Speaking Engagements & Private Workshops - Get Dean Bubley to present or chair your event

Need an experienced, provocative & influential telecoms keynote speaker, moderator/chair or workshop facilitator?
To see recent presentations, and discuss Dean Bubley's appearance at a specific event, click here

Wednesday, October 31, 2012

Interoperable does not mean "federated" - lessons from email

When I've recently debated telecoms "traditionalists" about #TelcoOTT and the Future of Voice, I've noticed that a familiar theme is around interoperability.

The argument sometimes goes along the lines of insisting that we shouldn't have "islands" for messaging or voice or social networks. There is then a rapid leap to conclude that the traditional model of interconnected telecoms networks/services is something to be emulated in future. (And, indeed, enshrined in platforms like IMS).

In general, I think that fragmentation is always a stronger and more powerful trend than convergence. I have been presenting slides highlighting the importance & value of "divergence" for more than 15 years now.

But let's leave aside the hypothetical discussion about silos for a minute. There is definitely value in having *some* services or applications interoperate at least *some* of the time, that I will agree.

But there is a critical distinction that is not made by many:

Interoperability does NOT imply "Federation".

Federation is where every network has its own, dedicated service or application platform, and they interoperate via standardised network-network interfaces. This is familiar from services such as today's telephony and SMS. In those cases, there is a direct link between network infrastructure and services - so each telco has its own services and network, and interoperates/interconnects at their "border". A call starts on one telco's network, and ends on another's - AND, critically, it uses BOTH telcos' application platforms as well. The user has a service identifier (number) which is directly linked to the access line.

But that is not the only model of interoperation

There is another hugely-popular form of fully-interoperable communications, which is not "federated" in the same way. It operates in a fashion completely divorced from the network and access mechanisms.

That application is email. It interoperates almost perfectly. But it is not tied to an access provider (although your ISP can give you a dedicated email address too). It can be accessed on any device, via various protocols (POP3, IMAP and so on). It has been enhanced over the years. it can be web-based or client-based. And it works pretty flawlessly, most of the time.

A good way to think about it is that you can email yourself, using multiple accounts or apps on the same device - or across multiple devices. You, the user, might choose a primary email account, but it's decoupled from the access part. In theory, you can have private email "islands", for example inside companies, that don't interoperate.

I think the email model is a possible way to evolve voice telephony and make it more useful and enduring, especially in mobile. You could have multiple "lines" from multiple service providers, on a single device. At one level they would interoperate perfectly, but they might have separate special features or business models, in the same way that Gmail is different to Hosted Exchange or assorted others. (I still pay a subscription for Yahoo Mail Plus, because I like the disposable email aliases & the spam filtering is really good).

However, for this to happen, there needs to be a disaggregation of phone numbers from SIM cards, and will likely need to be done via Telco-OTT and LTE networks (or perhaps WiFi). It might be possible to have multiple "VMVNOs" on a single SIM as well, I guess - perhaps using multi-IMSI.

I think that future service/app interoperation will be driven by business model needs, or customer demand. If 100m users demand that Skype interoperates with Google Voice (or VoLTE), I'm sure Microsoft will consider it. Various IM and VoIP services already interoperate, either directly like MSN-Yahoo messengers, or via an exchange like Xconnect.

But while interoperability will continue to have value, I see zero - or perhaps negative - value in the legacy federated model. If anything, it enshrines business model and technical rigidity. It would be difficult to have an interconnect agreement between a fully-paid and a freemium telephony service, for example, as payments would have to depend on the status of each party.

In my view, email is the "forgotten" ubiquitous service. While it might be deeply unfashionable, and with revenues that are both small and hard to extricate from wider Internet usage, it is worth examining as a model for future interoperability. In particular, its standards do not require a specific type of access service, or enshrine a business model. Also, because it is easy to sign up for 2nd, 3rd, 4th or n'th accounts, it is a low-key way to extend your brand or ecosystem, without the pain of negiotiating users' switching barriers.

As such, an interoperable but non-federated model could also be the solution to the "speaking agency" problem I outlined in yesterday's post.

Tuesday, October 30, 2012

Telcos' role as "speaking agents" in voice telephony will inevitably be disintermediated

Imagine, if you will, a business where hundreds or thousand of independent service providers link millions of customers with a complex network which enables long-distance communications, earning a commission each time the users make a payment.

I am, of course, talking about travel agents selling flights.

Like many other agency or brokerage-type business models, they have seen revenues and margins drop precipitously because of the Internet, especially where they historically occupied bricks-and-mortar premises. They have often been disintermediated by new web-based businesses with different cost structures (eg Expedia) or airlines selling direct (EasyJet, JetBlue etc).

The buying of airline tickets is not a "service" as such. It's just a normal function of the airline along with flying the planes, baggage-handling and maintenance. (Yes, I know they contract out certain bits, but that's not a service from the passengers' viewpoint).

Some offline travel agents have nevertheless survived, along with a new tier of online aggregators / affiliates (Kayak etc). If you want to book a round-the-world flight, you'll probably still go to an expert, although even there I've seen online systems gain in sophistication.

