In Today’s Issue: Some phone or other launched; “ZZZPhone” debunked; Verizon ODI dip stick; Launchcast “Dashboard” open to hackers, in a good way; Verizon - dangerously interesting?; Sprint pushes push-to-talk; iPhone Truphone; NTT DoCoMo on the unwise monster acquisition trail again; unwitting private equity fund sups with Richard Li, helps 3UK double its customer base; pass the separator, Mme Reding; Comcast in trouble with the FCC; open search at Yahoo!; the coming mobile data boom?
Apparently the 3G version of some device or other launched today….but beyond such trivia, there were far more interesting things going on in the industry. For a start, wouldn’t you like to design your own phone? It’s a cool idea, but you’re probably best starting with an OpenMoko; it turns out that the devices are actually a job lot of old ZTE stock and the orders tend not to be fulfilled.
Here’s something more, ah, fulfilling: the first Verizon ODI device is out! And it’s something genuinely interesting - a module that fits into a tank of liquid and reports the level of the contents by text message. That’s actually useful, and could actually make money for operators and users, which is why nobody’s going to promote it very much.
Meanwhile, Verizon Wireless shipped the first gadgets with its “Dashboard” portal on board, aka Adobe Launchcast. More interestingly, details of it are being published so developers can make things to fit in with it; is Verizon gradually turning itself into one of the most interesting telcos around? FTTH, P4P, ODI, and this…which’ll come in handy, as they surrender to the falling price of SMS.
Quietly, Sprint gets some work in on its core voice & messaging products, rolling out its push-to-talk service to 47 new markets around the US. On the other hand, though; Truphone launches an implementation of its SIP client for the iPhone, but you can’t use it over-the-top on the 3G data network, because Apple doesn’t want its users burning the bundled data connectivity competing with the operators who both buy and subsidise the Jesus Phone.
Mike Elgan of Computerworld, meanwhile, asks the right question. Mobile phones have improved beyond measure in the last ten years - but what about calls? Hear, hear….literally.
The first rule of telecoms investment: don’t buy a Japanese network operator. The second: don’t buy a Japanese network operator’s good idea for use elsewhere. The third: if you are Japanese, don’t buy a foreign network operator. It is with a heavy heart that we read that NTT DoCoMo wants to buy stakes in foreign telcos again; oh dear.
Another piece of good advice is not to become a minority shareholder in anything Hutchison-related; there have been so many asset trades inside the empire that result in one-off gains for them and dilution for the others. Here go some private equity funds, looking at a 45% share of HKT. After all, 3UK will be wanting some capital if it’s going to double its business by 2012.
In the fixed world, meanwhile, they are about to feel the force of Viviane Reding for a change; so far the mobile operators have been the prime target, but now the European Commission is looking to push functional or structural separation across Europe. It might hurt at first, but it’s for your own good…
In other regulatory news, the FCC isn’t happy about Comcast and their Chinese Firewall-style use of spoof TCP RST messages to spork BitTorrent users.
They’re trying to buy Yahoo! again - that’s Microsoft and Carl Icahn, who’s apparently temporarily had enough of kicking Motorola around the pub car park like a sack of waste. Perhaps this is why? Yahoo! is introducing a platform to let third parties build their own specialised search service using Yahoo! APIs….which sounds cool.
And Informa crystal-ball merchants reckon we’re going to see mobile data volume race past voice in 2011.
In a previous article we provided an introduction to what we believe is the template for future growth in telecoms: two-sided markets. Having got the basic facts laid out, now we can take a closer look at some of the consequences of moving from one-sided to two-sided markets.
Two-sided markets in a nutshell
A brief reminder of what we’re talking about. In a one-sided market, merchants buy in equipment and services, taking on inventory risk. They combine them in some value-adding way, and sell the result on to end users (or other intermediaries in a value chain). The suppliers and customer do not interact directly. Most of telecoms works within a one-sided model today.
In a two-sided market, the middleman facilitates two groups on either side to interact with each other via some platform. This lowers transaction costs and builds scale. Critically, the price structure of using the platform is set to encourage participation from the most price-sensitive side, maximising platform revenues overall, rather than separately for the two groups. For example a newspaper typically charges a cover price well below that which would maximise reader revenue alone, because it needs a big audience to attract advertisers.
A good example of a company that moved from a one-sided to two-sided model is online bookie Betfair. Originally they only offered their own set of bets and odds online. They then allowed other bookies to offer bets in competition with one another, at which point Betfair’s business took off very rapidly. (Another in-depth example on job sites is here.)
In telcoland, i-mode is an example of a two-sided market, joining application developers to users. Our hypothesis is that operators need to focus on developing capabilities and services that facilitate a much wider range of business processes than content retailing. In each case, ‘upstream’ parties wish to interact with the telecommunicating public in some way, and the telco takes friction out of this process. The data by-products of the current triple/quad play products are key enablers, along with assets in the ‘edge’ devices (handsets, home hubs, set top boxes, smart meters, etc).
Elephants are born big
The first observation we have about two-sided markets is that they always need scale. That means you need to kick-start the market in some way to make it attractive to your initial customers and users. Typically some kind of trend-setters or marquee users are used to pump prime the platform. For example, an upmarket shopping centre looking to attract both retailers and shoppers will want a John Lewis or Nordstrom as anchor tenant. Every music platform will wants the Universal Music Group catalogue. In telcos we see this effect with peering and interconnect sites, with large anchor tenants attracting the smaller players.
This leads us to conclude that telcos are better advised to try building new two-sided revenue streams off their existing core voice, messaging and broadband businesses. This contrasts with the current approach of building whole new propositions, particularly around entertainment media, from a base of zero. Ultimately a collection of transaction platforms — for advertising, payments, and customer service — will be converging from multiple industries, such as online search, e-commerce sites or banking. Better to compete off a strong base when engaging such powerful rivals.
Gasoline, girls and guys
It would be nice to think that users will appreciate your wonderful new entertainment products and gladly surrender some more money for value-added services. Sadly, they seem to have a budget in mind that they want to stick to, and you end up dissipating a lot of your profit in marketing expenses.
A more compelling proposition is through creating efficiency and cost savings in a broader range of industries. In particular, saving labour costs or energy are obvious targets. It is possible to create one-sided solutions for specific verticals, e.g. a fleet tracking solution using cellular location. However, we feel that a two-sided market offers a more defensible opportunity, which means business process services (e.g. advert targeting) that involve interacting with the mass retail customer base in some way.
This is not to disregard the traditional focus of telco platform efforts, which is to optimise the supply chain of purely digital applications and content to users. This is necessary, but ignores the wider opportunity, and the consequent chance of spreading the risk and cost of building the platform across a much larger range of revenue sources. It’s the “analogue” world of people, trucks and raw resources where most of the economy still lives.
What do you know about the customer?
The ‘upstream’ party that wants to interact with the telco customer will have some data on that customer and their own relationship. However, this is likely to be significantly different from what the telco knows about the customer. The data assets of the telco, and the permission to use it derived from the customer relationship, are every bit as critical to optimising business processes as the ability to transmit data over distances via networks.
Where this data is most valuable is the ‘real time’ intelligence the telco can provide. Every time UPS delivers a yellow sticky ‘sorry you were out’ notice, instead of a parcel, resources are burnt to no productive end. The telco’s role is to use customer information/data — such as whether your mobile is associated with your home femtocell, if you’re in the middle of a call, or are roaming abroad — to help time interactions and facilitate transactions.
This is a common feature across two-sided markets: the more the middleman can help personalise the interaction, the more the platform can charge for its services. Operators need to reconsider how they manage such data assets, gather user consent, and extract value from them.
No bronze medals
Another property of two-sided markets is that they tend to follow ‘winner-takes-all’ economics, with a small number of dominant platforms: Windows and Mac OS; Visa, Mastercard and Amex; or Google and Yahoo!. In general, individual telcos will struggle to achieve scale in two-sided markets. The implication is that they will need to co-operate to either create clearing houses or hubs themselves, or work through existing aggregators or transaction networks (e.g. mBlox).
The goal for the platform must therefore be either to aim for general monopoly (in much the way the PSTN/PLMN voice network is the platform for all personal communications), or differentiate to dominate a niche.
