segments. This is to minimize the drain on upstream resources of uploaders
and facilitate faster downloads. Presumably this also makes apportionment of
culpability more difficult for those seeking to enforce intellectual property rights.
Overnet also makes use of a “horde” function, which seeks to match downloaders of
the same file with the four fastest nodes on the network currently trying to download
the same file. This creates partner relationships on the network, and allows those
contributing the fastest uploads to also benefit from the fastest downloads.
Unh@ppy with the size of So far all the examples we have looked at involve self-forming and largely self-
y0ur c0mmunications controlling networks, where the individual user’s level of control over what
t0ol? Cho0se d@rknet! information passes through his node is either limited (Overnet) or non-existent
(Freenet). In the case of Overnet, at least, good network etiquette is rewarded with
more generous bandwidth. However, we are particularly interested in the darknet
principal as applied in a highly personal context, in other words, defined as a
distributed peer-to-peer network which is created by, and known only to, its members,
and is effectively invisible to the outside world. The examples we have been able to
discover to date seem to be the product variously of a combination of frustration with
spam (and instant messaging spam, or “spim,” which is estimated to have doubled last
year, to 500m incidents), concern over privacy/freedom of information, and fear of
prosecution in the wake of the RIAA lawsuits.
US online adults level of annoyance with spam

2000 2002 May-03 Nov-03
Somewhat/very annoying 69% 96% 93% 90%
Very annoying 49% 80% 64% 65%
Somewhat annoying 20% 16% 29% 25%
Not annoying at all/not very annoying 27% 4% 7% 10%
Not very annoying 8% 3% 5% 7%
Not annoying at all 19% 1% 2% 2%
Source: Harris Interactive, December 2003
We think this phenomenon may become increasingly common in the years ahead,
aided by the apparent growing desire among internet users (reflected in the success of
Friendster) to define and control communication within one’s personal network. No
doubt darknets will also be of increasing interest to law enforcement and intellectual
property guardians, as they can just as easily serve to facilitate criminal activity and
illegal file sharing as to provide secure communication, protect freedom of expression
from oppressive regimes or foster a new age of content creation and sharing between
users. We remain non-judgmental, and believe the issue deserves more attention as a
driver of incumbent telco revenue erosion. Below we discuss two relatively obscure
applications which we believe epitomize the direction personal communications may
take in future.
We need to have a heart- WINW – This is an encrypted P2P networking application, developed as a sort of
to-heart P2P successor to the WASTE project (which was withdrawn immediately upon its
accidental release by AOL subsidiary Nullsoft, headed by the co-founder of the
Gnutella file-sharing network, in mid-2003), though with different aims. Developer
Mike Tsao told us that the change in direction of the WINW project arose partly from
concerns over the growing aggression of the RIAA, but was also partly inspired by the
success of social networking sites such as Friendster and Tribe. In its present version,
WINW supports 256-bit AES encryption with RSA session key exchange. Currently
only IM functions are active over a direct P2P connection, but future versions should
include routing and meshing capabilities, file transfer/browsing/sharing, and chat
rooms. Our guess is that it is not much of a stretch to assume that voice can be added to
this list at some point. WINW’s “30-second directions” give an idea of the likely
appeal of such do-it-yourself applications to users:
WINW “30-second directions”
How to use
30-second directions:
Install. Run. Generate key.
E-mail your public key to a
friend, along with your IP
address. Make sure port
8888 is open through your
firewall. When you get your
friend's public key, go to
"Add Friend..." and paste
his key in. Also enter his IP
address in the hostname
field. Make sure your friend
does the same for you. You
should now be able to chat
with each other.
Source: WINW.org
BadBlue – Another emerging darknet application we have come into contact with is
badblue, developed by Atlanta-based Working Resources. BadBlue’s mission
statement is pretty clear as to its intent:
“The mission of BadBlue is simply this: To establish a democratic platform for web
serving publishing, and sharing; to enable anyone to easily connect with others; and to
provide for bi-directional information flow for dramatically easier collaboration.”
Working Resources has seen 750,000 downloads of BadBlue since launch three years
ago, with 250,000 conversions to registered users. BadBlue is offered in both a
Personal and Enterprise version. The Personal edition is free, though advanced features
expire after a trial period and must be resubscribed for a license fee, starting at $29.99
per user. BadBlue enables a variety of web-publishing functions, as well as P2P file
searching and sharing within an entirely user-defined group of participants in a highly
encrypted environment. In practice, the individual user would download the software,
encourage friends to do the same, and all the members of the “network” would then
configure their machines to allow P2P interaction with their contacts on the basis of IP
address. Non-registered persons would not have access. As with WINW, messaging
and VoIP are not currently functions of BadBlue, but we expect in this case also that
adding these features is a natural extension of the kind of highly personalized
networking which the application enables.
If the shoe fits… We think the key appeal of applications such as these is in allowing the user to define
and control his communications network, either for the completion of a specific task, or
to create closed communities which share specific interest. The World Internet Project
report asked internet users to assess the impact of the Net on communication with
friends/family, people of the same religious group, people with similar political views,
and people sharing the same profession. The response was strongest in relation to
communications with people of the same profession, suggesting that many users
already view the web in this way – as a place to share valuable information toward a
purpose. How much better to do it in a secure, controlled environment, where
collaboration tools are included, and global voice may be free.
Internet increased/decreased contact with people who share their profession
90%
80%
70%
60%
50% Increased or great ly increased
40% Same
30%
20%
10%
0% en
SA
n
e
ry
ea
y
n)
ly
or
ai
an
ed
ga
Ita
or
a
U
Sp
ap
rb
Sw
K
un
ng
(u
er
H
G
Si
na
i
Ch
• VoIP is increasingly well established in the enterprise, and research suggests
high prospective levels of uptake in both large (c.80%) and smaller businesses
(c.50%) over the next three to five years.
• In the consumer segment, commercial VoIP offerings to date have been limited
mainly to access-tied products, but feature highly aggressive/disruptive pricing
models.
• Access-independent VoIP products are on the way and offer increased
opportunities for price and location arbitrage. We think BT’s launch of a
Vonage-style product gives a very clear view of how the more forward-looking
PTTs may respond, but also highlights the difficult cannibalization issues such
responses may trigger.
• We think the market underestimates the potential of closed VoIP platforms in
their many manifestations. If our survey is correct in showing that more than
90% of phone calls involve the user’s direct social and family network, then
failure to interconnect with the PSTN may not be a significant stumbling block to
widespread adoption. We think consumers will be increasingly drawn to use such
services opportunistically as broadband proliferates, reserving the “rump” of
calling habits for legacy networks such as mobile/PSTN.
• We believe there is a mounting case for telco erosion based on the rise of
relatively new behaviors such as blogging, social networking communities,
invitation only communications platforms, and self-created/self-defined darknets.
7. A multi-factor approach to stock rating
In this section we introduce an approach to rating stocks which seeks to encompass
some of the factors discussed earlier in this report, in an attempt to balance valuation
issues with other less obvious drivers of share price performance.
Do the forecasts If we have been at all successful in the presentation of our ideas in this report, then we
encompass the have at least established that we believe incumbent telcos operate in a very uncertain
uncertainty environment, heavily influenced by changes in consumer psychology, technology shifts,
demographic trends, political and economic development policy at the national,
regional and local levels, and regulation. Accordingly, we think visibility is very low,
and investment in the sector, even in well-established incumbent players, is a highly
speculative decision. As we have sought to demonstrate, the rate and scope of change
occurring in the sector is impossible to gauge accurately, and we suspect that most
market and broker forecasts do not fully encompass all the uncertainty.
“Cheap” ain’t good With so much uncertainty underlying future assumptions about growth and cash
enough generation, for the long-term fundamental investor, we believe it is no longer sufficient
to see a stock trading 15% below the DCF-derived fair value, decide it is “cheap,” and
buy it. The reality of the sector is more complex than that, and growing more so each
day. In such an environment, we believe investors probably use a system similar to our
own, which we present here for the first time. Somewhat bombastically dubbed the
“Daiwa multifactor rating system” it is an attempt by us to incorporate many of the
themes we have discussed here, as well as some other key technical and valuation
parameters, to come up with a weighed score, on which a rating may be based. Below
is an example of the tear sheet we have created for each company, and our explanation
of each factor follows it.
Daiwa multifactor rating system explanation
Source: DIR
Our rating system is comprised of seven weighted factors. Each is scored on a one-to-
five scale, consistent with Daiwa’s stock rating system. These factors are:
• Fundamental upside – Based on a 10-year DCF valuation, and applying the

