The migration from dongles to embedded modems will have serious ramifications

This week I've been wrestling with the thorny issue of devices. What will the mobile broadband devices market look like in 5 years time and what are the implications?

The USB modem totally dominates today (although reports of the demise of the datacard have been much exaggerated). We can safely assume that in 5 year's time everything will have gone embedded. There are a lot of implications. Here are just a smattering that spring to mind:

1. IT channels become more important in distributing broadband connectivity. MNOs can't compete with the IT channels in range of devices, so they'll end up selling more contracts through those stores. Also they'll be selling an increasing amount of SIM-only mobile broadband.
2. Home hubs, such as the T-Mobile Sharedock or the Huawei D100, will become obsolete. Why share your connectivity when all laptops have an embedded modem? To save on cost? Actually that won't be necessary thanks to shared plans, as launched by 3 Australia recently.
3. No need for handset-as-a-modem functionality. Why connect to the wide area network via bluetooth to you handset when your laptop has its own modem?
4. WiFi won't die out. It'll be a brave laptop manufacturer that first strips WiFi out of their product. Plus, it's perfect for in-home higher capacity connectivity.
5. Rapid upgrades to end-user equipment will be difficult, mitigating the need for escalation of the technology arms race. As we get towards market saturation with most users using an embedded modem there's less of a requirement for MNOs to upgrade their networks. Modem upgrades will follow the laptop upgrade cycle, i.e. about every 3 years, making it more difficult for existing subscribers to take advantage of higher network speeds.

And there are many more, but you'll have to wait for the release of Mobile broadband devices: from USB modems to where? for the rest of my conclusions on this fascinating area.

Does network sharing make sense in a mobile broadband world?

Today saw yet another twist in the ongoing network sharing saga in the UK. Speculation is rife that Vodafone will soon announce that it is tying up with O2, forgoing its existing deal with Orange, who are allegedly looking to join up with T-Mobile and 3's MBNL project.

We can speculate on the reasons for changing partnerships. Possibly O2 was conscious of being the only one at the party without a date and so realised that some serious flirting was necessary to ensure it didn't go home alone. Alternatively, maybe they're conscious of 900MHz refarming and are realigning to accommodate that. The issue for me is whether the logic of network sharing is destroyed by the growth of mobile broadband.

In a voice/SMS-oriented market, network sharing makes total sense. MNOs want to cut costs and they've not really competed on network coverage for years. One operator's 99% population coverage is much like another's. So it makes sense to pool resources and compete where they can differentiate, e.g. propositions, price, customer service etc. This is particularly true in mature markets where growth opportunities are limited and there is significant pressure on cost.

In a mobile broadband market this is turned on its head. The service is so commoditised that network coverage and capacity are critical elements in the value proposition and MNOs must seek to differentiate themselves. If they are all using the same network(s) this becomes impossible.

The question is: can any single operator build up their network to such an extent that they build an unassailable lead over their competitors? It seems unlikely. The competition is unlikely to give up the chase. The capacity arms race will continue. As a result it will be very difficult to actually develop an effective differentiator. This limits the appeal of keeping a stand-alone network, particularly as network sharing will bring significant benefits in network economics.

So the conclusion must be that even with the growth of mobile broadband, network sharing is a logical choice.

Lies, damn lies and mobile broadband surveys

Not one but two surveys were released today on who provides the best mobile broadband in the UK and from two respected sources. Do they agree? Of course not. Which? reckons Virgin Media's offer to be the best on the market, while YouGov puts T-Mobile top of the heap. There is at least some consensus in that they both run on T-Mobile's network. By my reckoning, that takes to five the number of UK operators who have the best offering according to independent testers. Back in December Orange gained the accolade acccording to YouGov (again), while in October a survey by Expert.co.uk put 3 top. Vodafone has to rely on a June 2008 survey conducted by itself...ahem...I mean independent wireless engineering company LCC International to put itself in top spot.