The danger is that the telco is enveloped by other transaction and commerce platforms. A naive approach to exposing location, presence or other data could not only leave value on the table, but even worse have the telcos subservient to a small number of powerful external intermediaries.
A regulatory misfit
In competitive industries, and over the long run, prices reflect underlying costs. Competition will weed out inefficient players, making costs and prices converge.
Two-sided markets don’t work this way, and this can cause serious regulatory and competition issues, as one-sided rules are applied to two-sided markets. The heart of the problem is the way the platform subsidises one side to get them ‘on board’ in order to more than make it up on the other side.
The classic case that’s made its way through the courts in the EU’s anti-trust action against interchange fees on the Mastercard network. What happens is that when you walk into a store and pay a merchant with your your credit card, the merchant’s bank has to pay a fee to the bank that issued you the card. This fee doesn’t reflect any specific operational cost being incurred. The contention of the EU was that this is anti-competitive and welfare-destroying. The reality is that it is necessary to ‘bribe’ the public with zero-fee cards, low introductory rates, cashback rewards etc. to adopt credit cards and use them. Without mass adoption, the cost of the interchange fee becomes moot, since merchants lose sales (as customers don’t carry enough cash with them, or need credit the merchant doesn’t offer), and costs rise (as you have to handle lots of cash).
We see the same dynamics in telecoms, with roaming and termination fees. These exhibit some of the properties of two-sided markets. Imagine there are two groups, businesswomen (who travel internationally) and househusbands (who don’t). The high price of roaming could be seen as a ‘tax’ on the price-insensitive businesswomen, which can be used to subsidise pre-paid service for the househusbands, driving mutual benefit of service adoption. Likewise, calling-party-pays and high termination fees in Europe have driven adoption far higher than in the superficially more ‘equitable’ North American model.
Competition law concerns
These competition and regulatory issues will be a major battleground. You can charge more than cost, and it’s not necessarily a sign of lack of platform competition. You can charge below cost, and it’s not necessarily predatory. (You can read more on this in this academic paper.)
Indeed, even having competing platforms may end up dividing the market into two sub-scale platforms that ends up destroying value to all users. Whether platform predator or prey, it is common for a dominant platform to reach such scale that it starts to cause wider competition concern. We see this with Microsoft and Google, for example. But there may be no public welfare gain from breaking these platforms up or constraining the scope of their activities. Standard competition law doesn’t apply here.
Where competition law does become more of an issue is where a successful platform adds on features and leverages its existing base to move into other areas. We’ve seen this with Microsoft (Windows plus browsers, media players and security suites), and it’s an ongoing issue with companies like the BBC using distribution of public service broadcasting content to enter the pay-per-view market, much to the objection of commercial rival Sky.
Not a binary choice
You can have elements of both one- and two-sided business models, simultaneously. Wal-Mart acts as a retail platform (two-sided) as well as a traditional merchant that buys inventory at wholesale top resell (one-sided). The two-sided mode is favoured when the middleman can either eliminate uncertainty and risk for the upstream parties wishing to interact with the users, or has distribution economies of scale. The driver is a balance between inventory and risk of the one-sided merchant mode vs. the cost of affiliation to and use of the platform.
We see this in Blyk, which gives away a limited amount of free calling in return for receiving targeted adverts, but continues to charge for higher usage. Telco business models will be complex hybrids for the foreseeable future.
A pricing challenge
The pricing of platform services can be a tricky subject. The obvious question is how much to shift cost away from the price-sensitive side. You can’t charge less than zero, but you can start to give away products and services (e.g. Google’s search, mail, and other content sites). Where do you stop? Could all those service delivery platforms end up being used not to create new consumer value-added services, but instead to create the ‘give away’ services you can trade for customer data and permission?
There may also be a need for ‘signalling’ of commitment by one side or the other. Job sites for high-end positions can filter out under-qualified job applicants by charging to become a member. Normal dating sites typically charge both sexes, but adulterous men have to pay a huge premium to signal intent to wayward women. It remains to be seen how these kinds of exceptions might play out in identity, payment and content services offered by telcos, but collecting cash from people merely so they can signal intent sounds like a profitable proposition.
Finally, where you have competing platforms they typically need to follow opposite pricing models. So platform A charges end users and gives away service to the upstream parties, and platform B does the opposite. This lets them co-exist in two distinct segments, rather than over-dividing the pie.
A very different kind of business
As we’ve seen, two-sided markets require a very different way of thinking about telcos and their role in the economy. They have very different economics to most of today’s telco products, and break the assumptions behind today’s regulatory and competition rules. They require new skills, partners, channels and organisation structures. However, they also offer an escape route from the problems of today’s one-sided telco business model, and since they principally re-sell data (packaged in special ways), they don’t require masses of capex to implement and are often highly profitable.
For more information on the opportunity for telcos to build two-sided market business models, see our report The 2-Sided Telecoms Market Opportunity
In Today’s Issue: OFCOM moves towards BT’s line on fibre; £31,000 phone bill; GTalk for mobile, where’s the talk?; iPhone bank run; pity about the OS security patches, though; cross-platform widgets; Files On Ovi; MTN-Reliance rift; EU offers more 2.6GHz TDD, WiMAX Forum delighted, Ericsson furious; RIM shares dive after profits double; how do you value mobile ads?; GSMA’s funny figures
Big news: OFCOM director Ed Richards spoke to the UK IT trade association, Intellect, in terms that suggest he’s leaning towards offering BT concessions on regulatory pricing in exchange for deployment of fibre in Openreach’s access network. This is an instance of the two-level bargaining process we described here; it looks like BT is still succeeding in monopolising influence on the regulator, but the next step will be to see how this can be made compatible with the existence of a competitive ISP/Altnet market in the UK. Details are to be published in September. Here’s a telling quote:
One thing is certain: the government is very keen that taxpayers don’t shell out to make Britain’s internet infrastructure competitive with more advanced networks in countries such as South Korea and France.
The horror….the horror…
Whatever OFCOM concedes to BT, it surely won’t be anything like some of the pricing people still encounter for mobile data roaming. Marvel at the £31,000 phone bill, to say nothing of the “IT consultant” who downloaded a whole episode of Prison Break on a roaming UMTS link and was surprised when he got a huge bill…
Meanwhile, here’s another case of Voice & Messaging 2.0. Google has developed a browser-based Google Talk implementation for iPhones, which looks likely to dig into iPhoners’ SMS traffic. However, despite the fact that GTalk has basic voice capability and uses standard protocols like XMPP, there’s no voice yet; one wonders what the reason for this is.
O2 UK’s…err…keen iPhone pricing caused the carrier to run out of stock when the 3G Jesus Phone went on sale. No surprise there: consumer surveys in the US show that 55% of those currently after a smartphone want a fruity one.
How many of them realise that the iPhone’s OS is several months behind MacOS X on its security patches? Could be a PR accident waiting to happen.
In a sense, this gets less and less significant; here’s Access offering a widget interface for Windows Mobile phones, as well as its own Linux system. Back at MWC this year, we noted that the LiMo community was already running various Symbian things on top of their Linux system. As Adobe and others develop universal clients, will anyone care what OS is underneath?
Meanwhile, Nokia launches a “GoToMyPC” clone as part of Ovi. Well, that’s nice, but will they let you put scripts or other executables there? Mobility implies you’re not necessarily on line, so one of the major barriers to mobile/desktop integration is the need for some sort of persistent presence on the Net that can act like a POP3 mail server, and transfer your stuff to wherever you are (and not just view a screen remotely). A solution for this with activation energy low enough for everyday users is badly needed.
The MTN-Reliance deal is falling apart, due to the dispute between the brothers who own the founding stake in Reliance Industries. Stand by for more Vodafone rumours.
Two worlds collide; the European Commission, with the assistance of the usual spectrum allocation alphabet soup, has changed the rules regarding the 2.5-2.69 GHz band in Europe to give more space to time-division duplex radio technologies like WiMAX, UMTS TDD, and perhaps LTE. The WiMAX Forum is delighted (and probably OFCOM, which is keen), Ericsson, as befits a vendor focused ruthlessly on UMTS, is furious.