criteria used in Daiwa’s stock rating system. 15% weighting in overall score.
• Relative value – Based on the stock’s forward Price/cash EPS multiple versus the
coverage universe. 15% weighting.
• Liquidity – Based on the stock’s share of sector liquidity on a rolling 90-day basis.
The rationale here is that, no matter how good a story, for many investors stocks
trading below a certain level of liquidity are uninvestable. 10% weighting.
• Market environment – This is the broadest and most subjective of the factors we
look at, but also potentially the most critical, which accounts for its 30%
weighting. Here we attempt to assess the regulatory and competitive environments
in key markets where the company operates, as well as the susceptibility of those
markets to adverse shifts in technology and demographics, as described earlier in
this note.
• Corporate governance – At the end of last year, as part of the development of our
multi-factor system, we conducted some research into the corporate governance
practices of our coverage universe: Eurotelcorama Issue 4: Corporate
Governance Review of European Incumbents (December 4th). We developed a
scorecard in conjunction with two policy papers which were issued during the
course of 2003: The UK Combined Code of Corporate Governance and
Modernising Company Law and Enhancing Corporate Governance in the
European Union – A Plan to Move Forward. Once the scorecard was developed,
we graded our companies in terms of how well they performed in five main
categories including: reporting, the composition of the board of directors, the
remuneration policies of the board, accountability and how accessible the
information was to shareholders. Overall we thought that our companies
performed well on the scorecard with no company receiving less than 60% of the
available points and with BT reaching 82.9% of the available points. We would
have expected BT to rank at the high end of the scale due to the fact that it has
been a publicly quoted company since 1984 (long before any of the other
companies) and because the UK is one of the main centers for shareholder
activism with respect to corporate governance. Apart from that, we can’t draw the
usual North/South divide, which is usually prevalent in evaluating this group of
companies because TI ranked second in terms of percentage of points. Overall,
the reporting, remuneration and shareholders’ rights sections scored highly. Many
of these companies have listings in the US and are required to provide the detail
that we were evaluating in these sections. The directors’ section was the most
disappointing with most of the non-executive board members not able to meet our
definition of independence, although we admit that our definition of independence
far exceeded the definitions prevalent in most codes. We have included the
following table, which tabulates the scores for the various companies but for a
fuller discussion, please contact one of the authors of this report.
Head to Head Comparison – Breakdown of scoring

BT DT FT KPN PT SCMN TEF TI TLSN
Reporting (10) 7 7 8 7 8 8 8 10 8
Directors (8) 7 3 2 5 3 4 3 3 5
Remuneration (10) 9 7 8 8 6 7 7 8 7
Accountability and Audit (4) 3 1 2 2 2 2 2 3 2
Strengthening Shareholders’ Rights 3 3 3 2 2 3 3 3 3
(3)
Total (35) 29 21 23 24 21 24 23 27 25
% of total 82.9% 60.0% 65.7% 68.6% 60.0% 68.6% 65.7% 77.1% 71.4%
Source: DIR
Based on the above table we have rated the companies as follows:
Head to Head Comparison – Summary of points

BT DT FT KPN PT SCMN TEF TI TLSN
Multi-factor rating 1 3 2 2 3 2 2 2 2
Source: DIR
• Adaptability – Again highly subjective, but it is our assessment of the scope for
the company to respond to adversity. 10% weighting.
• Yield – It is fair to stipulate that, if investing in a sector characterized by a high
level of uncertainty, the compensation should be a nice dividend payment. 10%
weighting.
This is roughly how it works in practice, using the values appearing in the generic
example above:
• Fundamental upside – Our DCF for Company A shows 10% upside, while our
strategist thinks the market is fairly valued. On Daiwa’s rating criteria, the stock is
an outperform. Rating 2.
• Relative value – Company A is trading in line with the peer group. On Daiwa’s
rating criteria, the stock is a neutral. Rating 3.
• Liquidity – Company A’s share of sector liquidity on a rolling 90-day basis is
only 3%. This is below the threshold we consider to be attractive, and the stock
receives a 4 rating
• Market environment – Company A operates in a country with a highly
interventionist regulator (regulation rating 5), a number of facilities based
competitors in the fixed business, and three other mobile operators with roughly
equal share (competition rating 4). The home market is highly penetrated with
broadband cable and competitors in the ADSL market have begun offering
aggressively priced DSL access and VoIP packages (new technology threat rating
5). However, a substantial and increasing proportion of the operator’s cash flow
comes from its very shrewd investments in emerging market cellular businesses,
thus offering a shield from otherwise difficult market conditions (demographics
rating 2). Overall rating 4.
• Corporate governance – Our survey of corporate governance standards revealed
that Company A has substandard and inadequate financial reporting and risk
management procedures, an insufficient number of independent non-executive
directors, and unusual remuneration structures for the directors and chairman.
Rating 5.
• Adaptability – Company A’s management have demonstrated themselves to be
forward-looking in creating innovative new products and marketing strategies,
and eager to employ new technologies in the defense of the core cash flow
generating units. However, labor laws in its home market are very restrictive, and
40% of employees are classified as civil servants. Rating 3.
• Yield – The company reduced its dividend by 50% one year ago as part of a debt
reduction drive, and it yields only 2% at the current share price. Rating 5.
Outcome and ratings Through this process, a stock which started out as a 2 [outperform] rating ends up as a
actions 3.65, which translates to 4 [underperform]. By going through the process and assigning
ratings to each element of the story, we were able to factor in some uncertainties which
are not apparent in the forecasts. This is a process which investors and analysts have
probably engaged in anyway, whether consciously or intuitively. We have merely
sought to isolate and quantify some of the factors which we think should go into the
decision making process. We have performed this exercise on all the companies in our
PTT coverage universe, with the following outcomes:

