Article Info

Free FTTH in France

image Fiber to the Home broadband is not Free in France, but Illiad (Corporate Website) is rolling out FTTH in Paris through it’s broadband ISP Free (note the capitals).

We’ve written about Free before (see “FTTH vs. VDSL in France“) but had a chance to learn more about what Free is doing while having Dinner with Benoit Felten. He consults by day and blogs by night over at Fiberevolution… What an appropriate name for a French Fiber blog.

Here’s what is interesting about Free and their FTTH offering:

  • Iliad disrupted the French broadband business by offering cut rate pricing for broadband DSL and adding more free services over time like VoIP and Video. They are to Broadband in France what Southwest Airlines is to air travel. Iliad doesn’t spend a lot on customer acquisition and relies on word of mouth, high value, and low pricing. They have a cult-like user following and hold 20% of DSL market share and 45% of all unbundled lines in France.
  • According to Benoit, Free pays France Telecom about 10 Euros a month, and sells DSL, unlimited VoIP, and limited pay TV for 30 Euros a month. ARPU climbs to $35 a month with a few extras.
  • Iliad is NOT deploying PON, they are doing point to point fiber. This is a massive bump in the number of SFP modules needed and presents an opportunity for module makers. Free’s architecture uses almost 2x the optics of a PON install, and at least an order of magnitude more in cabling complexity. These are not FTTH modules but instead are Ethernet SFPs sold primarily by Avago and Finisar (FNSR).
  • Install capex is estimated by Iliad to average 1500 Euros a subscriber in Paris. That is a lot considering one building in Paris might have multiple subscribers.
  • Free is offering FTTH for the same price as DSL. 100M FTTH and all of the other features as before for 30 Euros. Their revenue model remains the same though their costs are going up.
  • They can justify this (and this is the debatable point) because once the FTTH is in place, they no longer have to pay 10 Euros a month (120 a year) for copper. According to Benoit, Iliad is stringing two fibers, one for themselves, and another for open access. If a competitor such as France Telecom (FTE) or Neuf Cegetel want to use the fiber, they pay Iliad15 Euros a month.

I’ve got to hand it to Free. I slammed them hard (see Expropriation is Not Competition) because of their success by cherry picking the best customers (otherwise known as redlining here in the US). But at least they are now putting up their own risk capital for the venture (even though they still get a regulated ride in France Telecom cable ducts).

The financial reality is that spending 1500 Euros today to save 120 Euros/year in operational expenses tomorrow is a bad investment. That’s a 5% return over a 20 year investment horizon which, pardon my French, sucks. Something has to change.

Here are what I see as likely outcomes:

  1. Iliad runs out of cash. The way things are set up now with razor thin return margins, this is a financial certainty in a rational market. Something has to give, or like Ferdinand de Lesseps, Free’s second act is going to be a disaster.
  2. Free charges more. Free is delivering massive value. And if FTTH is really what people want, they will pay a premium.
  3. They beg for the French government to guarantee bonds. This will eventually be needed to make the economics of FTTH work in areas of less density than France, unless the users themselves bear the costs.
  4. Do it cheaper. This is the most likely outcome. Free can put FTTB (Fiber to the Building) and use the copper from the basement (owned by the landlord) to the apartment. They could do this to existing ADSL subscribers transparently, and could use VDSL to reach 50Mb/s in the same manner. No messy wiring, less optics, no 10 euro a month tariff, minimal new fiber, same user experience. At some point in the future they could go fiber the whole way with PON or Pt to Pt.

It should be noted that when Free first proposed using FTTH, they claimed it would cost 1B Euros to pass 4M homes. In their latest September report they indicate costs are up to 1500 Euros/subscriber. A move to VDSL is the quickest way to rein costs back in, something that would benefit incumbent Sagem and their chip supplier, Ikanos (IKAN).

Author holds positions in Finisar, Ikanos, and NTT.


Bonus reading:

In order to understand Free’s business model we need to examine the three forms of regulation in play.

  • No regulation – This is what we have in the US at this point. Verizon, AT&T, Qwest own the copper lines from the Central Office (CO) to the home. They can lease them to competitive providers at any price they wish. This applied to residences only but rulings are emerging that will continue this trend to business DSL as well. The Telecom Act of 1996 forced unbundling at set prices and led to the explosion of CLEC’s. This regulation was taken away with multiple turns of legislation and eliminated by the Brand X decision.
  • Separation – Newer model that British Telecom follows as dictated by the OFCOM (British Office of Communications). BT still owns everything, but a separate, regulated subsidiary called Openreach maintains and leases the local loop copper connections to several competitive DSL providers like Carphone Warehouse (CPW) or a host of others. Or, Openreach leases it to BT itself. CPW and their competitors can choose to put their own DSLAMs in the BT CO and connect them to Openreach copper, or they can buy a wholesale DSL connection from British Telecom if they don’t have their own equipment present. Separation is the hot buzzword for Viviane Reding, who would like to use this model for the rest of the EU. We like to refer to her as the Telecom Witch Queen for her gross anti-capitalistic policies and affinity for central planning.
  • Unbundling – The norm for Europe and Japan. The old PTT monopolies own and maintain the copper but are forced to lease access rights to other providers at low rates. ISP’s like Free in France or Yahoo/SoftBank in Japan pay the incumbent a regulated tariff to run their signals over the copper and use space in the CO.

From an editorial standpoint I acknowledge the shortcomings of the the no regulation model but prefer it to the others. If incumbents or new entrants don’t appear in a market, municipalities should be free to install fiber. When sufficient demand and need exists the risk capital will appear to make things happen.


Comments are disallowed for this post.