RIM shareholders gave them the bird, after the company only doubled its profits; apparently no-one believes in their new consumer-focused strategy. Neither do we, really; they could be improving their voice features and integration with communications-enabled business processes, rather than trying to get into the ‘shiny’ market just as Sony Ericsson finds it’s drying up.
Open Gardens has some interesting thoughts about how to value mobile advertising opportunities. As is traditional over there, they manage to work IMS in somehow. Not so sure about that…
And finally, the GSMA tells off the EU over regulation. Apparently mobile operator CAPEX has fallen from 13 per cent of revenue in 2005 to 11 per cent today. Well, that’s nearly a whole 0.7% a year!
More seriously, could this possibly have something to do with the fact that voice revenue, coverage, and penetration were still rising in major Western European markets in 2005, and now they’re not simply because there are only so many phones one person can use, there is only so much land to cover, and voice is competing with people who offer it FREE? Perish the thought.
The big news in mobile this week is that Nokia has bought Symbian, the mobile operating system provider, so that it can give it away. In the first article on this news we looked at the deal from the view of the shareholders and competitive threats. In this second article we take an in-depth analysis of the nuts and bolts of software licensing and governance, to see if Symbian really lives up to its ‘open’ headline.
Symbian: open in parts
With operators being pushed towards ‘open’, what can they learn from Nokia and Symbian’s approach?
A ship with a small hole still sinks
Like it or not, technocratic choices of licensing schemes really matter. The plan is to release the entirety of the Symbian OS and S60, UIQ, and MOAP under the Eclipse Public Licence by the end of 2009. What is an Eclipse Public Licence? Well, it’s a development of IBM’s Common Public Licence, originating when IBM Canada released the software development environment known as Eclipse. It differs from the purist open source GNU General Public License in that it doesn’t require work released under it to be purely open-source. For example: you can build something containing both Eclipse-licensed code and code you wish to keep as your intellectual property, and release the whole thing under Eclipse.
This implies that you can’t include anything licenced under EPL in a “more open-source” project — the GPL bars you from imposing restrictions on the user beyond the requirement to maintain the GPL status of the work, so the fact Eclipse licensing permits IPR restrictions would preclude it. Similarly, as you cannot legally assert restrictions on GPL code, you can’t include it in an EPL project. You must choose.
So, this is somewhat less open than Google’s rival OS, Android, whose guts are subject to the Apache licence. It is also significantly less open than the mobile Linux OSes LiMo and OpenMoko. Both of these are GPL, and the latter is open enough to satisfy the most rigorous free software fundamentalist. It’s glatt kosher software. On the other hand, though, it’s a big step forward in terms of openness from most vendor OS so far.
Even if only a very small part of the code is tainted with restrictions, it is likely to be a new and important feature with IPR held by some third party. As Linux users know to their cost, it doesn’t matter if every other part of your PC is working if you can’t get a driver for your particular network or graphics card.
Who is really steering the ship?
Software licensing and IP law, despite or even because they are the sexy issues here, are far from the whole story. There are a lot of other questions. Governance of the project is one, and another is the status of the developer infrastructure around it. The Nokia announcement allows us to know quite a lot about the organisation structure…
The members of the foundation’s Board of Directors will also have seats in each council, with additional seats in the councils available for other foundation members. The foundation will operate as a meritocracy [sic]. Device manufacturers will be eligible for seats based on number of Symbian Foundation platform-based devices shipped, with the other board members selected by election and contribution. The initial board members will be AT&T, LG, Motorola, Nokia, NTT DOCOMO, Samsung Electronics, Sony Ericsson, STMicroelectronics, Texas Instruments and Vodafone.
That’s certainly an interesting definition of ‘meritocracy’, one you could easily mistake for ‘plutocracy’. Note the detail that, as well as shipments, you can get on the board by “contribution” - this seems to mean that the more lines of code your developers check in, the better you are. (All the traditional objections to measuring code value by the line are well and truly in effect here.) And what are Moto doing in there, as a company that has a proprietary OS, some Windows Mobile products, and a major role in LiMo — but is about to sell its handsets division and cut half its R&D staff?
It’s also worth noting that two pureplay chip makers are involved; perhaps their low-level microcode is the proprietary treasure the Eclipse licence is meant to protect.
Who is navigating the ship?
So what will this big carrier/big vendor/big software house club be responsible for?
The foundation will be responsible for managing the software roadmap and releasing the software platform, with the source code available to all foundation members. The development of the platform will be the responsibility of the foundation members, with the foundation coordinating development projects and managing the master code line.
So, not that open; if you’re not on the board, it doesn’t look like you’ll have much input, even if you’re an important stakeholder, like the operator who has to subsidise these products. And does the provision regarding the source code being “available to all members” only last until the 2009 release to Eclipse? Or will there be chunks that the Foundation keeps to itself? It’s been pointed out that most Symbian Foundation people will be Nokia employees, but this doesn’t worry us much — OpenOffice, various Linux distributions, and Sun’s Java are all maintained to a large degree by Sun, Novell, and IBM coders.
Keeping the ship sea-worthy
A huge issue in all open-source projects is the status of the developer infrastructure. Things like standardisation working groups, the process through which new software releases and updates are handled, preparation of things like software installation packages and Windows installers, tech support, and the tool chain — software development kits, documentation, IDEs, debugging tools and compilers — often define their culture and success as much as the headline stuff about GPL & Co. In fact, a large part of the very first Free Software Foundation project, GNU, was the development of free utilities in order to make GNU independent of proprietary software throughout.
The Platform will be completely free and open to developers whether enthusiast, web designer, professional developer or service provider. Of course, membership is not required to develop services and applications on the platform. The Symbian Foundation’s developer program will provide a single point of access for developer support, providing a wide offering of tools and resources - most available free - including: Software Development Kits (SDK’s)DocumentationSample codeKnowledge baseApplication signing programIncident based technical support
Well, that sounds pretty good — except for the “most” tools and resources. Which ones won’t be? We suspect it’s probably the Symbian Signed process, which isn’t free (either in terms of free beer or free speech) today, and is widely hated by developers. Symbian, or should we say Nokia, is keen on keeping it going as a barrier to mobile malware, but it’s hard to see how it fits in an open-source project that’s a development platform rather than an application. Think of it like this: obviously, open-source means that I can do anything to my copy of Firefox or Linux I damn well like, and further that I can submit the alterations to the Mozilla Foundation.
Beware of customs checks and duties on arrival
Obviously Mozilla is within its rights to make me go through its development process before they let any of my code into the version that is available to the public, and certainly no-one would expect them to let J. Random Hacker offer their version of Firefox as an official Mozilla product through their release process. If I really wanted to, I could fork the project, and offer my own browser under some other name.
But the Symbian signing process doesn’t just apply to changes to Symbian itself — it applies to all applications developed for use on Symbian, at least ones that want to use a list of capabilities that can be summed up as “everything interesting or useful”. I can’t even sign code for my own personal use if it requires, say, SMS functionality. And this also affects work in other governance regimes. So if I write a Python program, which knows no such thing as code-signing and is entirely free, I can’t run it on an S60 device without submitting to Symbian’s scrutiny and gatekeeping. And you though Microsoft was an evil operating system monopolist…
May require proprietary fuel to operate
What else should a Symbian developer be careful of?
Furthermore, the platform will embrace the runtime technologies already used by the developer community allowing for efficient use of existing assets and skills. The platform will support an extensive offering of development environments including native Symbian C++, POSIX C and C++, Python and Web Runtime based on WebKit.
Webkit is open; so is Python, although the Nokia-contributed bits like the GUI toolkit and the wrappers for the Symbian API are debatable. There’s no mention of the Carbide IDE in there, or the standard Symbian C++ developer toolchain. It’s also not clear what influence the Foundation’s directors will have on the signing process.
Curiously, the mobile industry is adopting this sort of semi-open model just as the IT industry has given up on it. Sun has just finished open-sourcing the entirety of the Java world, and has even chosen to use the open-source version of Solaris instead of its own. But both Google and Nokia are choosing to be slightly less than open, or not quite the same way. Criticisms of Google over Dalvik (their homebrew Java virtual machine, which underlies Android, but unlike Java itself isn’t GPL-licenced) are perhaps overdone - Dalvik is covered by the Apache Public Licence, and nobody really doubts the openness of Apache.