Summary of ratings actions
Price DCF Target Upside Relative Multifactor rating Stock rating action
BT £1.82 1.70 -6.56% -4.99% 2.7 NEUTRAL [3] MAINTAINED
DTE €16.28 16.24 -0.26% 1.32% 3.1 NEUTRAL [3] MAINTAINED
FT €24.88 22.95 -7.77% -6.20% 3.5 NEUTRAL [3] → UNDERPERFORM [4]
KPN €6.42 6.81 6.06% 7.63% 3.5 NEUTRAL [3] → UNDERPERFORM [4]
PT €8.74 10.86 24.22% 25.79% 2.7 OUTPERFORM [2] → NEUTRAL [3]
SCMN CHF423.5 374 -11.58% -10.01% 3.8 NEUTRAL [3] → UNDERPERFORM [4]
TEF €13.09 13.93 6.44% 8.01% 2.4 NEUTRAL [3] → OUTPERFORM [2]
TIT €2.553 2.81 10.10% 11.67% 2.8 UNDERPERFORM [4] → NEUTRAL [3]
TITn €1.758 1.89 7.58% 9.15% 2.8 NEUTRAL [3] MAINTAINED
TLSN SEK37.50 37.68 0.48% 2.05% 3.3 NEUTRAL [3] MAINTAINED
Benchmark Central case Bear case Bull case
DIR STOXX target 235 210 260
Upside from current -1.57% -12.04% 8.90%
Source: DIR
Let’s build our own As for the specifics of our rationale in applying these ratings, we intend to follow the
darknet “darknet” principle described earlier in this report, and take the discussion offline, in
peer-to-peer fashion. For Daiwa clients and prospective clients, please contact us or
your Daiwa sales representatives for a one-on-one presentation or conference call.
DIR multifactor rating summary on BT Group (£m)
C ompany data Financial summary 2003 2004E 2005E 2006E
C ompany name B T Group Divisional turnover
Daiwa rating 3 W holesale 11,247 11,256 11,348 11,384
B loomberg ticker BT/A LN Retail 13,882 13,543 13,406 13,022
R euters ticker B T.L Global Solutions 5,417 5,829 6,467 7,284
P rice (GBP ) 1.82
M arket cap 15,698
EV 25,103
Technicals Group turnover 18,727 18,775 19,069 19,355

52 week high 210.25 Divisional EB ITDA
52 week low 140.50 W holesale 3,681 3,767 3,868 3,944
Average daily turnover $US, 100D 211.12 Retail 1,729 1,652 1,485 1,432
Share of sector liguidity, $US, 30D 6.20% Global Solutions 238 436 513 579
M oving average, 30D 183.35
5-day deviation from 30D M AV -1.15%
2 year beta vs. STOXX 600 0.92
STOX X Telecom index weighting 5.69%
Local index weighting 1.46% Group EB IT DA 5,805 5,957 5,965 6,066
Adjusted E PS 14.18 17.62 18.25 21.10
Cash EPS 51.30 58.43 56.99 57.42
M ultifactor rating analysis R ating DPS 6.50 8.20 11.50 13.13
Fundamental upside 4 Capex 2,444 2,528 2,618 2,757
R elative value 1 OFCF 1,976 2,496 2,158 1,994
Liquidity 3 Ratios
M arket environment 5 EV /EBITDA (adjusted) 4.71 4.33 4.21 4.07
C orporate governance 1 P /OFCF 7.94 6.29 7.27 7.87
Adaptability 1 EB ITDA/OFC F conversion rate 34.0% 41.9% 36.2% 32.9%
Yield relative to sector 1 P /E (adjusted earnings) 12.83 10.33 9.97 8.63
Yield 3.6% 4.5% 6.3% 7.2%
EB ITDA/capex 2.38 2.36 2.28 2.20
Net debt/EB ITDA 1.65 1.34 1.22 1.14
M ultifactor rating 2.7 EB ITDA/net interest 5.07 7.24 8.91 10.04
Source: DIR
DIR multifactor rating summary on Deutsche Telekom (€m)

C ompany name Deutsche Telekom Divisional turnover
Daiwa rating 3 T-C om 26,491 25,423 24,940 24,083
B loomberg ticker DTE GR T-M obile 18,339 21,480 23,985 26,093
R euters ticker DTE Gn.F T-Systems 6,894 6,846 7,034 6,575
P rice (E UR ) 16.28 T-Online 1,391 1,644 2,465 2,908
M arket cap 68,295 Headquarters and shared services 573 325 400 445
EV 117,884

53 week low 9.04 T-C om 10,092 10,189 9,795 9,994
Average daily turnover $US, 100D 356.83 T-M obile 5,047 6,617 7,647 8,741
Share of sector liguidity, $US, 30D 12.81% T-Systems 1,151 1,316 1,277 1,221
M oving average, 30D 15.173 T-Online 76 325 503 411
5-day deviation from 30D M AV 6.37% Headquarters and shared services 30 -35 -20 40
Adjusted E PS 1.84 0.81 1.21 1.70
Cash EPS 2.96 3.60 3.71 4.15
Liquidity 2 Ratios
Yield 0.00% 0.00% 1.43% 2.39%
EB ITDA/capex 2.39 3.08 2.65 3.08
Net debt/EB ITDA 3.76 2.64 2.10 1.49
Source: DIR

DIR multifactor rating summary on France Telecom (€m)
C ompany name France Telecom Divisional turnover
Daiwa rating 4 Fixed telephony and data France 18,703 17,997 17,813 17,264
B loomberg ticker FTE FP Orange 16,660 17,486 18,699 19,765
R euters ticker FTE.P A W anadoo 1,935 2,505 3,085 3,658
P rice (E UR ) 24.88 Fixed telephony and data outside France 9,332 8,764 8,732 9,383
M arket cap (EURm) 59,587
EV (EURm) 94,580

52 week high 25 Divisional EB ITDA
53 week low 14.81 Fixed telephony and data France 7,387 8,008 7,829 7,596
Average daily turnover $US, 100D 214.61 Orange 5,146 6,380 7,041 7,503
Share of sector liguidity, $US, 30D 5.60% W anadoo 90 356 468 704
M oving average, 30D 23.173 Fixed telephony and data outside France 2,312 2,606 2,781 2,721
5-day deviation from 30D M AV 2.23%
Adjusted E PS 0.04 1.71 1.99 2.37
Cash EPS 13.04 5.41 5.99 6.33
Liquidity 3 Ratios
Yield relative to sector 3 P /E (adjusted earnings) nm 14.54 12.51 10.50
Yield 0.0% 0.0% 3.6% 4.4%
EB ITDA/capex 1.95 3.11 3.01 2.98
Net debt/EB ITDA 4.58 2.68 2.13 1.82
Source: DIR
DIR multifactor rating summary on KPN (€m)

C ompany name KP N Divisional turnover
Daiwa rating 4 Fixed 8,679 8,638 8,402 8,563
B loomberg ticker KPN NA M obile 4,737 5,131 5,367 5,271
R euters ticker KPN.AS Other 1,225 996 1,016 1,036
P rice (E UR ) 6.42
M arket cap 15,992
EV 23,503

53 week low 5.30 Fixed 2,724 3,173 3,095 3,116
Average daily turnover $US, 100D 107.50 M obile 1,613 1,720 1,644 1,644
Share of sector liguidity, $US, 30D 2.80% Other 42 121 40 40
M oving average, 30D 6.26
Adjusted E PS 0.09 0.39 0.58 0.70
Cash EPS 1.28 1.45 1.66 1.58
Liquidity 4 Ratios
Yield nm 1.87% 3.46% 4.42%
EB ITDA/capex 3.85 3.95 2.75 2.62
Net debt/EB ITDA 2.82 1.83 1.48 1.16
Source: DIR