Other travel agents have gone down a bundling route (eg lowest-common denominator package holidays) or specialised in unusual destinations, demographics or unique adventures or experiences (eg The Adventurists). Often they will make money from ancillary services (accommodation, visas etc) rather than the flights. Some package operators have vertically-integrated and now own their own airlines.

Why is this relevant?

Well, firstly because this illustrates an important problem with the Future of Voice that many of my august peers seem to overlook. We over-focus on increased supply of telephony (so-called OTTs, MVNOs, new entrants like Free in France) and don't focus enough on peaking or decreasing demand for telephony. When was the last time you actually phoned a travel agent to book a flight? There are simpler better (not just cheaper) ways to do it.

But it's the second point that's more important - philosophical, even.

Today's telcos are often just acting as "speaking agents". They just intermediate between you and the person you want to talk to, over a distance, taking (effectively) a commission from the value of your conversation.

That is not a sustainable business model. It is ripe for disintermediation.

Speaking at a distance is not obviously (or exclusively) a "service" proposition. You don't need an agency involved, unless it adds significant value. In some cases, bundling can be "value" but only if it's cheaper or much more convenient than buying the components separately. Some of colleague Martin Geddes' ideas on Hypervoice (adding context & actions to voice streams) are valuable. DoCoMo's cloud translation service is value. Numbers have value for instances where you're calling someone new, or to a place (eg for a pizza). Emergency calls have value. There is some value in "quality", but it's not really enshrined in simple network QoS.

But connecting two friends or work colleagues together for a basic phone call does not involve any provision of value beyond the access layer. In fact, the restrictions of the phone call format may detract from the value of the conversation.

I know the analogy is not perfect. But the "speaking agent" model is going to become ever more niche. We only think of it as a service because of the ancient history of the telegraph, and then the use of manual intervention to connect you to your recipient's line.

(That model still exists in some instances with personal assistants "Hello, is that Mr Bubley? I'm connecting you to Mr X now". Ironically the only time in recent memory that's happened to me has been when speaking to representatives of the ITU before the recent conference in Dubai).

While telecom users might sometimes be lazy in switching, they're not stupid. Trying to eke out the last bits of growth in voice telephony-agent business makes sense, but blaming it on those "dastardly OTTs" is completely missing the point. Voice communications is already moving to cheaper/richer applications (eg Skype) and it's about to become embedded in the web (via WebRTC - I'm speaking at the conference in SF in Nov).

The idea that regulators (who are usually tasked with improving value to consumers, as well as competition fairness) will happily sustain a basic speaking-agency model long into the future are over-optimistic. Once ministers and regulators pick up on the idea that "voice" doesn't have to be a service, but can just be a function or application, the world will likely change rapidly. We will see efforts to decouple the valuable aspects (eg emergency calls) and provide them perhaps as a standalone service or basic citizen right.

If telecom operators want to continue to fight their corner, they need to:

  • Think deeply about the "agency" dilemma. Are you really just brokering (and metering) peoples' conversations?
  • Work out how to add real value to conversations. This will need careful segmentation of *why* people make calls, and look for unmet needs for specific contexts. It's ridiculous that we use the exact same product for a sales call, as we do for calling a relative overseas.
  • Promote the use of telephony and other voice services much better. 
  • Stop focusing myopically on the OTT bogeymen on the supply-side, or you'll miss the real elephant in the room, which is falling demand for an ageing and clumsy product.
  • Decouple the number from the service and access. While there is still some value in E164 numbers, the perspective of number=identity is extremely flawed. 
  • Review how your accountants do revenue allocation of voice telephony from bundles. I believe that it is often massively overstated to begin with
  • Understand the difference between voice & telephony, and between services & apps/features
  • Ensure your billing & OSS systems are up to the challenge of new business models - freemium, sponsored, differentiated or affiliated services etc.  Stop thinking that the "minute" is the fundamental unit of telephony.
  • Tell regulators to stop thinking in minutes, and to understand that the very nature of voice comms is changing.
  • Warn your investors & explain what you're doing (if you think Utility valuations are bad, have a look at Agencies of various types)
  • Get up to speed on the threats & opportunities from WebRTC. It's probably the most disruptive thing I've seen in more than 5 years. I'm doing a presentation on what it means for telcos, and also sitting on a panel at this conference next month.
There's a ton of other stuff I could add here. But it's critical to avoid the complacency from some of my rival analysts that it's all fixable, if you just hang on to the number and do some clever bundle-pricing. That is pure wishful thinking.

Martin Geddes & I did a Future of Voice / Telco-OTT workshop in London last week. Sign up to both his & my mailing lists and we'll let you know about our 2013 schedule soon. Or if you'd like to arrange a private workshop or brainstorm session, contact me at information AT disruptive-analysis DOT com .

Friday, October 12, 2012

Essential reading for ITU World attendees: Ubiquity is EARNED not ASSUMED

I've written before about the “Death of Ubiquity”.

In the run-up to next week’s ITU World summit in Dubai, where I’m on two panels on ecosystems & future service platforms, I thought it was a theme worth revisiting. I’ve also just spent two days at the VoLTE andWebRTC joint events in Paris, which has given me further food for thought.