But Symbian could be much more open, in more ways than one. Is Nokia under pressure from the chip vendors or the carriers to maintain certain special restrictions? And what would have happened if Psion had gone open source back in 2001?
In our Voice & Messaging 2.0 Report we listed over 70 new services we’d come across in our travels. One that’s come to our notice since publication is SMS GupShup. Whilst there were many ‘me too’ Internet messaging services we reviewed, this US/Indian start-up is noteworthy for its business model.
As operators find voice and messaging markets mature and revenues stagnate, they are looking for new growth. One route is to try to create elaborate new services and persuade consumers to part with money for them. The other is to find ‘upstream’ parties willing to pay to interact with telco retail customers directly with the telco as an intermediary. Advertising is the starting place for many such initiatives, and GupShup as a template for this begs the question: what is the role of the operator? Bit pipe, enabling platform or complete services provider?
Humans are tribal creatures — hairless monkeys with a grooming instinct
Today’s core telephony and messaging products suffer from many limitations, but perhaps the most central is that humans live and interact in groups, and that these products don’t support such activities well. Conference call systems are notorious for their poor user interface, and your phone’s address book never seems to learn that you message the same three people over and over.
But perhaps the most critical limitation is pricing: telcos are determined to scale price linearly with the number of participants in the conversation, but the sender of a message doesn’t see the value scale the sale way. In emerging markets, where alternative Internet and PC-based forms of communication are much more limited in penetration, this forms an important barrier to usage.
Fixing the pricing problem with adverts
GupShup is an SMS (and Web) based group messaging service available only in India, and with 7 million active users. Superficially the functionality is similar to Web 2.0 poster child (or enfant terrible) Twitter, minus the downtime. Users can send messages to a group, and can choose to follow up subscribe to multiple groups. Messaging is push-pull — you send the message into the cloud, but recipients can total control over what they receive. Like Twitter, the result is a stream of banal human existence. Fortunately, that’s what the users want, and is the basis for an SMS industry worth approaching $100bn/year.
The service has a diversified revenue model, comprising:
What makes it special is how the advert is managed and how every advert immediately provides value to the user. Group messages are limited to one hundred characters, with the remaining 60 in an SMS given over to brand advertisers. Sending a message costs the same as your usual mobile rate for one message, but the cost of forwarding that message is then picked up by the advertiser. Everyone wins, and unlike media advertising’s bait-and-switch, there’s a powerful social driver behind it, and the potential for personalisation and innovation.
With ads, everyone really must win prizes
This approach contrasts with the greedy attitude of carriers in the developed world. Many are trialling ad-serving technology that personalises adverts based on the user’s demographics and click stream. Such trials have been secretive, and failed to get user opt-in. Most importantly, they never answer the user’s issue: so, what’s in it for me? The user feels they’ve already paid the full rate for a broadband connection, and what are you doing wiretapping my Web browser and fiddling with the ads?
No wonder the result is a PR disaster and carriers are back-peddling fast.
The lesson is simple. You want to use the customer’s data and the customer’s device and create new revenue streams from them. Note the apostrophes — it’s not ‘customer data’. You’ve got to offer something in return for what you take. And there’s nothing better than the reward being immediate.
So what should carriers do?
As we wrote in our report Telcos’ Role in Advertising Value Chain, overall we are sceptical of operators trying to provide completely ad-funded services, or generate their own advertising inventory. Operators like Blyk are addressing a narrow, high-risk market. That said, as GupShup only cannibalises a rarely-used feature — messaging to multiple recipients — and is likely to stimulate new usage, it could be one worth emulating.
Alternatively, we would consider differentiating our retail pricing by (at least pretending) there’s no more cost to sending to multiple recipients than one (with the reality being you’d probably drop your bucket size). Another approach would be wholesale deals with online services that are heavy SMS users, again to facilitate some creative retail pricing to undo the “group penalty”.
However, a more Telco 2.0 approach is to ask not how to compete with such services, but how to become a supplier to them. Indeed, those very same ad-serving technologies become a lot more attractive in this scenario. Services like age verification, cash collection, credit management, customer care — there is a long list far beyond just the bit pipe. [Ed - which of course you can read all about in our report on The 2-Sided Telecoms Market Opportunity.] It just requires a new mindset around high volumes, ‘horizontal’ business process and value creation — not rent-seeking on the underlying access assets, or dazzling media services.
Telco 2.0 readers will be well aware that we’re very keen on any application that uses telco capabilities to remove friction and inefficiency from the wider world of business - perhaps the fundamental insight in the 2-sided business model is that the telco doesn’t only sell telephone calls as a finished product to end users, but also a much wider range of functions for upstream businesses to integrate into their production process. In terms of economics, these communications-enabled business processes usually exist to reduce transaction costs and thus facilitate trade that would otherwise not happen. Alternatively, they help larger enterprises to overcome their internal diseconomies of scale.
This use case is of the first kind; the telco platform as a trading hub, allowing the many companies that would never be able to build the mass-production IT systems that their bigger competitors use to benefit from increasing returns to scale.
You’d be surprised how significant backloads are for the transport industry and the economy more broadly. At first sight, the economics of a truck route would seem trivially simple; you pay for diesel and wages and capital, collect rates from customers, and costs scale by the mile - right? But it becomes a lot more interesting when you realise that the profitability of a marginal load is highly dependent on whether it is carried on the way out or back; as the truck has to come back anyway, the costs of both trips must be accounted for in the price of the trip out, so the margin on the trip back can be 100%.
This can have profound consequences - the phenomenon of fresh fruit and vegetables from eastern Africa showing up in European supermarkets, for example. Airlines providing cargo service to Kenya and other places noticed that the return trip tended to be empty, and unsurprisingly the answer was to drop freight rates dramatically. Any price at all was better than simply shipping air. It therefore became possible to export produce on a large scale, which has become one of these countries’ biggest sources of foreign exchange.
At a more micro level, the problem for an individual firm or owner-driver is finding a backload in the first place. Unless you can arrange it in advance, this is a major source of uncertainty, and one that may grant bigger companies increasing returns to scale - they can afford a monster IT system to track all their vehicles and customers and match them up, and they have more locations, staff, and vehicles out there looking for backloads. They can dedicate salesmen full-time to searching out and buttering up regular backload customers.
The text-book approach to this is to start an exchange; if everyone brings their supply (i.e. trucks looking for loads) and demand (i.e. loads looking for trucks) to the same place, the chances of a good match are dramatically increased for everyone. The more business the exchange does, the better it gets; prices are more stable, the range of deals on offer wider, and you can have greater confidence that you can buy or sell when you need to. Liquidity goes to liquidity. This dynamic is well-known, and can be perceived in stock markets, Internet exchanges, ports, produce markets, Web search engines, and dive bars. It’s interesting that some of the very first commercial exchanges of this kind were for freight - specifically shipping. There’s a good reason why the London freight bourse is called the Baltic Exchange when it trades in shipping to every port on the planet; when it started, that was where the trade went.
Telcos might create such an exchange, have a share in it, or simply be suppliers to it - this will vary between markets and territories, just as it makes sense for an operator to be a bank in Kenya but not in Germany, or it’s possible to make money from MMS in a human-machine application but not in a human-human one.
But doing this raises some big technical challenges, as set out in this slide from the 2-Sided Business Model report.
If you’re Maersk Logistics, these aren’t such big issues. You can afford to build this kind of capability, and you can pay IBM Global Services to do a lot of it (which is what Maersk did). And you’re a big enough customer that your friendly local telco is likely to be receptive to a wholesale deal. If you’re Charlie Cox, not so much. But Telco 2.0 could change this. It’s all about commercialising the core telco assets by making the key capabilities that spring from them available in forms which allow third-party businesses to use them in new ways, right? In this case, we’re using the telco’s secure messaging capability, its location capability, its identity capability, and its payments capability, just as all of these would be used to deliver an SMS message; we’ve just taken apart that finished product and built something new out of the parts, which we can only do if the telco doesn’t supply it in a sealed box.