DIR multifactor rating summary on Portugal Telecom (€m)
C ompany name P ortugal Telecom Divisional turnover
Daiwa rating 3 P TC 2,301 2,291 2,239 2,220
B loomberg ticker P TC PL P T Prime 335 0 0 0
R euters ticker PTCO.IN TM N 1,475 1,543 1,620 1,683
P rice (E UR ) 8.52 Telesp C elular/V ivo 1,218 1,400 1,608 1,789
M arket cap 10,687 P TM 676 665 706 742
EV incl pension and minorities (2003) 1,189 Other -424 -95 -101 -100
53 week low 5.55 P TC 957 915 937 931
Average daily turnover $US, 90D 35.74 P T Prime 32 0 0 0
Share of sector liguidity, 30D 0.98% TM N 623 691 731 765
M oving average, 30D 8.22 Telesp C elular/V ivo 512 531 629 703
5-day deviation from 30D M AV 4.25% P TM 76 128 165 206
2 year beta vs. STOXX 600 0.81 Other -24 -260 -121 33
STOX X Telecom index weighting 2.44% Group EB IT DA (Adjusted) 2,176 2,005 2,340 2,639
Local index weighting 21.63%
Adjusted E PS 0.38 0.48 0.66 0.85
Cash EPS 0.41 1.11 1.85 2.04
Fundamental upside 1 Capex 1,124 679 695 728
R elative value 3 OFCF 566 687 1,391 1,577
Liquidity 5 Ratios
Yield 1.5% 2.0% 2.6% 3.3%
EB ITDA/capex 1.9 3.0 3.4 3.6
Net debt/EB ITDA (includes pension liability 2.6 2.5 2.1 1.5
Source: DIR
DIR multifactor rating summary on Swisscom (CHFm)

C ompany name Swisscom Divisional turnover
Daiwa rating 4 Fixnet 6,258 5,799 5,806 5,759
B loomberg ticker SC M N V X M obile 4,115 4,204 4,324 4,360
R euters ticker SCZM n.VX Enterprise Solutions 1,523 1,372 1,383 1,383
P rice (C HF) 423.5 debitel 4,112 4,543 4,602 4,662
M arket cap 28,037 Other 1,464 1,311 1,277 1,291
EV 27,274 Corporate 703 661 660 660

52 week low 367.00 Fixnet 1,870 2,056 1,989 1,850
Average daily turnover $USm, 100D 31.18 M obile 1,978 2,030 1,946 2,049
Share of sector liguidity, $US, 30D 1.00% Enterprise Solutions 97 104 114 231
M oving average, 30D 410.30 debitel 158 106 115 233
5-day deviation from 30D M AV 0.17% Other 110 145 121 110
2 year beta vs. STOXX 600 0.55 Corporate 200 96 80 0
Adjusted E PS 28.93 30.53 33.31 34.01
Cash EPS 52.28 68.96 61.90 64.76
Liquidity 5 Ratios
Yield 2.8% 3.5% 4.0% 4.5%
EB ITDA/capex 3.61 3.82 3.13 3.08
Net debt/EB ITDA nm nm nm nm
M ultifactor rating 3.8 EB ITDA/net interest 14.21 nm nm nm
Source: DIR

DIR multifactor rating summary on Telecom Italia (€m)
Company data Financial summary 2002 2003E 2004E 2005E
Company name Telecom Italia Divisional turnover
Daiwa rating 3 Domestic wireline 17,035 17,016 17,353 17,576
Bloomberg ticker TIT/TITR IM M obile services 10,867 11,899 12,430 12,940
Reuters ticker TLIT/TLITn.M I Internet and media 1,991 1,262 572 634
P rice ordinary (EUR) 2.56 South America 1,409 1,147 1,099 1,128
P rice savers (EUR) 1.76 IT market 994 684 690 704
M arket cap 36,455 IT Group 996 941 976 1,000
EV incl pension and minorities (2003) 84,423 Other -1,885 -2,384 -2,335 -2,318
Technicals (0rdinaries) Group turnover 31,407 30,564 30,785 31,664
52 week high 2.63 Divisional EBITDA
53 week low 1.78 Domestic wireline 7,966 8,101 8,362 8,516
Average daily turnover $US, 90D 220.6 M obile services 5,039 5,531 5,764 6,074
Share of sector liguidity, 30D 5.64% Internet and media 593 309 30 44
M oving average, 30D 2.49 South America 450 387 344 337
5-day deviation from 30D M AV 1.93% IT market 111 64 95 97
2 year beta vs. STOXX 600 0.87 IT Group 98 74 139 143
STOX X Telecom index weighting 5.73% Other -1,000 -921 -1,197 -1,218
Local index weighting 6.59% Group EBITDA (Adjusted) 13,257 13,545 13,537 13,992
Adjusted E PS 0.2 0.2 0.2 0.2
Cash EPS 1.4 0.6 0.6 0.6
M ultifactor rating analysis Rating DPS (ords) 0.1 0.1 0.1 0.1
Relative value 2 OFCF 7,819 4,956 4,828 4,686
Liquidity 3 Ratios
Corporate governance 2 P /OFCF 5.8 6.8 7.6 7.8
Adaptability 3 EBITDA/OFCF conversion rate 59.0% 36.6% 35.7% 33.5%
Yield 1.4% 4.0% 3.8% 3.9%
EBITDA/capex 2.7 2.9 2.7 2.7
Net debt/EBITDA (includes pension liability 1.4 2.5 2.2 2.1
M ultifactor rating 2.8 EBITDA/net interest 8.2 5.3 7.2 5.6
Source: DIR
DIR multifactor rating summary on Telefonica (€m)

C ompany name Telefonica Divisional turnover
Daiwa rating 2 Telefonica de Espana 10,272 10,177 10,333 10,540
B loomberg ticker TEF SM International fixed line 6,954 6,364 6,802 7,365
R euters ticker TEF.M C C ellular business 9,449 10,270 10,971 11,949
P rice (E UR ) 13.09 Data 1,731 1,760 1,967 2,157
M arket cap 64,873 M edia 1,076 1,314 964 991
EV (2003) 86,487 Terra Lycos 600 530 550 572
Directories business 551 569 581 601
Atento 571 475 478 504
Other -2,793 -3,366 -3,594 -3,594
53 week low 7.75 Telefonica de Espana 4,517 3,161 3,836 4,074
Average daily turnover $US, 90D 515.07 International fixed line 3,347 3,070 3,317 3,653
Share of sector liguidity, 30D 19.11% C ellular business 3,830 4,641 5,026 5,511
M oving average, 30D 12.224 Data 170 270 346 397
5-day deviation from 30D M AV 6.18% M edia 115 204 180 184
2 year beta vs. STOXX 600 1.00 Terra Lycos -142 -56 10 57
STOX X Telecom index weighting 15.15% Directories business 151 180 187 197
Local index weighting 22.39% Atento 54 59 88 111
Other -318 -300 -284 -284
Group EB IT DA 11,724 11,231 12,706 13,902
Adjusted E PS 2.28 0.44 0.62 0.79
Cash EPS 2.96 1.94 2.50 2.38
Liquidity 1 Ratios
Yield 3.1% 3.4% 3.1% 3.6%
EB ITDA/capex 3.1 3.1 3.1 3.4
Net debt/EB ITDA 2.1 1.9 1.5 1.2
Source: DIR
DIR multifactor rating summary on TeliaSonera (SEKm)
C ompany name TeliaSonera Divisional turnover
Daiwa rating 3 M obile 32,294 35,894 39,582 41,649
B loomberg ticker TLSN SS Fixed 41,643 40,286 39,495 39,124
R euters ticker TLSN.ST International C arrier 5,577 4,154 4,349 4,576
P rice (SEK) 37.50 Holding 1,952 1,483 1,468 1,512
M arket cap 175,321
EV 183,615

