Original URL: http://www.theregister.co.uk/2006/11/15/aol_cpw_analysis/

Carphone Warehouse rescued with AOL buy

There's method behind the madness

By Steve Kennedy

Posted in Broadband, 15th November 2006 14:13 GMT

Analysis Mobile phone operator Carphone Warehouse (CPW) has now moved into fixed line telecoms and broadband.

Originally, the company's phone offerings were based on a system known as IDA (Indirect Access) whereby customers dialled a prefix and then the number they actually wanted. All the real work was done by another provider and CPW just did the marketing and billing (and took a cut of revenues).

It then acquired Opal Telecoms who was (and still is) a provider of business telephony services. This gave it the ability to handle calls on its own network and it started offering Carrier PreSelect (CPS) services so customers no longer had to dial a prefix to utilise CPW's tarrifs.

CPW then acquired OneTel and Tele2 which gave it scale in voice customers (around 1.3 million new customers). As all the customers would be on CPW's network, they could all call each other for free.

However, CPW's broadband base was low and the acquisitions only gave it around 60,000 new customers.

CPW also has a growing mobile base of its own as it set-up a UK MVNO (mobile virtual network operator) - known as TalkTalk Mobile. It is also setting up MVNOs in other countries directly as TalkTalk or combined with other brands (with Virgin Mobile in France, for example).

Broadband push

Though fixed line telecoms offers short term revenue, it's not a long-term winner as phone calls are becoming free. CPW currently wins because BT has been so uncompetitive for so long (partly from being regulated and partly because it's BT). Times are changing and the market has become much more competitive, so CPW needed another hook to win customers, and broadband was the answer.

Though CPW's goal was to roll-out a service based on local loop unbundling (LLU) putting its own equipment in BT exchanges, it takes a long time to do this (anywhere over six months, but considerably longer if the exchange needs any work doing to it). LLU has only recently proven to be economic due to price cuts from BT (which have been forced upon it by regulator Ofcom). However, LLU relies on massive scale for those economies to come into play.

Due to the delay in deploying LLU, CPW actually utilises BT Wholesale's IPStream service for broadband (the majority of broadband in the UK is actually on IPStream). There's a fixed cost to connect end-users and then further costs to connect the users back to CPW.

CPW offer "free" broadband as part of a package, whereby the user signs up for a talk package (TalkTalk3) and a line rental package. The line rental is a service known as WLR (wholesale line rental) whereby CPW takes over the copper wires that go into the customer's premises. This also removes the "blue bill" so the customer is now completely owned by CPW and BT can not market to them.

Scale and then there's scale

Carphone Warehouse (CPW) recently bought AOL UK for £370m, which may seem a huge amount for its 1.3m users (about £285 per user), but that's only half the picture.

If you add all the costs together (WLR, broadband etc) it actually costs CPW around £22.17 per month per customer (there's several assumptions here, see the end of the article for more information).

CPW's TalkTalk3 package (broadband and phone) is sold at £19.99 per month, including VAT, which means it actually gets £16.49 per customer.

Putting the two scenarios together means CPW is losing £5.69 per month per customer. Assuming 800,000 customers, that's a whopping £4,542,919.44 per month and the more customers they get, the bigger that loss gets.

The only way to reduce those costs is to move those customers onto its own unbundled broadband solution.

Even if CPW had a workable LLU plan, it is using Chinese broadband kit vendor Huawei, which makes cost effective equipment that tends to be very feature rich. Huawei, however, has little resource in the UK to do CPW's roll-out. Many other broadband vendors do the complete install and maintenance for the broadband provider.

The second major problem is the delay in getting the equipment into the exchanges.

The third problem CPW has is provisioning, or actually the lack of a workable, scalable provisioning system - so customers have been left without service for days, or even weeks.

AOL to the rescue

AOL has 1.3m broadband customers, which gives scale (and LLU is all about economies of scale). It also has 400 DLEs unbundled with a roll-out plan to 1,000. This is all based on Fujitsu kit, and it has an engineering base in the UK and has been doing the roll-out for AOL.

Now CPW can instantly migrate its users onto its own unbundled DLEs (the 400 will be in the most population heavy areas, so probably most of CPW's customers too) and turn that £5.68 per month loss into a profit (albeit a small one).

Though the above figures are accurate in terms of costs (see assumptions below), it's very difficult to know what CPW is actually doing and how it loads the pipes, etc. According to its half-year results it made a (six month) loss of £37.7m (for the broadband service) and had 461,000 broadband customers, which works out to a loss of £14.94 per user per month. However, that loss probably covers LLU roll-out and other factors.

A big advantage in buying AOL (apart from maybe making some money out of LLU migrations) is that it has a working, scalable provisioning system. You don't hear of problems with AOL users losing connections when they sign-up.

So though £370m may initially seem a lot, its saved CPW probably a year's worth of unbundling and a provisioning system that works. It just needs to migrate its customers over to the AOL platforms as soon as possible.

Of course, now CPW has to go the whole hog and start running triple or quad plays, internet and free calls won't cut it for long. It'll be video and mobile integration and other services which make consumers pay and allow CPW to generate a profit for its shareholders. AOL brings a part of the solution with it, content which puts it in a strong position against the likes of BSkyB. Given it also has a mobile solution, this should give it the ability to offer a strong quad play solution of voice, video, mobile and broadband - which means it may well be able to offer competitive competition to players such as NTL:Telewest (soon to be Virgin Media) and BT.

Consolidation is going to continue within the internet and telecoms industry, but it looks like Carphone Warehouse is going to be one of the survivors. ®

Assumptions

IPStream is BT Wholesale's broadband service that allows an ISP to offer IP connectivity to their backbone. There are two elements to the service, the end-user connection, which is a fixed cost per user, and the connection (or connections) "fat-pipe(s)" to the ISP which are based on bandwidth (known as capacity based charging).

A BT 622Mb/s "fat pipe" costs £124,000 per month with a set-up £150,000 fee. These are limited to 32,000 sessions (ie, simultaneous users). Though the ISP can decide their own contention, it's likely that each pipe will be max'ed out (CPW encourages the sale of a modem. When someone switches off their PC, the connection will go down and the session can be re-used).

Unfortunately, this means that if everyone is using their connection at full speed (or trying to), they'll only get a usable 20Kb/s or so of bandwidth. However, it's assumed CPW is loading them to the full, which means it needs 25 pipes for 800,000 customers. They'll also get about 1.5% per cent discount for the volume.

The cost of the ADSL connection is £8.40 rental and £12.40 set-up.

Wholesale Line Rental costs CPW £8.95 per month.

It can be assumed that the set-up charges are absorbed over 18 months (which is the length of the customer contract).

Therefore, the total costs per month are the ADSL and WLR costs (including the part of the set-up costs) and the split of the fat-pipes -which is where the £5.69 loss comes in.

There are various other costs that have been completely ignored, they include any ISP infrastructure (mail/web/etc.), transit costs (ie, the costs associated with getting traffic off CPW's network - with 25 fat-pipes, that means potentially 15Gb/s of bandwidth).

Revenue which has been ignored are things like phone call revenue. CPW must be making some money from people making calls to mobiles/premium rate numbers etc.