Requiem for GSM

This is the content of the fifth Future of Communications newsletter. To get future newsletters, sign up on the right.

Last month I took out my intellectual geometry set, charted the trajectory of telecoms, and noted we’re at the top of the curve. This Peak Telecoms effect is an unfolding process, not a single moment in time. It has taken around 170 years to get here from the beginning of commercial telegraph service, and is happening slowly over a period of a few decades. It is that slowness that makes its progress hard to see, and its timing hard to call.

The process of decline is also paradoxically a measure of success. Just as we no longer toil in fields to feed ourselves, the ideal amount of labour and capital needed to communicate globally is zero. From the perspective of humanity this is something to welcome, even if telco shareholders suffer.

One of these slow alpha-waves of change in the telecosm is the upcoming decline and death of the GSM ecosystem. As the biggest and baddest technology ecosystem in the world, this is very significant. Let me point out how it might gradually corrode and collapse – albeit over a long period – and what might grow in its place.

We come to celebrate GSM

GSM is arguably the single most impactful technology on everyday human existence since the wheel. (OK, since the axle and second wheel – the first wheel was a confusing novelty.) Superlatives like “astonishing” are appropriate. In a mere two decades GSM has created a connected planetary populace. The spread and impact of even the printing press cannot compare. The core offer is a perfect packaging of human voice and simple text into GSM’s mobile telephony and SMS standards. A $20 GSM phone with a $3 service plan is near-miraculous source of value. The world’s richest man is an emerging-market GSM entrepreneur, not a software mogul or energy tsar. This is a technology that has outpaced even the spread of clean water and mains electricity.

This achievement cannot be understated, and should not be diminished. Too many Web-heads dismiss the benefits that GSM has brought. It wasn’t the Internet that connected billions, it was GSM.

Yet by looking back at the causes of GSM’s success, we can see some of the consequences of the telecoms tide turning. The same human urge to connect and belong that drove the epic spread of GSM will drive the next wave of communications, and the next after that.

I believe these forces will rip apart the GSM ecosystem, taking a lot of the telecoms industry with it, and pushing us down the far side of the curve past Peak Telecoms.

Sources of GSM success

GSM delivers a basic human need to communicate at a distance, a need that was previously unmet. The essence of GSM is a careful technical and economic co-ordination of the value chain to deliver that basic human need. GSM brought together devices, radio networks, communications services and distribution via SIM cards. If there was a case to be made for intelligent design over competitive evolution in technology, this is it.

It simultaneously delivered two things: an interconnected means of delivering bits, and interconnected services that allowed humans to create and interpret those bits. This is critical. Vertically-integrated and special-purpose services precede the spread of modular and general-purpose ones like the Internet.

To recap: GSM addressed a compelling need; brought together all the necessary pieces; was simple to use; and solved the problem of both distribution (SIMs and networks) and service (telephony and SMS). We’re going to come back to these.

GSM ecosystem is put on suicide watch

Whilst the marketing people would like us to treat UMTS (aka 3G) and LTE (aka 4G) as something different from GSM, these are really just iterations of the same basic pattern of voice, messaging and data, delivered by broadly the same players. Whilst there is lots of good work on radio standards and interoperability going on via the GSMA and 3GPP standards bodies, it is the ecosystem as a whole that is under threat.

My Future of Communications co-conspirator Dean Bubley has been on a bit of a roll recently, and proposes many means by which the GSM ecosystem unravels in the face of ‘over the top’ rivals.