53 week low 23.70 M obile 14,623 16,633 17,511 18,309
Average daily turnover $USm, 100D 61.36 Fixed 12,439 14,888 14,322 13,971
Share of sector liguidity, $US, 30D 2.06% International C arrier -1,225 94 217 229
M oving average, 30D 37.76 Holding 243 198 147 493
Adjusted E PS -1.92 2.31 2.76 3.10
Cash EPS 5.81 5.10 5.36 5.60
Liquidity 4 Ratios
Yield 1.07% 1.60% 2.23% 2.92%
EB ITDA/capex 2.17 4.01 3.56 3.30
Net debt/EB ITDA 1.50 0.64 0.22 -0.14
M ultifactor rating 3.3 EB ITDA/net interest 21.36 nm nm nm
Source: DIR

Companies mentioned in this report:
4G Networks (unlisted)
AOL (TWX US)
Apple Computer (AAPL US)
Arbitron (ARB US)
AT&T (T US)
B2 Bredbandsbolaget (unlisted)
BT Group (BT/A LN)
Cablecom (unlisted)
Casema (unlisted)
Clear Channel Communications (CCU US)
Deutsche Telekom (DTE GR)
D-Link (2332 TT)
Essent Kabelcom (unlisted)
France Telecom (FTE FP)
Free.fr/Iliad (ILD FP)
Free World Dialup (unlisted)
Friendster (unlisted)
Google (unlisted)
IAXtel (unlisted)
IPtel (unlisted)
KaZaA (unlisted)
KPN (KPN NA)
LinkedIn (unlisted)
Matchnet (MHJ GR)
MetaMachine (unlisted)
Microsoft (MSFT US)
Nielsen//NetRatings (NTRT US)
NTL (NTLI US)
PacketFront (unlisted)
Portugal Telecom (PTC PL)
Public Mind (unlisted)
Qwest Communications (Q US)
SafeMessaging.com (unlisted)
SIPPhone (unlisted)
Skype (unlisted)
Swisscom (SCMN VX)
TDC (TDC DC)
Tele2 (TEL2B SS)
Telecom Italia (TIT/TITR IM)
Telefonica (TEF SM)
Telewest (TWT LN)
TeliaSonera (TLSN SS)
TextAmerica (unlisted)
Theglobe.com (voiceglo) (TGLO US)
Tribe (unlisted)
UPC Austria (unlisted)
Voicepulse (unlisted)
Vonage (unlisted)
Working Resources (unlisted)
Xten (unlisted)
Yahoo! (YHOO US)
DIR Eurotelcorama 105

Sites/resources of interest:
www.fibertothehome.info
www.breedbandproeven.nl
www.utopia.org
www.cenic.org
www.bryggenet.dk
www.djurslands.net
www.kenniswijk.nl
www.packetfront.com
www.badblue.com
www.winw.org
www.textamerica.com/weblogs
www.databank.it/STAR
www.eurescom.de/e-living
www.europa.eu.int/information_society
www.pewinternet.org
www.free2air.org
www.consume.net
www.freenetworks.org
http://tv.seattlewireless.net
www.personaltelco.net
www.blogstreet.com

Analyst coverage:
James Enck
BT Group (BT/A LN)
Deutsche Telekom (DTE GR)
France Telecom (FTE FP)
KPN (KPN NA)
Swisscom (SCMN VX)
TeliaSonera (TLSN SS)
Jacqueline Millan
Portugal Telecom (PTC PL)
Telecom Italia (TI IM, TIR IM)
Telefonica (TEF SM)
Telefonica Moviles (TEM SM)
Vodafone (VOD LN)

DIR Global Telecoms Contacts
Global
Kaoru Uchida Global Coordinator (813) 5683 5989 k.uchida@rc.dir.co.jp
James Enck Global Sector Strategy (44 20) 7597 8455 james.enck@dir.co.uk
Europe
James Enck Communications, all flavors (44 20) 7597 8455 james.enck@dir.co.uk
Jacqueline Millan Communications, all flavors (44 20) 7597 8460 jacqueline.millan@dir.co.uk
USA
Kenji Nishimura Integrated/Cellular/Equipment Manufacturers (1 212) 612 6108 nkenji@dira.com
Japan
Shinji Moriyuki Integrated/Cellular (813) 5683 6958 s.moriyuki@dir.co.jp
Jun Hasebe Alternative Operators (813) 5683 5109 j.hasebe@dir.co.jp
Hiroyuki Masuko Equipment Manufacturers (813) 5683 6957 h.masuko@dir.co.jp
Asia-Pacific (ex-Japan)
Andrew Jobson Regional/Korea/Taiwan (852) 2848 4463 andrew.jobson@dir.com.hk
Jenny Szeto, CFA Hong Kong/China (852) 2848 4465 jenny.szeto@dir.com.hk
Jerry Yeu Australia (613) 9916 1354 jerry.yeu@daiwasmbc.com.au
Brenda Lee, CFA Singapore (65) 6321 3039 brenda@dir.com.sg
Vijay Chander Asian Credit Strategy (852) 2848 4411 chander@daiwasmbc.com.hk