While we were researching this for the 2-Sided Business Model report, we calculated that improved backload finding could be worth up to £218m a year in the UK, on the basis of a 5% improvement in load factors. So there’s serious money to be had in there. It’s not surprising we also encountered a number of start-up freight exchanges trying to do just that - one of them is even offering an application for Windows Mobile devices in order to mobilise their IT system. That sounds like an ideal partner for a telco.
This is a special case of our general model for the telco future. At the bottom of the stack, telcos own huge legacy assets which we’ve characterised as pipes, packets, and platters. These produce certain functional capabilities that grow out of them - we’ve described those as the seven questions and various other things. Traditionally, these were then combined into a standardised product by telco engineers and commercialised by telco internal marketers direct to end users. In the future, however, we think they will be sold in three ways - as plain APIs for third-party developers, as integrated products created by third parties in partnership with the telcos, and within products the telco offers under its own brand.
The recent purchase of Symbian by Nokia highlights the tensions around running a consortium-owned platform business. Obviously, Nokia believes that making the software royalty-free and open source is the key to future mass adoption. The team at Telco 2.0 disagree and believe the creation of the Symbian Foundation will cure none of the governance or product issues going forward. Additionally, Symbian isn’t strong in the really important bits of the mobile jigsaw that generates the real value to any of the end-consumer, developer or mobile operator.
In this article, we look at the operating performance of Symbian. In a second we examine the “openness” of Symbian going forward, since “open” remains such a talisman of business model success.
Background
Symbian’s core product is a piece of software code that the user doesn’t interact with directly — it’s low-level operating system code to deal with key presses, screen display, and controlling the radio. Unlike Windows (but rather like Unix) there are three competing user interfaces built on this common foundation: Nokia’s Series 60 (S60), Sony Ericsson’s UIQ, and DoCoMo’s MOAP. Smartphones haven’t taken the world by storm yet, but Symbian is the dominant smartphone platform, and thus is well positioned to trickle down to lower-end handsets over time. What might be relevant to 100m handsets this year could be a billion handsets in two or three years from now. As we saw on the PC with Windows, the character of the handset operating system is critical to who makes money out of the mobile ecosystem.
The “what” of the deal is simple enough — Nokia spent a sum of money equivalent to two years’ licence fees buying out the other shareholders in Symbian, before staving off general horror from other vendors by promising to convert the firm into an open-source foundation like the ones behind Mozilla, Apache and many other open-source projects. The “how” is pretty simple, too. Nokia is going to chip in its proprietary S60, and assign the S60 developers to work on Symbian Foundation projects.
Shareholding Structure
The generic problem with consortium is typically not all members are equal and almost certainly have different objectives. This has always been the case with Symbian. It is worth examining the final shareholder structure which has been stable since July 2004: Nokia - 47.9%, Ericsson - 15.6%, SonyEricsson - 13.1%, Panasonic - 10.5%, Siemens - 8.4% and Samsung - 4.5%. At the bottom of the article we have listed the key corporate events in Symbian history and the changes in shareholding.
It is interesting to note that: Siemens is out of the handset business, Panasonic doesn’t produce Symbian handsets (it uses LiMo), Ericsson only produces handsets indirectly through SonyEricsson, and Samsung is notably permissive towards handset operating systems.
SonyEricsson has been committed towards Symbian at the top end of its range, although recently is adding Windows Mobile for its Xperia range targeted at corporates.
Nokia seems almost committed though has recently purchased Trolltech — a notable fan of Linux and developer of Qt.
The tensions within the shareholders seem obvious: Siemens was probably in the consortium for pure financial return, whereas for Nokia it was a key component of industrial strategy and cost base for its high-end products. The other shareholders were somewhere in between those extremes. The added variable was that Samsung, Nokia’s strongest competitor, seemed hardly committed to the product.
It is easy to produce a hypotheses that the software roadmap and licence pricing for Symbian was difficult to agree and that was before the user interface angle (see below).
Ongoing Business Model
Going forward, Nokia has solved the argument of licence pricing — it is free. Whether this passed to consumers in the form of lower handset prices is open to debate. After all, Nokia somehow has to recover the cost of an additional 1,000 personnel on its payroll. For SonyEricsson with its recent profit warning, any improvement in margin will be appreciated, but this doesn’t necessarily mean a reduction in pricing.
It also seems obvious that Nokia will also control the software roadmap going forward: it seems to us that handset operators using Symbian will be faced with three options: free-ride on Nokia; pick and choose components and differentiate with self-build components; or pick another OS.
We think that given the chosen licence (Eclipse — described in more detail in next article), plus the history of Symbian user-interfaces, and the dominance of Nokia, all point towards other handset operators producing their own flavours of Symbian going forward.
Competition
Nokia may have bought Symbian, even without competitive pressures, purely to reduce its own royalties. However, the competitive environment adds an additional dimension to the decision.
RIM and Microsoft are extremely strong in the corporate space and both share two features that Symbian are currently extremely weak in — they both excel in synchronizing with messaging and calendaring services.
Apple has also raised the bar in usability. This is something where Symbian has stayed clear, but is certainly not one of the strengths of S60, the Nokia front end. The wife of one of our team — tech-savvy, tri-lingual, with a PhD in molecular biology — couldn’t work out how to change the ringtone, and not for lack of trying. What do you mean it’s not under ‘settings’? Some unkind tongues have even speculated that the S60 user interface was inspired by an Enigma Machine stolen to order by Nokia executives.
Qualcomm is rarely mentioned when phone operating systems are talked about, and that is because they take a completely different approach. Qualcomm’s BREW would be better classified as a content delivery system, and it is gaining traction in Europe. Two really innovative handsets of last year, the O2 Coccoon and the 3-Skypephone, were both based upon Qualcomm software. Qualcomm’s differentiator is that it is not a consumer brand and develops solutions in partnership with operators.
The RIM, Microsoft, Apple and Qualcomm solutions share one thing in common: they incorporate network elements which deliver services.
Nokia is of course moving into back-end solutions through its embryonic Ovi services. And this may be the major point about Symbian: it is only one, albeit important piece of the jigsaw. Meanwhile, as we’ve written before, Ovi remains obsessed around information and entertainment services, neglecting the network side of the core voice and messaging service. Contrast with Apple’s first advance with Visual Voicemail.
As James Balsillie, CEO of RIM, said this week “The sector is shifting rapidly. The middle part is hollowing — there are cheap, cheap, cheap phones and then it is smartphones to a connected platform.”
Key Symbian Dates.
June 1998 - Launch with Psion owning 40%, Nokia 30% & Ericsson 30%. Oct 1998 - Motorola Joins Consortium
Jan 1999 - Symbian acquires Ronneby Labs from Ericsson and with it the original UIQ team & codebase.
Mar 1999 - DoCoMo partnership
May 1999 - Panasonic joins Consortium. Equity Stakes now: Psion - 28%, Nokia / Ericsson / Motorola - 21%, Panasonic - 9%.
Jan 2002 - Funding Round of £20.75m. SonyEricsson tales up Ericsson Rights.
Jun 2002 - Siemens Joins Consortium with £14.25m for 5%. Implied Value £285m
Feb 2003 - Samsung Joins Consortium with £17m for 5%. Implied Value £340m.
Aug 2003 - Five Years Anniversary. Original Consortium Members can now sell. Motorola sells stake for £57m to Nokia & Psion. Implied Value £300m.
Feb 2004 - Original Founder Founder Psion decides to sell out. Announces to Sell 31.7% for £135.5m with part of payment dependant of future royalties. Implied Value £427m. Nokia would have > 50% control. David Potter of Psion says total investment in Symbian was £35m to-date, so £135.5m represents a good return.
July 2004 - Preemption of Psion Stake by Panasonic, SonyEricsson & Siemens. Additional Rights issue of £50m taken up by Panasonic, SonyEricsson, Siemens & Nokia. New Shareholding structure: Nokia - 47.9%, Ericsson - 15.6%, SonyEricsson - 13.1%, Panasonic - 10.5%, Siemens - 8.4% and Samsung - 4.5%.
