Simon Rockman, Contributor
The Spectrum Policy Forum has engaged mobile business experts Real Wireless to look at the effects of switching off 2G. It sounds like a no-brainer. When was the last time you bought a 2G phone? Even dodgy dealers using burner phones will pick up a thirty quid Alcatel Android.
Typical phone replacement rates are something like every twenty months, so killing off something which saw it’s heydays twenty years ago really shouldn’t affect anyone should it? It’s time for 2G mobile phones to go the way of the typewriter and cathode ray tube.
What’s more the frequencies associated with 2G are fabulously valuable. When cellphones started in 1985 they were given the radio frequency of 900 MHz. At the time this was considered super high, at the edge of what was useable. As technology has improved we’ve moved to thirty times that and more[ blu wireless link], but the advantage of lower frequencies is that they travel further. Even the 2G frequency of 1800 MHz which was launched with One 2 One (now part of EE) and Orange (ditto) is much lower and longer travelling than the 2300 Mhz spectrum used for 3G and 4G and the 3.4 and 5 GHz “mid band” used for 5G. The further the radio waves travel the fewer cell towers a network needs to build and so there is a huge cost saving in using 900MHz as opposed to 3.4GHz.
You can understand why the networks want to shut down 2G and use the radio spectrum for 4G and 5G.
By changing technology there are other advantages for the network. The kind of processor chips which go into a 2G phone date back to the mid 1990s. The radio systems in the phones don’t compress the anything like as much as the newer technologies, so the cost per minute or megabyte for the network to support a 2G user is far higher than for a 3G, 4G or 5G user.
Not so fast
It all points towards wanting to close down the 2G network as soon as possible and use the radio spectrum for something newer.
There is of course a problem. Quite a lot of people and things are using 2G. It’s widely discussed that most of the connected water meters in the UK use 2G. Switching off the connectivity to those will be a problem, but at least we know where they all are – which cell sites they communicate with. A more pressing issue is the devices which are mobile. Notably vehicles. Lots of automotive applications use 2G. There is 2G navigation, fleet management and stolen vehicle tracking.
While a phone might have a life of twenty months, a car will last decades. This is set to come down as technology reduces the life of cars. If an engine control unit for a shed dies the car is going on the scrap heap. There are however a lot of cars which when they were built ten years ago have at least another 15 years life in them and which use 2G for their navigation systems. It’s not just navigation. The Nokia 6310 has a great eBay market because it’s the standard fit phone for 1990s Mercedes. Switching off 2G will upset a lot of older car users.
The fleet management problem is bigger. Thousands of vans and lorries use a system where their business dispatches jobs to them over 2G and then tracks where they are using 2G. The technology was chosen because it’s cheap. Most of the essential patents for 2G are cheap as chips, while 3G – where the patent holder is significantly more expensive and 4G, where a phone will have nigh on 4,000 families of patents required, is very much more expensive. This is reflected in the difference between a 2G module costing less than $3, while a 4G module costs more than $15. To upgrade all the vans and lorries using 2G to 4G would be fabulously expensive for the fleet management companies. One person working in the industry told me it was his biggest nightmare. When the US closed down 2G at short notice, giving only three years’ warning, several telematics companies went broke. Deciding which customers to upgrade and which to lose would be tough on a telematics vendor.
Equally bad is the stolen vehicle tracking. These are devices which ping the network from time to listen for text message, when they receive one which says “You’ve been stolen” they repeatedly send their location back to the tracking company.
None of these vehicle applications make much money for the networks. They wouldn’t mind much if they were switched off. But the chap with the retro Mercedes or stolen BMW might.
It’s very hard to find figures for the number of 2G devices used in telematics, but a good finger-in-the-air guess is more than a million but less than two million.
And there are still handset sales. The company which tracks who is buying what is GfK because I approached it for data on 2G and it was quite parsimonious in what it was prepared to tell me. This might be because I wasn’t paying or because all the money is in tracking smartphones and it just doesn’t have the data on the dregs of the market. I suspect a bit of each. The figure GfK was prepared to give me was that 6% of the phone and smartphone sales. Allowing for the nature of the market it’s probably fair to assume that GfK undercooks the numbers and 10% is more reasonable. A different way to express it is that Ofcom’s 2018 state of the market report gives a figure of 78% of the population having a smartphone. So 10% of people using a non-smartphone, the figure of 10% looks low. The vast majority of non-smartphones will be 2G so that’s another few million 2G users.
But who are the people who are still using 2G feature phones? Some are second phones, some are people just visiting the UK and have picked up a local SIM and cheap handset, but most are older people.