Disclaimer
This material was produced by Daiwa Institute of Research Europe Limited for and on behalf of Daiwa Securities SMBC Europe Limited. Daiwa Institute of Research is the trading name of Daiwa Institute of Research Europe
Limited. Daiwa Institute of Research Europe Limited is an appointed representative of Daiwa Securities SMBC Europe Limited which is regulated by The Financial Services Authority and is a member of the London Stock
Exchange. This material is distributed in the United States by Daiwa Securities America, in Japan by Daiwa Securities SMBC Co. Ltd, Tokyo, and elsewhere in the world by Daiwa Securities SMBC Europe Limited, or another
authorized affiliate of Daiwa Securities SMBC Europe Limited (such entities and their associated persons, the “Companies”). The Companies may, from time to time, to the extent permitted by law, participate or invest in other
financing transactions with the issuers of the securities referred to herein (the “Securities”), perform services for or solicit business from such issuers, and/or have a position or effect transactions in the Securities or options
therefore and/or may have acted as an underwriter during the past three years for the issuer of such securities. In addition, employees of the Companies and of Daiwa Institute of Research and its affiliates may have positions
and effect transactions in such securities or options and may serve as Directors of such issuers. Daiwa Securities SMBC Europe Limited may, to the extent permitted by applicable UK law and other applicable law or regulation,
effect transactions in the Securities before this material is published to recipients.
This document is intended for investors who are not private investors within the meaning of the Rules of The Financial Services Authority and should not, therefore, be distributed to private investors in the United Kingdom. This
document is not to be used or considered as an offer to sell or solicitation of an offer to buy any securities. Information and opinions contained herein have been compiled or arrived at from sources believed to be reliable, but
none of the Companies, Daiwa Institute of Research and/or its affiliates accepts liability for any loss arising from the use hereof or makes any representation as to their accuracy or completeness. Any information to which no
source has been attributed should be taken as an estimate by Daiwa Institute of Research. This document is not to be relied upon as such or used in substitution for the exercise of independent judgement. Should you enter into
investment business with Daiwa Securities SMBC Europe’s affiliates outside the United Kingdom, we are obliged to advise that the protection afforded by the United Kingdom regulatory system may not apply; in particular, the
benefits of the Financial Services Compensation Scheme may not be available. Additional note to our U.S. readers: Daiwa Securities America is distributing this document in the United States and, in connection therewith,
accepts responsibility for its contents. Any U.S. customer wishing to effect a transaction in any Security should do so only by contacting a representative at Daiwa Securities America in the U.S.
Recommendation Guidance
1 - expected to outperform by more than 15%, 2 - expected to outperform by between 5% and 15%, 3 - expected to deviate no more than plus or minus 5%, 4 - expected to underperform by between 5% and 15%, 5 - expected to
underperform by more than 15%. All DIR recommendations are made on a 6 months basis. Our benchmark is the DJ STOXX 600 index.
North America
This report is distributed by Daiwa Securities America Inc. (DSA). It was prepared by Daiwa Institute of Research Europe Ltd (DIREL), a UK company and an affiliate of DSA. It may not be accurate or complete and should not
be relied upon as such. It reflects the preparer’s views at the time of its preparation, but it is provided with a time delay and does not reflect events occurring after its preparation, nor does it reflect DSA’s views at any time.
Neither DSA nor the preparer has any obligation to update this report or to continue to prepare research on this subject. This report is not an offer to sell or the solicitation of any offer to buy securities. Unless this report says
otherwise, any recommendation it makes is risky and appropriate only for sophisticated speculative investors able to incur significant losses. Readers should consult their financial advisors to determine whether any such
recommendation is consistent with their own investment objectives, financial situation and needs. Unless applicable law permits otherwise, non-U.S. customers wishing to effect a transaction in any securities referenced in this
material should contact a Daiwa entity in their local jurisdiction. U.S. customers wishing to obtain further information or effect transactions in any securities mentioned in this report should contact DSA (Financial Square, 32 Old
Slip, New York, New York 10005 Telephone 212 612-7000).
Ownership of Securities. DSA and its affiliates beneficially own one percent or more of a class of common equity securities of the following issuers: Belx Co., Ltd.; Daiko Denshi Tsushin Ltd.; Ehime Bank Ltd.; Imperial Hotel,
Ltd.; Japan Securities Finance Co. Ltd.; Japan Securities Agents, Ltd.; Kokusai Densetu Co., Ltd.; Ltd.; Meisei Electric Co., Ltd.; NIF Ventures Co. Ltd.; Nomura Co., Ltd.; Shinkin Central Bank; Usen Corp.; Taito Corporation;
Taiyo Life Insurance Company.
Investment Banking Relationships. DSA or an affiliate has managed or co-managed a public offering of securities in the past 12 months, or has received compensation for investment banking services in the past 12 months, or
expects to receive or intends to seek compensation for investment banking services in the next three months, for or from the following issuers: ABN Amro; AC Real Estate; Acom Co. Ltd.; A&D Co Ltd.; Admiral Systems;
Advance Create Co. Ltd.; Aeon Credit Service Co. Ltd.; Aichi Bank, Ltd/The; Akindo Sushiro Co. Ltd.; Akzo Nobel NV; Alliance & Leicester Plc; All Nippon Airways Co. Ltd.; Ajinomoto Co, Inc.; Alpha Credit Group, Inc.;
American Golf UK, AnGes MG Inc.; Anglo American Capital; Anritsu Corp.; AOL Time Warner, Inc.; A-One Seimitsu Inc.; Apamanshop Co. Ltd.; Aplix Corp.; Arata Corp.; Arrk Corp.; Asahi Breweries Ltd.; Asahi Broadcasting
Corp.; Asahi Denka Kogyo KK; Asahi Glass Co. Ltd.; Asahi Pretec Corp.; Asatsu-DK, Inc.; ASB Bank Ltd.; Associant Technology, Inc.; A&T Corp.; Au Optronics Corp.; Australia and New Zealand Banking Group Ltd.; AviChina
Industry and Technology Company Limited; Banca Comerciala Romana SA; Banco BPI S.A.; Bandai Co Ltd.; Bandai Networks Co. Ltd.; Bandai Visual Co. Ltd.; Bank of Nova Scotia Ltd.; Bank of Western Australia; BAT
International Finance; BCP Finance Bank; Belluna Co. Ltd.; BHP Finance USA Ltd.; BHP Billiton Ltd.; BHW Bausparkasse AG; BMW AG; BNP Paribas; Boeing Co.; Boise Cascade Corporation; BOS International Australia Ltd.;
Bradford & Bingley; Bunadarbanki Islands; Business Bank Consulting Co. Ltd.; C4 Technology Inc.; Caja Duero; Calsonic Kansei Corp.; CAM Global Finance; Caixa Catalunya International Finance; Casio Computer Co. Ltd.;
Cassina Ixc. Ltd.; Cawachi Ltd.; Central Japan Railway Co.; Chiba Bank Ltd./The; China Life Insurance Co.; China Resources Power Holdings Co., Ltd.; Chloride Group; Chi Mei Optoelectronics; Chohung Bank; Chubu Electric
Power Co. Inc.; Chugoku Electric Power Co Inc/The; Chugoku Marine Paints Ltd.; Chunghwa Telecom; Citigroup Inc.; Citizen Watch Co. Ltd.; Cleanup Corp.; CNA Co. Ltd.; CNOOC Ltd; Cognis Group; Commonwealth Bank of
Australia; Concorde Securities; Continental AG; Cosel Co. Ltd.; Cosworth Tech/LDV/ MSX Int.; Credit Saison Co. Ltd.; Crymson Co. Ltd.; Culture Convenience Club Co. Ltd. Cyber Communications, Inc.; Daicel Chemical
Industries Ltd.; Daidoh Ltd.; Daihatsu Diesel; Daiichi Pharmaceutical Co. Ltd.; Daikokutenbussan Co. Ltd.; Dainippon Screen Mfg; Daio Paper Corp.; Daiwa Securities Group Inc.; Daiyu Eight Co. Ltd.; Depfa Bank plc.; Depfa
Deutsche Pfandbriefbank AG; Deutsche Bank Finance; Dexia Credit Local; Diamond City Co. Ltd.; Diamond Lease Co. Ltd.; Dip Corp.; Dorma Holding GmbH & Co.; Dow Chemical Company; Dydo Drinco, Inc.; East Japan
Railway Co.; Ebara Foods Industry Inc.; EFG Hellas plc; Elecom Co. Ltd.; EMTEC Consumer Media GmbH, Endo Manufacturing Co. Ltd.; Eneserve Corp.;. ES-CON Japan Ltd.; Eurohypo AG.; Excel Co. Ltd.; Exel; Finconsumo
SPA; First Financial Holding Co., Ltd.; First Juken Co. Ltd.; Ford Credit Australia Limited; Freddie Mac; F T Communications Co. Ltd.; Fuji Co. Ltd.; Fujicco Co. Ltd.; Fuji Corp. Ltd.; Fuji Country; Fuji Electric Co Ltd.; Fuji Heavy
Industries Ltd.; Fuji Logistics Co Ltd.;Fujimori Kogyo Co. Ltd.; Fujitsu Ltd.; Furukawa Electric Co. Ltd.; FuKoKu Co. Ltd.; Gambro, GECC; GE Capital Australia Funding; General Packer Co. Ltd.; Gentosha; GK Japan Agency
Co. Ltd.; Global One Real Estate Inverstment Corp.; Glory Ltd.; GM; GMAC; Goldman Sachs Group, Inc.; Gulliver International Co. Ltd.; Haba Laboratories, Inc.; The Hachijuni Bank Ltd.; Hana Bank; Hankyu Corp.; Harakosan
Co. Ltd.; HBOS Treasury Service; HCA Inc.; Heiwa Real Estate Co. Ltd.; Hellenic Bank; Henkel; Hertz Corp.; Hikari Tsushin, Inc.; Hiroshima Bank Ltd./The; Hitachi Ltd.; Hitachi Construction Machinery Co. Ltd.; Hokkaido