• Vast over-supply: There are over 800 providers of the commodity service we label “telephony”. This can be turned from a network service into a cloud application. This breaks the federation and interconnect model for voice and messaging, and undermines the raison d’etre for the GSM ecosystem.
• Lack of demand: The process of iterating telephony and SMS into RCS is entirely technology-driven. No customer has ever asked for a bad imitation of an early instant messaging client. No Jobsian figure integrates it into a wider design picture.
• Lock-in to obsolete usage patterns: The more the industry extends the life of telephony and SMS, the harder it gets to incorporate new patterns of communications. VoLTE in particular replicates the limits of 2G voice, which replicates many of the limits of analogue fixed-line telephony. New paradigms are emerging all the time; for instance Twitter gives us a model of semi-public interwoven conversations.
• Blocked route for enterprise up-sell for telcos: The “bring your own device” trend in enterprises means “bring your own service provider”. Apple and Google don’t have this problem, as they offer network-independent services with direct user relationships. Thus BYOD becomes a threat to GSM, since the whole point is to lock value-creation to the network. Oh, and enterprises are your best customers.
• Network-centric rather than customer-centric: We see one fixed operator, Rogers in Canada, leaving the ‘official standard’ with a ‘Semi-OTT’ solution. Notably, it uses proprietary protocol extensions and does not rely on tight coupling of device and network standards. Indeed, it uses a standard PC SIP client for unified communications from CounterPath, not a telco IMS one. This stuff isn’t new (BT & Mobilkom got there first around 2006-7), but that we’re supposed to be living in a multimedia services nirvana built on telco standards by now (unless you read what I wrote in 2006.) It isn’t happening, and we’re moving to a Web/OTT stack instead.
• Contradiction between voice and data strategies: Got data QoS to sell? Then you don’t need telephony, or RCS, or any other communications service tied to your network. This in turn calls into question the IPX programme the GSMA has created to support interoperable IP data services.
• Prisoner’s dilemma defections: The first mobile operator to defect away from the formal standards approach, and gain a substantial user base of ‘Telco-OTT’ users from other carriers, gets the biggest reward. Fiduciary duty to shareholders over-rides collective loyalty to the GSM system.
• Dud ‘big data’ strategy: The day will come when all voice conversations are by default recorded and computer-analysed and indexed. There is a mis-alignment between the expectation of users and telcos on this ‘big data’ future of voice and messaging. Rules like CPNI that cloak telephony in expensive regulation make it hard to evolve. Google and Facebook can get away with things telcos cannot.

As Dean notes, there is a real danger of misguided 4G voice and messaging programmes causing real damage to existing 2G and 3G voice and messaging revenues. Device fragmentation, complex user interfaces, and customer confusion may make people seek out simpler alternatives. Everyone on 2G and 3G was a talker and texter. Roaming generally worked. Pricing was predictable. None of this may be true in 4G.

The GSMA is potentially sponsoring the demise of its own greatest achievements – mobile telephony and SMS. And nobody is going to stop it.

Indeed, some of the structures of the GSMA has created are now becoming a liability. Take roaming, for example. This month Maxroam launched a multi-IMSI SIM card with cheap global data roaming. Maxroam breaks the lock-in that SIM cards have traditionally offered, and potentially tips the balance of power towards wholesale aggregators and away from retail telcos. Again, not new (Truphone were first), but the pricing shows how the balance between wholesale and retail is shifting. Once you lose that customer relationship, it’s hard to get it back, and you lose margin.

So apart from that, what else has the GSMA done for us recently?

GSMA considered harmful

The GSMA has two major initiatives that could do significant further damage.

The first is OneAPI. This is a programme to create a standard set of APIs into telco networks and billing systems. Unfortunately, the way these are structured is like writing “device drivers for the cloud”. Just as Windows commodified PCs and their components by a device driver abstraction layer, the GSMA is making it really easy for powerful cloud players to commodify networks (which, as you will remember from last time, are just large distributed supercomputers).

There are ways of avoiding this by clever structuring of APIs. Don’t bet on it happening.

The other initiative is WAC, which is a standard platform for wholesale applications. The downside is that it is been positioned to compete with Apple-style “packaged software” rather than telco-style “packaged communications services”. So it will cost a lot of money, and generate some sales (carrier billing has value, particularly for pre-paid), but comes at the opportunity cost of solving the problem that really matters: value-added features for telephony and SMS.