  1. * good write up …
    * Iliad reminds me sooo much of Yahoo BB Japan 5 years ago.
    * They are fun to read about.
    * Majorly aggressive in services. They even have their own residential boxes.
    * Back then I couldn’t figure out how Yahoo BB was going to survive. I feel the same way about Iliad.
    * But I really like that they are innovating like mad instead of wishing the Internet away.

    * I sure hope they are a success.

    Posted by iain | October 12, 2007, 5:43 PM
  2. Agreed… but Yahoo never went on a major capex binge.

    Posted by Andrew Schmitt | October 12, 2007, 8:00 PM
  3. Andrew

    I happen to be in Paris to visit Free. The first observation is that they actually are delivering what they promise – unlimited DSL, unlimited voice to 30 countries (except mobile) and 80 channels of TV – for 30 euro. They can do that and still be so profitable that they pay serious corporate income taxes. One reason it works is that Xavier runs the slimmest and most efficient operation I’ve seen in eight years of covering telecom. His staff, especially engineering, is enormously productive.

    His bet on fiber makes more sense than the quick numbers suggest. To begin with, the 1500 figure is a red herring. The initial build is in Paris, where the man-sized sewers mean they won’t need to dig. Future builds will primarily be in areas they can deploy on polls. Free hasn’t given any estimates, but I’d guess this will bring the cost down to 600-800 in three years.

    Additionally, the benefits are far more than the 10 euro/month saving from FT. Knowing they wold one day get fiber, Free new subscribers recently doubled from already healthy growth. These are incremental subs on an already built network, highly profitable. When he actually deploys, expect a very high take rate. If a building has a choice between 100 megabit fiber and DSL at the same rate, most will take Free’s fiber. Free saves a lot of money with customer acquisition costs well below 100 while others spend 200-300. Giving more for the same price is a better sales tool than advertising.

    Add to that the capability of offering more and better service, including HD TV at full quality rather than the over compressed form of cable, satellite or other DSL providers. Every customer has a hard drive, so VOD can be pre-downleaded to play instantly. …

    Factoring in the reduced customer acquisition, higher penetration, lower opex, and extra service capabilities, Free is probably seeing 15 euro/month benefit.

    180 euros per year benefit on a 700-1000 euro investment is much more reasonable, especially for someone who can afford to wait for the longterm return. Niel owns the majority of stock, worth $billions. He can look ahead.

    Other side benefits include that Niel becomes a national chapmion, able to compete with FT, Alcatel, etc. whose political power could be overwhelming. He is curently being offered the fourth national wireless license, and is negotiating the price. That’s one heck of a return on an announcement.

    Dave Burstein

    Posted by Dave Burstein | October 13, 2007, 12:14 AM
  4. In Australia we’ve moved from “Unregulated” to “Unbundling” and now looking at “Separation”. There are different grades of Regulation, and I agree that too much is a bad thing, however the appropriate grade depends on the environment. In the US “Unregulated” may be appropriate, but in a small market like Australia with only one incumbent “Separation” is the only way to create fair competition.

    Posted by Julian | October 13, 2007, 12:44 AM
  5. A very nice writeup indeed, Andrew. Thanks. Missing in this account is any mention of IMS, FMC or wireless convergence of any type. I suspect that this avoidance of back office tonnage affords Free a level of savings that makes it easier for them to move forward with an all-you-can-eat VoIP regime while making available speeds that would make most incumbents cringe. Were these IMS-related ‘capabilities’ discussed at dinner, as well, assuming ‘capabilities’ is the right term to use here? Curious …

    Posted by Frank A. Coluccio | October 13, 2007, 1:53 AM
  6. Dave’s comments are, on the whole, spot on. However, I should point out a number of things seeing as it was largely things Andrew and I discussed which prompted his posting about this.

    First of all, on the issue of 1500 being a high figure, I must respectfully disagree with Dave. Remember that this is 1500 per existing customer to cover all houses (customers and non customers) in an area where Free has 15% coverage. So what it comes down to is most actually more like 300-400 per home connected, which sounds about right for Paris, but will rise sharply with lower density and absence of available sewer space or access to sewers.

    Regarding aerial, it’s the first I’ve heard about it, and it seems like a surprising proposition to me considering the cities that Free has announced so far. Unless they go for a radical strategy of non-dense area coverage, I can’t see aerial being a significant component of deployment in France…

    Where I wholeheartedly agree with Dave though, is that what makes this business model work is actually that it pays back for itself in six years with no acquisition and no ARPU increase. But that doesn’t mean that there won’t be acquisition and it doesn’t mean their won’t be ARPU increase. It just means that, as important as these may be, they’re gravy.

    Now again, the figures for Paris are, in my opinion, realistic. So if, as Dave says, we’re looking at twice the price outside of Paris on average, that means payback in 10-12 years, in which case acquisition and ARPU increase become mandatory for the business model to work.

    One final point: breaking news here in France is that a merger between Neuf and Free has just failed to happen. That would have put the merged ensemble in a much stronger position with an average of 40-45% penetration in existing COs. It’ll be interesting to see what Free does next…

    Posted by Ben Felten | November 8, 2007, 5:43 PM
  7. I see your point – $1500 with 15% penetration could become $1000 with 25% penetration.

    $1500 a user doesn’t pay itself back in 6 years…. not even close.

    Posted by Andrew Schmitt | November 9, 2007, 9:41 PM
  8. I’m presenting my view of the business model in Amsterdam next week. It’s rax, but I’ll send it over to you once I’m done. I’ll publish my slides on my website anyway.

    Posted by Ben Felten | November 10, 2007, 4:15 AM