Electric Power Co. Inc.; Hokuriku Electric Power Co.; Honeys Co. Ltd.; Hon Hai Precision Industry Co Ltd.; Hoosiers Corp.; Hopewell Highway Infrastructure Ltd.; Horiba Ltd.; Hosokawa Micron Corporation; HSH N Finance;
Hyundai Motor Co. Ltd.; Ibiden Co Ltd.; IK Co. Ltd.; Impexbank; International Business Machine Corp; International Hotel Group; Ishikawajima-Harima Heavy Industries Co. Ltd.; Itochu Corp.; Itochu Europe; Ito En Ltd.; Ito-
Yokado Co. Ltd.; Izumiya Co Ltd.; JFE Holdings Inc.; Jaccs Co. Ltd.; Japan Airalines Co. Ltd.; Japan Petroleum Exploration Co.; Japan Prime Realty Investment Corp.; Japan Pulp & Paper Co. Ltd.; Japan Real Estate
Investement Corp.; Japan Wind Development Co. Ltd.; JSR Corp.; KGaA; Kajima Corp.; Kakaku.com, Inc.; Kanebo; Kansai Electric Power Co.; Karula Co. Ltd.; Kappa Create Co. Ltd.; KBC IFIMA NV; Keisei Electric Railway
Co. Ltd.; Kinki Nippon Railway Co. Ltd.; Kintetsu World Express Inc.; KOGAS; Kohnan Shoji Co. Ltd.; Kohsoku Corp.; Komatsu Ltd.; Kookmin Bank; Koram Bank; Kourakuen Corp.; Koyo Futures Co. Ltd.; Koyo Seiko Co. Ltd.;
KT Corp.; Kubotek Corp.; Kureha Chemical Industry Co. Ltd.; Kyushu Electric Power Co. Ltd.; LAND Co. Ltd.; Landbanki Islands HF; Lehman Brothers Holding; Leopalace21 Corp.; LG Electronics, Inc.; Lian Hua Supermarkets,
Linde AG; Link Consulting Associates-Japan Corp.; Mag Garden Corp.; Maeda Corp.; Macquarie Bank Limited; Mandom Corp.; Marubeni Corp.; Maruetsu, Inc./The; Matsushita Electric Works Ltd.; Matsushita Electric Works
Information Systems Co.; Mazda Motor Corp.; MDM Holdings GmbH; Meadwestvaco Corp.; Mediatek Inc.; Meiji Seika Kaisha Ltd.; Meisei Electric Co. Ltd.; Meisei Industrial, Meitetsu Transportation Co.; MG Technologies;
Mikikogyo Co. Ltd.; Mikuni, Millennium America; Mitsubishi Corp.; Mitsubishi Chemical Corp.; Mitsubishi Electric Corp;. Mitsubishi Heavy Industries Ltd.;. Mitsubishi Kakoki Kaisha, Mitsubishi International GmbH; Mitsubishi
Tokyo Financial Group Inc.; Mitsui & Co. Ltd.; Mitsui Chemicals, Inc.Mitsui Mining & Smelting Co Ltd.; Miyoshi Oil & Fat Co. Ltd.; Moc Corp.;Modec, Inc.; Nafco Co. Ltd.; Nagoya Railroad Co. Ltd.; Nanya Technology Corporation;
National Bank of Canada; National Grid Co. Plc.; National Rural Utilities Cooperation Finance Corp.; Nationwide Building Society; NEC; NEC Corp.; NEC Electronics Corp.; NEC System Technologies Ltd.; Nestle Australia
Limited; Nichirei Corp.; Nidec Copal Corp.; Nihon Trim Co. Ltd.; Nihon Unisys Ltd.; Niitaka Co Ltd.; Nikon Corp.; Nippon Cable System Inc.; Nippon Electric Glass Co. Ltd.; Nippon Information Development; Nippon Oil Corp.;
Nippon Restaurant Corp.; Nippon Restaurant System, Inc.; Nippon Sanso Corp.; Nippon Sheet Glass Co. Ltd.; Nippon Steel Corp.; Nippon Telegraph & Telephone Corp; Nippon Unipac Holding; Nippon Yusen Kabushiki
Kaisha; Nishimatsu Construction Co. Ltd.; Nishimatsuya Chain Co Ltd.; Nishi-Nippon Railroad Co. Ltd.; Nissen Co. Ltd.; Nissho Iwai Corp.; Nissin Food Products Co. Ltd.; Nittyu Co. Ltd.; Niws Co. Ltd.; Noritsu Koki Co. Ltd.;
Noritz Corp; Northern Rock Plc.; NTN Corp.; NTT DoCoMo Inc.; Obayashi Corp.; Oenon Holdings, Inc.; Orix JREIT, Inc.; Odakyu Electric Railway Co. Ltd.; O&H Capital; OJI Paper Co. Ltd.; Okamoto Glass Co. Ltd.; Okinawa
Electric Power; Okuwa Co. Ltd.;Onco Therapy Science Inc. Onoken Co. Ltd.; ORIX Corp.; Osaka Gas C. Ltd.; Overnight Corporation; OX Information; Pasona, Inc.; Peoples Insurance Company of China; PHH Corp.;; Pla
Matels Corp.; Pochet SA; Point, Inc.; POSCO; Postabank rt; Precision System Science Co. Ltd.; Premier Investment Co.; Promise Co. Ltd.; QP Corp.; Rabobank Nederland; Rakuten Inc. R.S. 2 Software; Rann International;
Reins International Inc.; Renaissance Inc.; Renault S.A.; Renown; Renown; Rentokil Initial; Rinnai Corp.; Roland DG Corp.; Royal Bank of Scotland; Ryowa Life Create Co. Ltd.; Sakata INX Corp.; Samsung Electronics Co.
Ltd.; San-A Co. Ltd.; Sanko Marketing Foods Co. Ltd.; Sanyo Electric Co. Ltd; Sapporo Breweries Ltd.; Sapporo Drug Store Co. Ltd.; S&B Foods, Inc.; Sawai Pharmaceutical Co Ltd.; Sega Toys Ltd.; Seiko Epson Corp.; Senko
Co. Ltd.; Sharp Corp.; Shikoku Electric Power Co. Inc.; Shimamura Co. Ltd.; Shiomi Co Ltd.; Shiseido/Milbon; Shochiku Co. Ltd.; Showa Denko, Showa Shell Sekiyu KK; Shuei Yobiko Co. Ltd.; Silex Technology, Inc.;
Siliconware Precision Industries Company; Silicon Sensor International AG, Singapore Telecommunications Ltd.; Site Support Institute Co. Ltd.; SK-Electronics, Ltd.; Skylark Co. Ltd.; Softbank Investment Corp.; Sohken
Homes Co. Ltd.; Soiken Inc.; Sotsu Agency; Sparebanken Rogaland; Spectris; SPK Corp.; Stanley Electric Co. Ltd.; STB Leasing Co. Ltd.; St. Marc Co. Ltd.; Studio Alice Co. Ltd.; Sumitomo Chemical Co. Ltd.; Sumitomo Light
Metal Industries Ltd.; Sumitomo Metal Industries Ltd.; Sumitomo Metal Mining Co. Ltd.; Sumitomo Mitsui Banking Corp.; Sumitomo Mitsui Financial Group Inc.; Sumitomo Osaka Cement Co. Ltd.; Sumitomo Realty &
Development Co. Ltd.; Sumitomo Trust & Banking Co Ltd/The; Sumitomo Wiring Systems Ltd.; Sunpot Co. Ltd.; Suzuken Co. Ltd.; Suzumo Machinery Co Ltd.; Ltd.; Sysmex Corp.; Tac Co. Ltd.; Taisei Corp.; Taiyo Life Insurance
Company; Taiyo Tuden Co. Ltd.; Takachiho Electric Co. Ltd.; Takara Holdings, Inc.; Tanaka Co. Ltd/Tokyo; Techno Medica Co. Ltd.; Telemedia; Telstra; Telstra Corp Ltd.; T Hasegawa Co. Ltd..; Tietech Co. Ltd.; Tocalo Co.
Ltd.; Toho Gas Co. Ltd.; Tohoku Electric Power; Tokai Rika Co. Ltd.; Tokyo Electron Device Ltd.; Tokyo Electric Power Co.; Tokyo Gas Co. Ltd.; Tokyu Corp.; Tokyu REIT, Inc.; Tokyu Store Chain; Tomen Cyber-Business
Solutions Inc.; Tomen Devices Corp.; Toray Industries Inc.; Toray Medical; Toshiba Corp.; Touei Housing Corp.; Toyoda Boshoku Corp.; Toyoda Gosei Co Ltd.; Toyoda Machine Works Ltd.; Toyo Radiator Co. Ltd.; Toyota
Credit Canada, Inc.; Toyota Finance Australia Ltd.; Toyota Motor Corp.; Toyota Motor Credit Corporation; Toyo Tire & Rubber Co Ltd.; TT Electronics; Tullet & Tokyo; UFJ Holdings Inc.; Ulvac Inc.; Union Oil Co. of California;
Union Pacific Corp.; United Microelectronics Corporation; United Urban Investement Corp.; US Cellular Corp.; Valic Co. Ltd.; Valor Co. Ltd; USC Corp.; US Cellular Corp.; VeriServe Corp.; VeriSign Japan KK; Verizon Global
Funding Corp.; Viacom Inc.; Vivanco Gruppe AG/Vivanco Electronic; Volkswagen Financial Services AG; Volkswagen International Finance NV; Wacom Co Ltd./Japan; Warabeya Nichiyo Co. Ltd.; Webzen Inc.; Wella AG; West
Japan Railway Co. ; Westpac Banking Corporation; Wintest Corp.; World Co. Ltd.; XNET Corp.; Yahoo Japan Corp.; Yokohama Rubber Co. Ltd./The; Yumeshin Co Ltd.; Yusen Air & Sea Service Co. Ltd.; Ziphyr Co. Ltd.
The statements in the two preceding paragraphs are made as of January 1, 2004.
DSA Market Making. DSA made a market in securities or ADRs of the following issuers at the time this report was published: Mitsui & Co.; NEC; Millea Holdings; Trend Micro; Crayfish; Makita; Internet Initiative Japan; Fuji Photo
Film; Kirin Brewery; Nissan Motor; Sanyo Electric; CSK Corp.; Daiei Inc.; Wacoal Corp.; All Nippon Airways; Bandai; Bridgestone; Casio; Fujitsu; Isuzu; Sega; Sharp; Shiseido; Q.P. Corp.; and Sumitomo Corporation.
Research Analyst Conflicts. The principal research analysts who prepared this report have no financial interest in securities of the issuers covered in the report and are not aware of any material relevant conflict of interest
involving the analyst or DSA, except as noted: no exceptions.
Hong Kong
This publication is produced by Daiwa Institute of Research Europe Limited, and distributed by Daiwa Institute of Research (Hong Kong) Limited and/or its affiliates. This publication and the contents hereof are intended for
information purposes only, and may be subject to change without further notice. Any use, disclosure, distribution, dissemination, copying, printing or reliance on this publication for any other purpose without our prior consent or
approval is strictly prohibited. Neither Daiwa Institute of Research (Hong Kong) Limited, nor any of its affiliates, nor any of its or their respective directors, officers, servants and employees, represent nor warrant the accuracy or
completeness of the information contained herein or as to the existence of other facts which might be significant, and will not accept any responsibility or liability whatsoever for any use of or reliance upon this publication or any of
the contents hereof. Neither this publication, nor any content hereof, constitute, or are to be construed as, an offer or solicitation of an offer to buy or sell any of the securities or investments mentioned herein in any country or
jurisdiction nor, unless expressly provided, any recommendation or investment opinion or advice. Any view, recommendation, opinion or advice expressed in this publication may not necessarily reflect those of Daiwa Institute of
Research (Hong Kong) Limited, except where the publication states otherwise.
Daiwa Institute of Research (Hong Kong) Limited, its parent, holding, subsidiaries or affiliates, or its or their respective directors, officers, servants and employees may have positions in or may be materially interested in the
securities in any of the markets mentioned or may have also performed investment banking and other services for the issuers of such securities.
Additional information may be available upon request.

