1. The (Slow) Fiber Revolution
Benoît Felten, CEO
benoit@diffractionanalysis.com
2. Why a Fiber Revolution ?
• 2007 saw the birth of www.fiberevolution.com
• Why is it a revolution?
• Abundant supply of bandwidth
• Future proof wireline broadband
• Opportunity for new entrants in the market
• Opportunity to rethink social organisation around
universal connectivity
• In markets with advanced fiber penetration, the online service
economy is thriving.
3. The reach of 100 Mbps service over Fiber is expanding
Finland
Sweden
Norway
Baltic Russia
Netherlands States
France Slovenia South Korea
North-Eastern USA (VZ) Romania
Japan
Chattanooga, TN Portugal Italy Turkey
Taiwan
New Delhi Hong Kong
Singapore
Sao Paulo
Places where 100 Mpbs FTTP service is available
with more than 100k homes passed.
5. The Reach of Gbps Internet Service Still Limited
Telia
Telia
Bredbands Bolaget
Bredbands Bolaget
& 5 other ISPs
& 5 other ISPs
T2 Slovenia NTT
NTT
T2 Slovenia
EPB Fiber ZON Superonline KDDI
KDDI
EPB Fiber ZON Superonline
HKBN
HKBN
PCCW
PCCW
Places where 1Gpbs FTTP service is available
with more than 100k homes passed.
7. And yet...
Average Speeds per Region
(Source: Akamai State of the Internet)
7.00
6.00
5.00
Average Speed (Mb/s)
Europe
4.00
APAC
Middle East
3.00
Latin America
North America
2.00
1.00
-
Q42008 Q12009 Q22009 Q32009 Q42009 Q12010 Q22010 Q32010 Q42010 Q12011 Q22011 Q32011 Q42011
1-2 quarter decline accross the board
8. Is wireline quality of experience degrading?
• Too early to draw definitive conclusions on the basis of 2 quarters
of decline, but worth keeping an eye on.
• If there is continued decline, explanations could be:
• Akamai’s data gathering methodology has changed
• Economic crises causes customers to downgrade their
subscriptions
• Quality of experience is really degrading
• Content is shifting to higher standards of quality, the number of
devices pulling on set network resources is increasing but
broadband access isn’t following suit.
9. The Incumbents’ Passive Resistance
• Incumbents that are not facing a very serious competitive threat on the
network layer (from cable or from another FTTP new entrant) have no
incentive to aggressively deploy NGA:
• They already own a very profitable (though obsolete) infrastructure
• They are no longer structured for long-term investment
• Incumbent resistance to FTTP takes different forms.
Fiber to the Manipulating Future Siphoning
Press Release Needs Public Subsidies
The incumbent makes The incumbent insists that the The incumbent convinces
numerous broadband needs of users is local and national
announcements, but doesn’t not so high (and even gets governments that any public
actually implement much if broadband redefined in that subsidies for broadband
anything (France…) direction…) (US…) should go to the incumbent
(UK…)
Litigate, litigate, litigate…
10. Mythbusting: There is no FTTH/B “demand issue”
• Despite some variation on a project by project basis, take-up is generally
a factor of time.
Take-up vs years in operation
• Acquisitions 90%
strategies aim to Take-up across deployment
80%
accelerate take-up
70% Altibox
60% Bredbands
• Premium 50%
Networx
Bolaget
strategies tend to
40%
slow it down Superonline
30% Verizon HKBN
KPN
20%
• Larger projects TEO
10%
and incumbents Orange PT
have generally 0%
lower take-up 0 2 4 6 8 10 12 14
Years in operation
11. Mythbusting: FTTH/B ARPU on average is 46% higher than DSL
• When comparing FTTH/B Average Revenues Per User (ARPU) with DSL
ARPU for the same player (or, in case of FTTH/B only players for the
incumbent in the same market), across the sample we see that on
average FTTH/B ARPU is 46% higher.
DSL ARPU vs FTTH/B ARPU
(normalized USD)
60
50
40
30
55
20 38
+ 46%
10
0
Average DSL ARPU Average FTTH/B ARPU
12. Why FTTH/B ARPU is higher?
• Overall, FTTH/B prices are
higher for equivalent service
propositions (bandwidth
excepted)
• FTTH/B customers tend to
subscribe to more services (ie.
more revenue generating
units for the service provider)
• FTTH/B customers tend to buy more options and pay as you go services:
premium channels, VoD movies, multi-screen options, etc.
• Examples:
• On average 2.5 VoD per month on DSL and 7 on FTTH/B
• 90% of FTTH/B customers have triple play vs. 15% on DSL
13. Competitive ARPU vs Incumbent ARPU
• Incumbent operators tend to be Competitive ARPU vs Incumbent ARPU
more cautious regarding (base 100 = average DSL ARPU)
FTTH/B pricing strategies. 200
• Competitive operators have to 150
reduce the gap with DSL ARPU
to develop their customer base. 100
178
• Incumbent operators have 114
50 94 104
lower incentives to sell
aggressively due to the
0
revenues from copper local
Average DSL ARPU Average FTTH/B ARPU
loop. They call it “managing the
transition”. Competitive Incumbent
Competitive + 121%
DSL FTTH/B
Incumbent ARPU + 172%
ARPU
14. Original and Compounded Policy Sins
• The original sin of telecom policy makers was not understanding that
infrastructure and services operate:
• On different timeframes
• With different investment structures
• With different skillsets
• With different market dynamics
• The compounded sin of telecom policy makers is thinking they can find
incentives for incumbents to invest in next-generation infrastructure.
• Policy makers need to wake up and smell the coffee: only by encouraging
infrastructure renewal through new entrants (municipalities, utilities, new
infrastructure players…) can they hope to
15. The Missing Piece
• Telecom funds don’t want to
finance long-term
infrastructure.
• Infrastructure funds don’t
want to finance telecom
players.
• Awareness of the
infrastructure opportunity is
rising!
16. The Revolution Will Not Be Television
• Betting on incumbents to drive next-generation access networks is misguided
and will not work.
• Policy makers should focus on eliminating all barriers to alternative players –
public or private – deploying infrastructure.
• Television is no longer the driver for NGA deployment and adoption. NGA
should be viewed as very profitable long term infrastructure investment and
funded on that basis, which would ensure:
• a competitive landscape (optimal take-up realised through open access)
• A universal deployment (if the right compensation mechanisms are put in place)
• A future proof technology (as long-term investors seek revenue assurance)