The market for phones for senior users is one which tends to go without recognition. Numbers are hard to get hold of but I’ve heard figures of 800,000 a year from one manufacturer, and of 300 a week from a reseller for another. The two biggest suppliers in the UK of phones for seniors are Doro and TTFone. The mainstream device Doro sells is the 6520 , which is a 3G phone but it still sells plenty of its it’s more recent 6030 2G model. TTFone doesn’t currently have any 3G phones it its portfolio. And there is a long list of other companies whose products are sold through organisations like the RNIB, Age UK and in the back of gardening magazines. Audioline, Amplicom, Binatone, Beafon. I could go on through an alphabet of companies you probably have never heard of.
And the users are loyal. Not to the manufacturer but to the individual device. Doro hasn’t sold the 410 for the best part of a decade, but you can be sure that if a customer throws one in the washing by mistake, he or she will want, not a shiny new 6520, but another 410. These customers don’t upgrade. Ever. A sensible timescale for holding onto a phone is 84 months, seven years. They are vulnerable users. What’s worse is the way they use the devices. It’s all about “just in case”. Having a phone is very important to them and their families. The phone is owned only to be used in case of an emergency. Many have a button on the back which acts as a speed-dial to a cluster of friends and family who can be called in case they fall, the car breaks down or or their bird feeder is empty.
With hundreds of thousands of phones a year being sold into a market which keeps the devices for a long time there is a substantial installed base of users but finding out how many is hard. I’m probably the country’s leading expert on this and I don’t know.
How many do you say?
We could ask the mobile networks – Vodafone, EE and O2. They could look at the electronic serial number, the IMEI, of every phone which made a call, look it up and see how many 2G phones connect to the network. Unfortunately this has two huge problems.
The first is that we cannot trust the networks. Remember the people using the most valuable spectrum are the least valuable customers; someone spending £10 every six months, isn’t a prime candidate for customer retention. It’s akin to a landlord who has a tennant in a valuable property paying a peppercorn rent. So if we were to ask the network it’s likely there would only be a quick scan of the usage rather than heavy data mining, and that would massively under-read.
The second problem is that the networks don’t really know. Many older people will only switch their phone on in an emergency. They leave it off “to save the battery”. The network never hears from them.
While it’s easy to work out where all the water and gas meters are, and it’s possible to locate the fleet tracking and navigation users it’s ironically hard to find the stolen vehicles – unless we send them all a “you’ve been stolen” text, and next to impossible to contact the senior users. You can’t even use online communications.
What we do know is that there are millions of devices out there which only use their 2G connection when they really need it.
Sometime, Always, Never
As the mobile network operators scrabble to move new technologies to the old frequencies there are three ways that could be adopted to manage the problem.
The first option is a long lead time. With long automotive replacement cycles and older people hanging onto their devices it would need new devices to be developed and all the 2G apparatus in the market to sell through well ahead of the end of the technology. Allowing three years for the first part of the process and seven for the second we’d need a decade more of 2G. Look for a 2030 farewell.
The second approach is to believe that causing the problem will make it wash through. Announce a five year deadline for a total switch-off and see what bubbles to the surface. This is probably what the networks want, but it’s likely that there will be some older users who only find out about this when they press their emergency button and fail to make contact. It’s quite conceivable that people will die as a result.
There are two options for the third “never” approach. They both leave some element of 2G running for the foreseeable future. All of the scenarios in which 2G is used have someone in common: a need for a lot of coverage and very little capacity. None of the networks want to use their valuable radio spectrum for 2G so they could band together and only leave a tiny veneer of 2G running across the whole country. All the networks sharing just enough radio spectrum to handle a few simultaneous calls or the minimum of a GPRS data service. Just one time divided channel might do it. Of course you’d need four channels, based on the four colour rule, nationally but that’s a tiny amount of spectrum and is the kind of co-operation the networks have been keen to indulge in with the new rural coverage schemes and the plans for 26 GHz sharing. This has been proposed in other countries and is probably the solution which best suits users. The engineering twist on this would be to build some custom hardware. In the areas where there are a handful of die-hard 2G users – such as a retirement home – it’s possible that you could put in a 2G to 4G bridge. Something like MiFi which allowed the users to carry on with 2G phones and then employed 4G as backhaul. Building such a device in small volumes would be very expensive but it would be worthwhile for the operators to assess the market need, particularly if they valued it by the spectrum which was freed up rather than the need to retain the customer.
The Real Wireless report on shutting down 5G comes out in October. It will make interesting reading.
Forbes Simon Rockman
Real Wireless Real Wireless