Of course it's all nonsense. Are we seeing regular radical overhauls of networks that would mean that network providers are overtaking each other in terms of network capacity? Of course not. So why the disparity? It's because there are so many variables associated with the user experience (distance from the base-station, contention ratio, height of the antenna, thickness of walls) of mobile broadband that generalisations about "the best network" are meaningless.

This won't stop MNOs from trumpeting their success in these polls of course, but in reality, user experience will bear little resemblance to the results of all these surveys. MNOs need to do two things.

Firstly they need to ensure that these polls continue to give out - at least - mixed messages. Ideally, of course, an MNO would want to be #1 in every poll. The bare minimum required is that they do not allow themselves to drop too far behind their competitors. Once there is consensus on 'the best' or 'the worst' network it will be a difficult barrier to overcome, even if there is no grounding in reality. A good reputation will allow MNOs to charge more, while those with a bad reputation will be forced to compete more aggressively on price. This requires that MNOs continue to invest in network upgrades.

Secondly they need to allow users to try before they buy. That way they can do the comparison that matters, i.e. how do the various offers match up in my front room, office, local pub etc. rather than depend on YouGov, Which? or whoever doing that research in a small random selection of locations.

TalkTalk customers want mobile broadband

I just spotted some news from the end of last week that UK ISP Talk Talk has issued an interest nugget from a recent user survey. Apparently 16% of their surveyed subscribers would be "likely" or "very likely" to take up mobile broadband. TalkTalk particularly specifies that users would be interested in doing so as a complement, but I'm not au fait with the actual wording of the question, so it's difficult to know exactly how they're qualifying the interest.

This is interesting for two reasons:

• Firstly that Carphone Warehouse is associating its TalkTalk brand with mobile broadband. We should expect an MBB complement to their TalkTalk DSL offering soon. That said, CPW's forays into the voice market as an MVNO/service provider have been largely unsuccessful. But it looks like they have realised that including a mobility element in DSL offers will become increasingly important.
• Secondly, Talk Talk customers are amongst the most price-sensitive in the market. Anyone prepared to go through this (and many peoples' experiences were equally terrible) for 'free' broadband must be. One-sixth of those users are apparently interested in mobile broadband. So if the service is affordable and priced right, it is definitely a mass-market phenomenon.

More capacity is a mixed blessing

I know it might sound like heresy, but could it be that more spectrum might do more harm than good to the nascent mobile broadband market?
In markets worldwide, huge amounts of additional bandwidth (freed up by activating more existing carriers, 900 MHz refarming, digital dividend and 2.6GHz spectrum auctions, HSPA/LTE upgrades) will soon be unleashed on waiting subscribers, with severe implications for supply, and thus demand.

Today demand exceeds supply. Obviously it's not an easy thing to prove, but I think it's reasonable to assume that I'm not the only one who finds himself fuming as his mobile broadband connection speed drops to a snail's-pace. This is indicative of substantial demand, without the capacity to meet it. Where demand exceeds supply there is little incentive for MNOs to engage in fierce price wars. They wouldn't be able to support the customers. So prices should stay artificially high.

However, if you put MNOs in a situation where they can support 3 or 4 times as many customers, won't they simply cut prices to attract them, ultimately compromising the profitability of their existing subscriber base. Experience from the voice market seems to suggest so: in many cases prices have been cut to little more than the marginal cost of providing the service. Making more bandwidth available to operators - and thus customers - will simply encourage MNOs to race for the bottom in terms of price.

OK, so I'm playing devil's advocate a little here. The addition of further bandwidth will improve the user experience and hopefully alleviate my fury. It also allows mobile to more closely replicate the broadband experience by enabling higher bandwidth applications such as over-the-top video. But, to be worth implementing, it must be possible for MNOs to monetise the increasing capacity and bandwidth. Giving users a better service must translate into additional revenue, either through more subscribers or higher ARPU. History suggests that it will be very difficult for MNOs to increase prices as they ramp up bandwidth. So far there has been an inverse relationship between price and capacity/speed.