The wrong kind of digital DNA?

I think there may be a deeper issue here that limits the viability of these GSMA initiatives. Telcos have operating and governance models that are adapted to eliminate risk and maximise “certainty of reward”. In other words, they are biased towards “optimise” rather than “explore”, which encourages herds of slow-moving herbivorous ‘fast followers’. That’s dangerous as the predators get stronger.

Paul Golding, who has done a lot to make Telefonica a leader in this space, notes that even they have struggled to acquire “software DNA” in the softco part of the business. My experience of BT’s Ribbit acquisition was that while white blood cells are excellent for repelling pathogens, they also reject donor organs vital to future wellbeing.

As HBR puts it, Apple’s success comes from putting passion for product ideas over penchant for profitable investments. That’s a firing offense at telcos. I know, because I keep getting fired for it.

So, in summary, the GSM ecosystem is largely incapable of absorbing these new technology advances, and if it does they may end up causing harm rather than good. The value really is created outside of the network, and you can’t undo it.

The best-case scenario for telcos is that suppliers like Ericsson, who are going all-in for Web technologies like WebRTC, to give their customers a profit transfusion. They may also help to rescue some of the stranded investments in IMS and related efforts with hybrid telco/OTT services.

The effect on the telecoms industry and its supply chain

It’s f*cked.

Ah, sorry about that. You wanted a more professional and considered piece of analysis.

For operators, I’ll quote a few words from Alan Quayle’s excellent report, The Services Domain (buy yours here):

Service Provider P/E (price earnings ratio) is currently 15-30, while for a commoditized industry, such as paper or cement, the P/E is 3-5. This difference impacts the industry's ability to invest, its head count, salaries, etc. Essentially an organization 5-6 times as small will be required.

Telcos have created immense complexity in their operations and back office. Watching the gyrations of BT to get this under control in the simpler fixed network domain, where everything looks like a circuit, makes me think that a lot more shrinkage is needed in mobile operators.

As for the vendors? There are a number of tier-2 network vendors like Metaswitch who “get it”. And boutiques like Voxygen who “deliver it”. (Assume I have done business with these unless told otherwise.) The point solutions for traffic management need to be integrated into holistic services (Ericsson wins, and Huawei follows).

And everyone building IMS systems, RCS clients and IPTV wonderlands? The future looks short and brutal. If you see a red handle near you labelled “consolidate”, please pull it.

Rise of a successor ecosystem

The smartphone is the platform. The iPhone is the first real smartphone usable by ordinary humans with ordinary lives. Everything else before was an interesting prototype.

The smartphone is a fundamental shift, not just another connected device. It merges man and machine. We can’t live without it. It has stunning sensory complexity and use potential. This matters because the bottleneck to creating value is now between human and devices, not devices and network. The way you alleviate that bottleneck is to focus devotedly on user experience.

Apple sells a digital lifestyle experience. The devices are just tokens to activate the experience. Nobody else is in the experience business yet. Apple is therefore quietly capturing privately the whole ecosystem effect. Chipset, device, services, supply chain.

The only missing component is network, which is a subject for another day. Apple will dictate the network features. Apple will define the core service capabilities.

The fundamental force driving telecoms is an oscillation between horizontal (modular) business models and vertically-integrated ones. Apple is the next upswing in vertical integration. In the words of a recent tweet from Horace Dediu, a brilliant telecoms analyst, it has taken Apple eight weeks to add one Facebook to its market cap. Do you get it now?

To recap: Apple addressed a compelling need; brought together all the necessary pieces; was simple to use; and solved the problem of both distribution (retail) and service (application store). Their stock still looks cheap. They haven’t yet begun to spend their money on anything really disturbing and disruptive to telcos. But then, why should they, when telco shareholders are so happy to subsidise Apple ones, to the point of turning white and stopping breathing?

When will everyone in the GSM world understand they are on an iceberg, not solid land?