Eurotelcorama 05 300104

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Eurotelcorama 05 300104

Uploaded by

Copyright:

Available Formats

DAIWA

DAIWA INSTITUTE OF RESEARCH LTD.

30th January 2004 (No. of pages: 109)

In such a complex and uncertain environment, we think stock selection requires a

• We are upgrading Telefonica from rating 3 [NEUTRAL] to rating 2

Important disclosures are provided on the last page of this report.

Global Equity Research

Executive summary 3-4

2. A tale of two markets 14 - 18

3. Understanding the customer 19 – 41

4. Europe – the demographic timebomb 42 - 50

5. Challenges for the telcos – the access layer 51 - 73

6. Challenges for the telcos – the service layer 74 - 95

The challenges in the access layer identified in this report are:

• Mobile substitution in access;

• VoIP - in the enterprise, as a consumer product tied to broadband access, or as

As a result of our review, we have made the following ratings actions:

• We are upgrading Telefonica from rating 3 [NEUTRAL] to rating 2

On a capitalization–weighted basis, our current individual stock ratings leave us with a

EU electronic communications market value (€bn)

Source: EU ninth report

UK communications as a % of total household consumption 1981 – Q2 2003

1.90% As % of household expenditure

Source: DIR, UK National Statistics

Household final consumption in the UK, shares and changes in shares

Source: DIR, UK National Statistics

Estimated average monthly expenditure of a standard PSTN user (€)

40.0 34.0 31.5 29.8 30.0 29.5 29.4

Source: EU ninth report

2004 EPS forecasts – ratio of upward to downward revisions

Meanwhile, for the worst-offending companies, concerns over the long-term

Net debt/EBITDA evolution of Europe’s four “problem” PTTs

Source: DIR, company data

6 0 0 ,0 0 0 2000 2001 2002

EBITDA per employee (EUR)

Source: DIR, company data. N.B. data based on half-year results

Reported personnel costs as % of sales

Source: DIR, company data.

Source: DIR, company data

European broadband market development, 2000 – Q3 2003

70% Cable market share

Source: DIR, from various

Source: DIR estimates, company data

Sequential % PSTN line loss Q1 2002 – Q3 2003

Source: DIR, company data

Source: DIR, company data

Outgoing minutes per channel per day on the PSTN

Source: DIR, from national regulator and operator data

Source: DIR, from national regulator and operator data

SMS per day per channel

Source: DIR from national regulator and operator data

National calling ex-

Source: DIR, from national regulator data

UK (left) and French (right) fixed-line call volume breakdown

Source: DIR, from national regulator data

France & UK mobile channel and mobile traffic shares

28% France mobile traffic/total (RHS)

Source: DIR, from national regulator data

France and UK voice yield mobile/fixed (x)

1.5 UK mobile yield/fixed yield voice

Source: DIR from national regulator data

France and UK mobile voice yield/SMS yield (x)

Source: DIR from national regulator data