It would be a very foolish MNO that did not pursue additional capacity given that competitors will be so doing. If more bandwidth means lower prices (e.g. through greater spectral efficiency), you don't want to be the MNO that can't compete. MNOs need to participate in the arms race, rather than risk being left behind.

Sharer plans will stimulate subscriber growth

Hopefully you will by now have seen the mobile broadband forecasts that I've put out in my capacity as Principal Analyst as Analysys Mason. In the forecasts I'm fairly bullish about the prospects for the growth in connections over the next 5 years. A number of factors will stimulate growth including increasing competition, faster networks and more appealing tariffs. I'm also assuming that MNOs will find increasingly innovative ways to spread connectivity out to the mass market, grabbing the opportunity presented by massively increasing laptop penetration and the virtual ubiquity of embedded modems in the future.

One example struck me last week. 3 in Australia (ever a testbed market for the rest of the mobile broadband world) launched a sharer plan for business, allowing corporate or SME accounts to share a bucket of data between multiple users. There's nothing new about sharer plans in voice and it's a natural extension to open this out to mobile broadband for businesses, and eventually families.

With group plans, mobile broadband becomes a more affordable option if individual usage is low. Today MNOs penalise and put off occasional users by requiring a minimum monthly spend or charging a higher per-GB fee. By bundling users into a group plan MNOs can provide novice users with the opportunity to experience and (hopefully) come to value mobile broadband, ultimately driving growth in usage and spend.

RIM and Nokia into the netbook market

Nokia seems to be plotting entry into the netbook space with a range of Symbian-based devices. RIM, meanwhile, has demonstrated more modest yet substantive ambitions in its plans announced so far. Rather than getting into the hardware market it is focusing on putting BlackBerry functionality onto devices, specifically Lenovo ThinkPads. It uses an Express Card to hook up via Bluetooth to the BlackBerry, synchronising mail to the laptop, even if the laptop is off. This is a fairly modest entry into the laptop space, but I suspect it is just a case of testing the water. Historically RIM's success has been tied to end-to-end management of the user experience. So a BlackBerry netbook is surely not far away.

The appeal of getting into the laptop space is obvious.

• The growth of PC mobility. Do I really need to go over the stellar growth of mobile broadband and the anticipated continuing growth? Not to readers of this blog surely?


• Device convergence looks to be occuring from both directions. So rather than just focusing on the increasingly smart handsets, it's natural to also focus on increasingly connectivity-oriented (dumb, arguably) PCs, in the form of netbooks. Will they ever meet in the middle? I doubt it. There is a fundamental shift between a mobile device and a portable device and you can't be optimised for both. Anything in the middle ends up being as successful as the Spork. Actually I may be being a little unfair to the spork, which graces countless prepacked salads and camping kits around the world.


• Keeping up with the Joneses. Other OSes look set to bridge the gap between handset and PC and provide a seamless user experience. Windows Mobile 6.5 is a huge leap forward for Microsoft in bringing Windows to the smartphone. It still demonstrates very little joined up thinking in terms of linking up the various platforms, but that will come with version 7 I'm sure. Similarly Google will doubtless be looking to transport Android onto PCs. In this environment, Symbian's lack of PC pedigree would stand it in poor shape.

Clearly some adaptation would be required to shift Symbian or BlackBerry OS into the PC space. In the case of BlackBerry the logical option would be to partner to run BlackBerry functionality on existing OSes on a BlackBerry branded device, rather than attempting the heavy lifting required to develop a PC version of BlackBerry OS. Nokia probably does have both the brand and the scale to make a success of the netbook.

EDIT: I forgot to mention in the initial post that I will be publishing a new report looking at mobile broadband devices "Mobile broadband devices: from USB modems to where?" in April.