BT’s Mobile goes Inside-Out

BT’s Mobile goes Inside-Out

Amongst the acres of press coverage on the UK auction results very little attention was paid to BT re-entering the UK mobile network market. Its subsidiary Niche Spectrum Ventures picked up one of the paired 2.6 GHz channels that will be suitable for urban LTE (4G) coverage. This could be the start of something really significant happening in the UK mobile market. For a mere £12m per MHz of spectrum (or thereabouts) BT has bought itself some very interesting mobile options. But that is not the only fascinating story from the 4G auction to have largely passed under the radar…

The story of cellular radio from the beginning has been adding a new base station roughly half way between two existing base stations as traffic has built up (it is called cell splitting). In the early 1990’s a 900 MHz network operator might only need 1600 base stations. Today the number is more in the order of 16,000. Eventually the end destination (at least in dense urban areas) will be a super high density wireless networks comprising millions of very low powered base stations each providing coverage of 100m radius or less. This follows from the laws of physics. It is a story of first getting the reach (coverage) and then the long haul of driving-up the density (capacity).

Where BT is being pioneering is that they have started their high-speed mobile network (currently called FON) at the very end destination and are working back to build-up useful mobile coverage. It has been built on a partnership with their fixed broadband customers to share the capacity of the customer’s WiFi unit between private use and public use. The public use comes as some of this WiFi coverage spills onto the streets (inside to outside). The up-side is that it offers very high speed mobile access at a very low unit cost. The down-side is that WiFi will reach out of a home for only 50-100m. So the coverage story is very weak.

Here is where BT’s success at the 4G auction could be game changing. It jumps the coverage story up by an order of magnitude. Where WiFi coverage peters out at 100m (or less) a typical 2.6 GHz cell will provide 1km of coverage. It will not be long before the marketing wizards will be coining buzz words like Super-Hot-Spots or Mega-Fons. From there it is an open road to gradually covering cities and large towns as gaps between the 2.6 GHz cells get filled-in.

But for most mobile customers that is still not enough coverage. We may spend most of our time within 50 miles of where we live but we all take the occasional holidays and business or family visits to the rest of the country. It would make absolutely no economic sense for BT to try to use 2.6 GHz to create national mobile coverage. Fortunately BT already have that piece of the jigsaw in place by other means – they have an MVNO deal with Vodafone (See note 1 below) where they buy capacity wholesale from them on their national network and resell it to their own mobile customers. It also gives BT the benefits of Vodafone’s real-time home location register…customers can be found…which WiFi can never adequately do.

That is why BT’s new spectrum from the 4G auction has the potential to be game changing – they now have so many options to attack mobile data capacity, much higher data speeds and lower unit costs. At one end BT’s customers are supporting the inside WiFi coverage spilling out to the public streets absolutely free of charge.

If anyone wants to compete with BT’s approach say by hanging a low power mobile base station on a local lamp post to do much the same street level coverage job…they are likely to find (to most people’s surprise) that many local authorities have outsourced the lamp post provisioning to private company who is likely to demand the usual sky-high site rents. Add the cost of back-haul…and that route is effectively choked off.

At the other end of the story the mobile coverage of up to the final 30% of the population loses money for the national mobile operators. That is not BT’s problem…they only pay for what they use. Somewhere in the middle BT’s mobile customers are leaking their traffic onto the Vodafone network in varying degrees. BT now have the option to keep dropping in 2.6 GHz LTE cells where that leakage is at its greatest to keep driving down their payments to Vodafone.

Setting out the possibilities does not mean this is what BT will actually do…but somebody on the strategy side of BT has done an outstanding job…BT is in a position to drive significant change in the UK mobile market. Their mobile strategy is being under-rated by analysts and that is probably because the impact will be minimal in the short term. It’s a long term play.

BT is not the only under-reported story from the 4G auction. The second was O2 picking up the 800 MHz licence that has the impressive national coverage obligation. They paid around £55m/MHz for it. The price that Everything Everywhere and H3G paid per MHz was not only very similar(why so high as their spectrum had no coverage obligation?) but the sting in the tail for them is that they only managed to come away with 5 MHz spectrum each. But 4G (LTE) in only a 5 MHz wide channel is a performance disaster. It does not deliver anything better than today’s 3G (HSPA).

This leaves O2 in a strong position with an exceedingly well advanced 900 MHz HSPA network…which they will now re-enforce with national 4G (LTE) network on the even more favourable 800 MHz spectrum. This will allow them in the future to take serious cost out of serving the last 30% of the population with a very high-speed mobile network (and do a fair job serving those inside buildings everywhere else). Trying to do the same job at 2.1 GHz (which H3G is trying to do) or at 1800 MHz (which EE’s is doing currently for 4G) is much more costly…and come the inevitable crunch…the most rational option to cut back on.

Nobody is currently positioned to over-take O2 on such a cost effective 4G national story. Vodafone was the only other mobile operator to come away with the minimum 10 MHz wide channel at 800 MHz necessary for 4G. But they have a site sharing agreement with O2 – so whilst they may pace O2 there is no incentive to outpace them. Therefore unless and until Everything Everywhere and H3G pool their two lots of 5 MHz into a single 10 MHz wide channel at 800 MHz…O2 (and Vodafone) finds themselves on route to a competitive advantage in providing high-speed mobile coverage in many rural and semi-rural areas.

What is poetic about this story is that BT used to own O2. It is a nice touch that the two entities now find themselves leading from the two polar positions of tomorrow’s mobile network development paths – the outside-in (cell splitting to build capacity) and inside-out (cell aggregation to build coverage)…and eventually cross-over somewhere along the way.

A final bit of the untold 4G auction story emerges in the prices paid for different spectrum at the UK 4G auction versus the prices paid for the identical spectrum in the Germany 4G auction. My rough calculations (See note 2 below) are that 2.6 GHz FDD spectrum costs around £12.4m/MHz in the UK but only £3.15 m/MHz in Germany. The 800 MHz digital dividend spectrum went for £102m/MHz in Germany but only £55m/MHz in the UK . So the UK auction appears to have driven the UK price for 800 MHz digital dividend spectrum to well below the German level but driven the 2.6 GHz spectrum considerably higher…a quite unexpected twist.

Since the 2.6 GHz spectrum prices went so high it makes it all the more puzzling why the more desirable digital dividend spectrum at 800 MHz appears less valuable in the UK than Germany…even when the size of population is taken into account.

 The most likely explanation is the way Ofcom divided some of the digital spectrum into 5 MHz packages instead of aligning with the LTE (4G) standard of 10 MHz packages at 800 MHz. BT’s sole interest was 2.6 GHz, so if the other new entrants are dismissed as non-starters  there were effectively less bidders than packages on offer…no wonder the Treasury were disappointed. 


Note 1: Since this blog was written BT have switched their MVNO deal to EE

Note 2: most published international comparisons will adjust for the difference in population size but that will not affect the point that the auction designs in Germany and the UK appear to have driven prices for 800 MHz and 2.6 GHz in quite different directions.


Comments are closed.