Agree to rise cost base to c. £100m/per annum and headcount of c. 1,200.
Feb 2007 - Agree to sell UIQ to SonyEricsson for £7.1m.
June 2008 - Nokia buys rest of Symbian with Implied Value of €850m (£673m) with approx. payout of - Ericsson - £105m, SonyEricsson - £88.2m, Panasonic - £70.7m, Siemens of £56.5m and Samsung £30.3m. Note, Symbian had net cash of €182m. The price quoted by Nokia of €262m is the net price paid by Nokia to buy out the consortium not the value of the company.
In Today’s Issue: WiFi in your car; connectivity included with O2UK iPhones; Nokia buys Symbian and gives it away; LiPS and LiMo; OpenMoko ships; Sony Ericsson struggling; Reding terminates 70% of termination fees; France Telecom hits the silk from the Telia deal; T-Mobile USA launches femtoVoIP; FemtoForum+NGMN=sense; where is the network API? probably not in the IMS; Blyk expands; FISA fight goes on; T-Mobile UK goes down; Vodafone to buy Ghana Telecom?
Just what I always wanted — WiFi in the car. Chrysler’s UConnect product will essentially give some vehicles a WLAN router with a 3G radio modem for the backhaul. The interesting question will, of course, be the business model. Chrysler says there will “be no tie-in to long term contracts”, but it’s going to be hard to get this off the ground without some element of two-sidedness — perhaps by bundling connectivity in the car as an optional extra, or doing something clever with GPS and localised ads.
You’d better hope the WLAN is encrypted by default, or the term “wardriving” will take on a whole new layer of meaning as hackers chase open WLAN vehicles down the freeways of California, trying to stay in range just long enough to finish that torrent download, or to spork the satnav display with a specially crafted packet. All your SUV are belong to us.
That’s if their attention isn’t distracted by one of this week’s wave of shiny gadget developments. For a start, pricing and launch details of the 3G iPhone are leaked; it looks like O2 in the UK will bundle data connectivity with the prepaid version for the first six months. It’s astonishing; you show a carrier an iPhone and they’ll sign anything.
More seriously, Nokia spent the equivalent of two years’ Symbian royalties buying out the other partners in Symbian, before immediately promising to transform the software house into an open-source foundation. We’d say more about it here, but we’re saving ourselves for a much more detailed post on the issue, so watch this space. It’s an interesting thought, however - what would have happened had Psion decided to open-source the whole thing when they gave up on mobile back in 2001? Another one for the “great moments in the failure of British industry” file.
Almost immediately, there was more mobile open source activity — the LiPS (Linux Phone Standards) Forum has merged with the slightly better-known LiMo Foundation. It’s a sensible move on the face of it - LiMo’s membership contained more carriers and developers, whereas LiPS had more hardware vendors, so the merger will help to provide a comprehensive Linux environment from the silicon up.
Compared to LiMo, the plans for the Symbian Foundation look…not very open at all. But there is one mobile-Linux group who are even open-er than LiMo - yes, up to 1000% more bigger openosity now! It’s the OpenMoko community, makers of the Neo1973 and FreeRunner all-open-source, glatt kosher mobile phone. Well, it’s in the shops, or rather, available through a group of selected e-tailers. Pricing in the UK is around £272 from Truebox. We’ll see your iPhone and raise you…
No wonder Sony Ericsson is looking green at the gills. It’s profits warning time; apparently the economic downturn has hit its speciality of midmarket fashion gadgets disproportionately hard. Smartphone sales are holding up reasonably well, cheaper devices are benefiting from trading-down, but the midmarket shinies suffer.
And the carriers have no business smiling, either; here comes Viviane Reding again. The European Commission has published more details of its plans on mobile termination fees — it wants them cut by around 70 per cent. Ouch! It’s enough to make you abandon a misguided monster merger. Like France Telecom just did. We didn’t think much of the attempt to buy TeliaSonera to begin with, and now it looks like it’s a dead’un. It just wasn’t a good idea in the first place…after all, you could be improving your crucial voice and messaging products, like T-Mobile USA. They’re offering a $10/month unlimited VoIP service to all their cellular customers, on condition they get a “router” which we think is probably a femtocell.
It’s a smart move; not only are they competing sharply on price, they’re doing so using their competitors’ costs. And, as we’ve often pointed out, CPE is an underexploited opportunity for fixed operators to astonish their customers and mobile operators to infiltrate the fixed subscribers’ front rooms. So we’re also pleased to see that the NGMN, the carriers’ talking shop on 4G, is talking to the Femto Forum about including femtocell support in the LTE or WiMAX standards.
As well as improving your core voice and messaging products, what else should you be doing? That’s right, exposing key enablers as APIs so third-party developers can create interesting new services, while redesigning your business to gain upstream revenues. Gary Kim of IPCarrier asks where the network API is. He seems to reckon IMS might help. And Blyk has announced its ad-funded virtual operator is expanding into Germany, Spain, and Belgium.
Finally, thanks to Senator Chris Dodd, the FISA fight goes on; T-Mobile UK’s data service goes down, hard; and Vodafone may buy into Ghana Telecom.
We were asked to present on a panel at the private marketing innovation conference of a UK mobile carrier last week. The subject was the “Need for Speed”: what are the real drivers for network capacity and speed, and thus where should an operator focus its investments?
Since our answers are generic to all mobile carriers, we thought we’d share them with you here. The format was five questions, addressed by all the speakers:
What user behaviour is driving the demand?
The first question noted that customers’ demands for increased speed and capacity appeared never-ending. But what exactly is driving the customer’s seemingly unquenchable thirst for more speed?
We took the view that you have to examine this issue from the perspective of value, not volume. There are three fundamental modes of communication: information-based applications; real-time personal communications; and entertainment (more depth, as always, in our Voice & Messaging 2.0 Report).
The Internet, and especially the Web, is shifting in role from a way of distributing stored information/documents (i.e. hypertext) to a way of interacting with applications. This is evidenced by the general “cloud” trend of software-as-a-service, Akamai’s Edge Computing initiative, online multiplayer games and virtual worlds. Related to this, the bandwidth requirements of common applications are storming up. This is particularly true for the uplink: every photo and video captured disappears into the cloud at full resolution, but not all are ever viewed, and those that are may only be viewed in thumbnail or extract form.
However, it’s not primarily about information, but about ‘presence’ — the sense that you’re with someone despite not actually being physically there. Communications systems try to replicate this ‘being there’. It should feel very strange talking to a lump of plastic at the side of your head that whispers back in your ear. Yet our brains are surprisingly good at smoothing out audio and video artifacts , and the illusion we’re really talking to someone else is maintained. However, consider when you want to share that 10Mb Powerpoint deck half way through the Skype conversation. This is the digital equivalent of pushing a document across the table in a meeting room. To make the illusion work you can’t have a pause. Therefore we tend to need very short bursts of high speed.
Entertainment is all about filling the void in between bouts of communication. It might be able to absorb network capacity, but it won’t be what drives the business model. Furthermore, such traffic is often best delivered by other means — broadcast, edge cached, or sideloaded — and competes against low-cost distractions such as games. Tellywood alone won’t pay your bills. Which access technologies to invest in?
Next we considered the plethora of different technologies available to fixed and mobile operators (e.g. 3G, HSDPA, ADSLx, Femtos, WiFi, Wimax, LTE, Fibre etc.). Which ones should mobile operators care about most?
Here, the job of the telco splits into two: the wholesale telco, that aggregates networks and provides the ‘logistics platform’; and the retail services provider who integrates and promotes the end user experience. Defining the products and services themselves is slowly migrating away from today’s operators towards other players (e.g. Nokia Ovi, Apple iPhone, Amazon Kindle). Many of today’s operator functions such as billing and customer care may become white-label inputs into these other businesses.
In choosing a network strategy, operators should prefer breadth over depth — coverage always wins. Verizon Wireless trounced Sprint, with identical technology, by emphasising network coverage and quality, whilst Sprint pushed advanced application services. Customers care little or nothing about network data speeds per se. Since femtocells and home hubs extend in-building coverage, and boost speeds too, we’d say that’s where the future lies. Furthermore, most of the ‘dead time’ filled with bulky low-value entertainment also occurs in the home, which makes it all the more important to lower your network costs there.
