To print: Click here or Select File and then Print from your browser's menu

This story was printed from silicon.com, located at http://www.silicon.com/

Story URL: http://comment.silicon.com/0,39024711,11027651,00.htm

Why unbundled broadband was never a goer
And oddly enough, it's not BT's fault...

By Ben King

Published: Monday 24 September 2001

Today is the day unbundled broadband was meant to start shipping in volume. Yet there are less than 150 unbundled lines in the whole country. As Ben King explains, unbundling was never going to work.

Today's deadline will pass without a blip on the radar screens of consumers thirsty for faster net access. Anyone hoping to buy broadband off someone other than BT will have to have a large chequebook, or prepare for disappointment.

Local loop unbundling, (LLU), once a key part of the blueprint for broadband Britain, has been largely written off for providing business customers with a niche product. And a quick glance at BT's telephone exchanges makes it clear what a long, difficult and expensive process unbundling is, and why it was never likely to make sense for consumers.

The local loop unbundling (LLU) process is often mistaken for DSL as a whole. Almost all of the 80,000 people who have DSL use BT's own DSL equipment, either provided by BTopenworld, or a reseller like AOL or Freeserve, who basically sell the same product with a different label.

'Unbundled' broadband, however, involves renting a berth about the size of two motorcycle parking spaces in a room rather like a rugby changing room in one of BT's local exchanges.

In this, an ISP has to put about £50,000 worth of its own DSL equipment which it attaches to BT's phone lines on one end, and the homes and offices of their customers on the other.

This investment means you can supply about 1,000 customers with DSL. The cheapest DSL for consumers costs about £30 per month, so that's a total of £30,000 per month a provider could expect to earn from the hostel.

Costings from one of the local loop entrants, Bulldog Communications, show it can get the cost down to a little over £10 per month per customer when the exchange has around 800 customers.

So, buy it for £10, sell for £30 - doesn't sound like a bad business. But after adding other costs like marketing, installation, customer support, churn and the fact that it will take a while to get the exchange full, it begins to look like a rather perilous exercise.

Even at £30 a month, it's almost impossible to even come close to filling all DSL capacity. BT has an average of 80 customers per exchange, and even in more mature markets like the US, exchanges are rarely more than half full.

At £30 a month, there's no way the figures can add up, and no one's really claiming they are. If home users want DSL at this price, they will have to buy it off BT or one of their resellers. For consumers, unbundling simply doesn't make economic sense.

However, £30 only buys a 512Kpbs connection, with no service level agreement. If it's down it's down. But businesses are already shelling out hundreds of pounds a month for a 2Mbps connection, with a service level agreement and various other bells and whistles.

The hope of the companies still in the local loop unbundling process is that they can fit DSL into some wider business service package which businesses will be willing to pay top dollar for.

As Nigel Pitcher, from DSL and network provider Fibernet, explained: "We're not providing a go-faster modem here. We're trying to take the premium voice services you can only get with a leased line down to the small and medium-sized business market."

Fibernet is selling 2Mbps lines, with quality of service agreements high enough to run traditional voice services on, at around 40 to 70 per cent of the cost of a leased line, which often runs to thousands of pounds per month.

It's not expecting to make a huge amount of cash on the DSL part of the service - the proposition only makes sense because they're selling the use of their national fibre network on top of it. Pitcher reckons the local loop will contribute about 20-30 per cent of the profit from a typical customer, with the 'middle mile' fibre network contributing the rest.

He reckons that BT won't be competing in this small niche for a good many years, but lower down the DSL food chain, for customers expecting lower data rates and lower service levels, no-one's really talking about selling unbundled DSL for consumers any more.

It's difficult to imagine that the LLU players would have a more compelling offering than the companies supplying DSL through BT's network, or using some other technology like fibre or even satellite.

Most press coverage (not least on silicon.com) has laid the blame at the door of BT for the failure of local loop unbundling to allow more than a small niche product, but in reality the picture is more complex than that.

While BT hasn't exactly bent over backwards to make LLU work, it can correctly claim that it has met all the deadlines set for it by Oftel and the EU.

BT and the regulators point the finger at the financial markets for the failure of LLU, and it's true that investor confident in telecoms has collapsed over the past year.

But fundamentally, the sheer complexity of the unbundling process and the merciless economics of the local loop business make it difficult to see how consumer LLU could ever have been anything other than a complete mess.

So, while the official BT line is "We're doing everything we can to make sure that LLU is a success," they don't sound like people who are expecting fireworks.

Nicholas Clegg, the European Parliament's Rapporteur on LLU, and one of the architects of the legislative processes that drove LLU through Brussels so fast, still maintains LLU broadband for consumer DSL is a viable option.

He said: "I don't see why a-priori it should be impossible for non-incumbents to deliver broadband over existing copper wires. Maybe in a couple of years we will see the fruits of our labours in the shape of more unbundled players."

For the moment, though, unbundling for consumers is a non-starter.