The wholesale platform needs a wide range of delivery capabilities, just as a physical-world logistics company needs road, sea, rail and air freight assets. This again points towards building ‘edge’ assets through the retail side, which can then be used to create wholesale solutions for partners wishing to cheaply and effectively deliver content and services to users.
The trick is to integrate all the various access technologies to present an experience to users that matches the capabilities and limitations of the different bearer technologies.
What we certainly wouldn’t be doing in mature, highly teledense markets, is rushing out to build new WiMAX macro networks.
What kind of devices will absorb this capacity?
Next we considered what kinds of devices that customers will use to quench their thirst for increased speed and capacity.
The hardware form factor of user devices will change least of all — because much of their design is determined by human ergonomics. Their software, however, is a different question.
The focus will slowly move away from the device onto the data it carries and can capture or display. Wireless Grids is an example of a company that lets you use the displays, input devices, printers, and other resources around you. For this to work, there will be an increasing emphasis on short-range communications and personal area networks. Synchronisation between devices is current a nightmare part of the user experience - solve this and print your own currency. Your ‘phone’ might be a primary source of data, identity, authentication etc. to lubricate these interactions.
Does wireless displace fixed access for data?
For the fourth question, it was noted that the last decade had shown that the “gap” between fixed and mobile networks in terms of data speed and capacity of retail propositions is closing fast (even if the fundamental technology gap remained, as optics leap ahead). The same decade saw an equivalent trend in quality and convenience of voice calling gap close, and displacement occur. So, will the need for fixed broadband data products “disappear” in the same way?
In general, we think not. The comparative advantage of fixed networks will, however, slowly erode without a fibre build-out. In some rural and dense urban areas we may see a switch to fixed wireless. Yet absent some leap in mesh networking, wireless doesn’t look like having the capacity or economics for video delivery. As we noted before, there is also more dead time in the home and therefore the consumption of fat video products will always be higher there.
The focus, however, needs to be on getting a lot more out of the infrastructure we have. Femto roaming, for example, could become a big deal — where carrier A recompenses carrier B for using your neighbour’s femto when you’re down the end of the garden.
How does the customer relationship evolve?
Finally, the mobile business is shifting from selling voice calling and text messaging to connecting customers to the Internet in better and faster ways. Where does this leave the operator’s role in the customer relationship in the future?
Ask your customers — do any of them want a ‘relationship’ with their phone company? Consumers don’t get excited about who supplied the batteries in their phone, so why should they be worked up about the radio network?
Ideally telcos should become better retailers — category management, packaging, promotion, segmentation, distribution. This is hard, as they lack many of the necessary skills and assets. For example, the ‘one size fits all’ mobile retail store doesn’t match the segmentation needs of an increasingly fragmented marketplace.
In reality the game is all about gaining a larger slice from the device makers and upstream players. This will only be done by offering services they both value and struggle to replicate in scale. This is best done by creating two-sided markets (see our report here for more information) to exploit the telco assets better. (It was clear in discussion afterwards that they see limited mileage in trying to squeeze consumers for more value-added content services, and that the growth will have to come from elsewhere.) Otherwise, the best relationship is no relationship, and become a specialist wholesale-centric carrier.
[Ed - We’ll be covering some of these issues in more depth at the 4-5 November Telco 2.0 event in London. Agenda now available here]
Following the popular article last week on FTTH prospects for the UK, Benoit Felton of Fibrevolution sent us some excellent material looking at the issues from a more international perspective. He says:
In April this year I attended a very interesting conference in Stavänger, Norway, organised by the OECD on Next Generation Access Networks where I conducted a number of interviews with experts. The end result is two podcasts:
The Stavänger Show - Part One: FTTH Policy focuses on the challenges raised by next generation access when it comes to government policy and regulation. The people interviewed are:
Marvin Sirbu, Prof. Engineering and Public Policy at Carnegie Mellon University; Antony Walker, Head of the Broadband Stakeholder’s Group; Grant Forsyth, Head of EMEA Regulatory Affairs at BT Global Services; Dimitri Ypsilanti, responsable for the OECD Working Party on Communications
The Stavänger Show - Part Two: Deployment Models focuses on the various approaches to NGANs, from municipal open access to incumbent vertical integration. The people interviewed are:
Herman Wagter, CEO of Glasvezelnet Amsterdam; Dennis Weller, Chief Economist at Verizon; Christian Berg, Consultant at Dansk Energi; Richard Clarke, AVP of Regulatory Planning and Policy at AT&T
In Today’s Issue: 60 years of computing - our Mancunian future; 25 years of DNS, 10 years of a post-Jon Postel world; securing the root DNS; Yahoo! loses clue to the wider environment; Apple’s outrageous iPhone margins; iPhone-RAZR culture shock; 1st VZ ODI gadgets; Moto tries to slim itself fat; Huawei handsets up for grabs; Telefonica leads misguided acquisition rush into China; Chinese bank buys Poland; first WiMAX.eu; Sprint XOHM goes live in September; Sprint offers enterprise e-mail; Nokia builds mapping capabilities; “IMS light”, again; communicating non-neutrality; Isenberg makes stabby over FISA
It’s been sixty years since the very first computer that accepted a stored program, Manchester University’s “Baby”, successfully determined the prime factors of a given number. The beginning of computing is one of those events it’s hard to date - in the UK alone, you’d have to consider the rival claims of the Cambridge Maths Lab, NPL, and Manchester, to say nothing of the code-breaking COLOSSI (although they weren’t capable of re-programming in memory), the US’s ENIAC, or Konrad Zuse’s work in Germany. But Baby’s special claim is because it was both a digital computer, and one which could read a stored program; you can date the beginning of the primacy of software to this point, and hence essentially everything that defines the IT industry and its distinctive culture.
Whilst we’re on the topic of history, it’s been 25 years this week since the first DNS server went on line, just one of the many contributions of the late Jon Postel as co-author of RFC882, which specifies the DNS, administrator of the .us TLD, president of the Internet Assigned Numbers Authority, and editor of the RFCs. This is a fine excuse to link to RFC2468, the tribute to Postel written into the Internet’s standardisation process (Postel’s own invention) by none other than Vint Cerf. And yes, that is as in “2,4,6,8, who do we appreciate?” More practically, the Renesys blog has some thoughts about the problems of securing the root servers, with a handy list of where they should be.
At Yahoo!, the problem is increasingly that executives who should be inside the building are not. Since the collapse of the Microsoft bid, people have been leaving in numbers, starting with the founders of the numerous whizzy start-ups they bought over the last five years - the founders of Flickr and Del.icio.us being the most senior to walk so far. (Can anyone imagine what the walk-out rate would have been like if Microsoft had bought the company?)
Yahoo! might, however, find some relief in the fact that the majority of mobile search requests come from iPhones (German link); perhaps they might cuddle up to Apple. That’s if Apple actually wants them - the latest forecast for the iPhone bill of materials puts its cost at about $100, which implies an outrageous profit margin.
It turns out that a quarter of all iPhone users switched from a Motorola RAZR, by far the biggest single group among them. It seems exactly the same gadget hipster twits who bought iPhones were the ones who flocked to RAZRs three years ago. No wonder they love the things - the RAZR V3 had some of the worst figures in industry history for customer satisfaction, and a reputation for poor quality control to boot. The comparison with Apple’s optical glass, burnished aluminium, and Unix can only be telling.
Verizon, meanwhile, has homologated the first ODI device.
Motorola, meanwhile, reached for the classic response to hard times - sack half the R&D lab. This is now the second major cut in product development and research at Motorola since the crisis began; it’s not the most obvious strategy for a company whose problems stem from a lack of good products. Perhaps they assume that whoever buys the handsets operation will already have their own lab? (hint - Nokia?)
If you want a spare handset factory, though, Huawei is selling a large chunk of its devices operation. The leading Chinese vendor is far better known for its network infrastructure products, and it looks like they are planning to specialise in them, like so many other NEPs. But however low the margins on Huawei gadgets are, you can bet their problems aren’t as bad as Moto’s.
We mentioned that the reorganisation of the Chinese telecoms industry is likely to trigger a rush by essentially all the NEPs to sell them CDMA2K and UMTS gear, and all the major telcos to try to buy into the new big three converged operators. Telefonica kicked it off this week, angling to come out of the China Netcom/China Unicom merger with a 10 per cent stake. There’s something slightly worrying about this business of very expensive minority stakes far from home - it sounds a lot like being a Western oil investor in Russia, and you have to remember the sad tale of BT’s Japanese investments.
Interestingly, this works both ways: here’s a Chinese bank funding a greenfield mobile operator in Poland. In fact, it’s more of a monster vendor financing deal than anything else, as all the equipment is coming from Huawei.
Europe’s first mobile WiMAX net launches in Amsterdam; Sprint confirms that its first commercial WiMAX network will be live in September in Baltimore. More usefully, perhaps, they also started offering MS Exchange and Lotus Notes e-mail for nonfancy devices.
Nokia, meanwhile, opened yet another web storefront; more interestingly, they also bought a rather impressive social mapping application, Plazes.
NEC launched something called “IMS light” (again) at NXTComm - about all that’s interesting here is that the applications they are pitching for it are mostly unified comms. A couple of years ago, the first thing any IMS person wanted to show you was an “interactive video-sharing” service, which were uniformly dire. I remember vividly the one the GSMA showed off with Nokia at 3GSM 2006, which was reduced to semi-functionality by latency - especially amusing due to the number of people involved busy explaining how only they could guarantee acceptable quality-of-service for video streaming.
Relatedly, Telephony Online asks how carriers who want to offer multiple levels of service by application will communicate this to their customers, given the number of people who complain of “slow” broadband within the first month of getting it.
And finally, the US carriers were officially shriven of responsibility for their part in illegal surveillance operations. David Isenberg is furious.
So, with two major US carriers rolling out fibre to the home, a string of European cities doing the municipal-fibre thing, Iliad fibreing-up their own network in France, and Japan and Korea having long started wiring up whole apartment buildings, how soon will the UK get cracking? Telco 2.0 went to the Broadband Stakeholder Group’s conference to find out.
Background to the issue
The broadband incentive problem tells us how there’s little incentive for network owners to invest in networks when they can’t capture much of the incremental value of the traffic. One way out would be to make a radical cut in the underlying incremental costs of bandwidth, and to stretch budgets further. And that’s precisely what we’re seeing all over the world, as operators upgrade in order to substitute new CAPEX for old OPEX.
There are many ways of doing this: deploying fibre, DOCSIS 3 cable systems, and advanced wireless in the access loop; moving to technologies like Carrier Ethernet inside their networks; and substituting peering for transit whereever possible. Mobile operators are increasingly pulling fibre to their cell-sites in order to cope with a rising tide of data traffic encouraged by the arrival of megabit-plus radio links.
Verizon estimates that it saves up to 70% of OPEX on every link it converts to FiOS. So you’d think the pressure would be on to get the fibre out there in Britain, a country criss-crossed with high-maintenance copper in a damp climate. The UK is also perhaps the guinea pig for the broadband incentive problem. But FTTH is further behind in the UK than almost anywhere else in Europe. So far there is literally no SOHO fibre access anywhere in Britain. What’s going on?
Correlates of Success
The first problem with a UK national FTTx roll-out is that nothing like that has been done before, so no-one really knows what is going to happen, who will benefit, or how much it costs. The BSG, however, has commissioned some research into experience from the existing fibre deployments.
You can draw a number of conclusions from this, although it’s worth noting they didn’t include the major PON deployments in Japan, Korea or the US. For a start, it seems clear that locally-owned fibre is more likely to succeed. Relatedly, the crucial variable in the success or failure of a deployment is take-up; there is a strong link between community ownership and adoption.
Secondly, open access is correlated with success. The most successful fibre builds are the ones that practice open or shared access at a low level in the protocol stack. Thirdly, Ethernet is to be preferred to GPON, etc. — it’s not clear whether this is inherent in the technology, or whether it is because the successful open-access, locally-owned builds chose it for other reasons. And finally, layer zero is king. Perhaps the defining factor is how easy it is to get access to civil works, which represent around 70% of costs. It’s easiest to deploy fibre by linking up apartment blocks, especially new buildings where the infrastructure can be put in during construction.
Unfortunately, this is much more helpful in Paris or Amsterdam than it is in a sprawling British suburb. Here, it’s unavoidable that anyone deploying fibre will have to cover a lot of trench mileage - so who’s going to pay for that? The owners of the existing infrastructure are of course BT and Virgin Media. BT has been politicking with the government for years about the fibre question, and we’ll come to that later. But first, let’s put on record that Virgin didn’t come to the BSG. With their own infrastructure, and no annoying open access requirements, plus a permit from OFCOM to call their co-axial network “fibre optic” in adverts, their plan appears to be simply to ignore the issue. So much for that option. And to be fair, the Virgin Media network seems to be excluded from the options considered in the Plum Report commissioned by the BSG.
The Politics of Openreach
The other national network is of course that of BT Openreach, chartered steward of the copper wires. In their role as wholesaler to all the ISPs and alt.nets, they are the obvious choice. But Openreach, and BT more broadly, is in a very strong bargaining position — and they are determined to extract a high price, in terms of regulatory concessions, State funding of the deployment, and actual pricing of both the future network and their existing one. BT shareholders would expect nothing less. There is a two-level game in progress, on the upper level of which OFCOM, DBERR (the Department of Business, Enterprise and Regulatory Reform - the former Department of Trade and Industry), and Openreach are negotiating about fibre, and on the lower level of which OFCOM, Openreach, and Openreach’s customers are negotiating about Openreach’s regulated pricing. This obviously strengthens the hand of BT, as it can play the two levels off against each other, by demanding higher prices in exchange for fibre deployment or holding fibre hostage to pricing negotiations.
The good news is that Openreach CEO Steve Robertson, going by his contributions at the BSG, is conscious of the telecoms industry’s crisis. Openreach (and BT Wholesale) face a complex optimisation problem in the lower level of the game. They quite reasonably want to maximise their revenue, and doing so requires first of all that they square the regulator. But if they push the price up too far, they will kill the ISPs, even if they don’t fall out with OFCOM first. So Steve Robertson has to simultaneously pursue a maximal regulatory goal whilst also persuading OFCOM that if they give him the pricing power, he won’t push it too far. He appears, fortunately, to be aware that there is a significant risk of ISPs collapsing, which would leave Openreach faced with perhaps two national ISPs, and thus be put in a very difficult negotiating position.
Customers, however, much as they would love to have fibre access, are not being heard in the upper level of the game, so they have no other option than to oppose any increase in regulated prices and hope that something will turn up. OFCOM is marginally more sympathetic to them than DBERR. The worrying thing is that unlike BT, neither Government agency seems to be aware that there is anything wrong. Both OFCOM and DBERR representatives at the BSG seemed to believe that the ISP market is currently in a stable equilibrium with genuine competition, rather than its costs exploding (and being substantially determined by a monopolist), while a price war holds down their revenues to unsustainably low levels.
The Broadband Crisis
A British DSL provider can be modelled as having two substantial costs — BT, and everything else — and one source of revenue, subscribers. Under BT there is the cost of IPStream service, which is billed per-bit, or LLU rental, and BT Wholesale backhaul (which also scales with usage). Under ‘everything else’, we have the costs of electricity, salaries, rent, and capital. Given that there is little difference in the CAPEX requirements of any two DSL providers of the same size, that their requirements for premises and power are essentially the same, and there is a free market in network engineers, it’s fair to assume this will be much the same for everyone. Similarly, the BT bill is largely determined by two factors — the OFCOM-regulated price, and usage. (There is potential for some variation due to different cost-structures, for example the proportion of LLU versus IPStream lines, and the fraction of backhaul which is subject to actual competition. However, these are in the nature of a one-off shift.)
With a small number of ISPs in the market, we have the classical conditions for oligopolistic price stability. Whoever raises prices first, or fails to match a price cut, lo