Many operators, industry associations and vendors refer to the “ubiquity” of the PSTN, and the universal “reachability” of E164 telephone numbers. This is also sometimes linked to the history of universal availability of emergency calling as well. This ubiquitous nature is taken to be, unarguably, a benefit, and something to be maintained at all costs.

I’ve identified three important implicit assumptions that don’t bear scrutiny:
1) The assumption that the phone call – and phone number – will remain our primary – and ubiquitous - means of communication in the future
2) That IP-based versions of those services (generally IMS-based) are somehow “entitled” to ride on the coat-tails of their ubiquitous circuit-based predecessors, and will also inevitably become ubiquitous. 
3) That the classical model of each telecom company owning its own (commodity) switches and applications, and interoperating/federating them, will remain central in future

In a week in which we learned that Facebook has passed a billion users, yet Verizon has pushed back commercial VoLTE rollout to 2014, there is clearly a reality-check needed about what will be “ubiquitous” and when.

Voice is more than just phone calls

Let’s start with the phone call. As Martin Geddes and I point out in our Future of Voice workshops** , it was a pretty decent idea 100 years ago – but it really isn’t a great reflection of how we’d ideally like to communicate in spoken form. It’s interruptive, rather unnatural and suffers from what is termed the “Hegemony of the Caller”. It’s fine for certain types of interaction, but looking increasingly poor for others – which is why many people now pre-schedule calls, or “escalate” from IM. (It also has numerous other limitations which I can discuss on request). We’re seeing the emergence of new forms of voice communication (eg ambient voice or app-embedded communications such as in-game chat). Some markets like the UK have now gone past “peak telephony”, with minutes-of-use falling as we discover other mechanisms that work better for certain functions.

**(a few spaces left for Oct 23-24 in London - sign up now!)

That said, I don’t expect the phone call to disappear quickly or entirely (although revenues will). What is less certain is whether the phone number will endure. I’m increasingly irritated that web-based forms expect me to enter a phone number, when often I’d rather be contacted by the service of my choosing such as Skype (or not at all by voice). As technologies like WebRTC start to turn voice communications into a function of certain websites or apps, rather than just a service, we’ll have ever more voice interactions that don’t need a numerical identifier attached to a subscription or SIM card.

Cut the number?

I don’t think it’s feasible yet, but I’d quite like to “cut the number” when I get the chance. I only use the phone for a few functions, and my call minutes diminish year on year, despite getting more in my mobile or fixed broadband allowance. It’s a hassle having to port the mobile number whenever I churn, as it’s linked to my account/SIM.

Overall I think that numbers could still remain quite useful – but only if they can be totally decoupled from access line and SIM. Increasingly, we will have multiple access providers anyway (especially as we use multiple devices and assorted 3rd-party WiFi connections), so there is ever-less argument to have a single “master” access against which everything is tied. (Separately, I think that WhatsApp and Viber are storing up problems by using the number as an identifier for their services).

But today, phone numbers and phone service are pretty ubiquitous, I agree. They grew up in an era in which there were no alternatives, offered good reliability, extra features like emergency connectivity, and has served us well and gained popularity.


This means that the PSTN has *earned* its ubiquity. Billions of people have seen it to be good, and bought into it. Mobile telephony (and SMS) has gradually usurped fixed telephony and extended its reach.

Fake it till you make it?

The problem is that the IP-based successors of telephony – such as IMS-based VoLTE – have conspicuously not earned their ubiquity yet. Some in the industry are assuming that will happen in the future, but it’s far from obvious. RCS is even further from being “entitled” to ubiquity as it doesn’t have a circuit-switched predecessor as heritage (no, it's not SMS 2.0).

That VoLTE and RCS5 are being combined by some operators undermines this assumed “right” even further – it has not been given a “mandate” by end-users yet, and in the new world of choice it is wrong to pre-suppose that it necessarily will.

Some vendors and the GSMA are taking a stance of “fake it till you make it”, in order to scare recalcitrant operators into adoption against their better judgement. But CFOs don’t want to invest in expensive new systems to service a declining market.

It’s indisputably clear that such operator-based services will not be the only games in town. Indeed, at this point in time, Facebook, Skype and WhatsApp are all more “ubiquitous” than IMS-based services. Even for fixed telephony, IMS-based VoIP solutions compete with simpler NGN-VoIP, 3rd-party services such as Vonage or Skype over “naked DSL”, and of course circuit telephony, which is still leading after 10 years of grindingly-slow substitution.

While many in the industry claim that so-called OTT players are “silos” or “islands”, that is neither accurate nor relevant. In its current state, it is IMS that is the silo, albeit one managed by an arguing committee rather than an individual company. Not to mention that there are many ways by which Internet-based services can and do interoperate – not all the time, or for all examples, but it is a trivial problem where there is demand. (Indeed, email is the best example of an OTT communications application which interoperates perfectly).

Which brings us to a more important issue: it seems abundantly clear that users positively like silos. (Note to regulators/ministers: users are also voters). By and large, people don’t seem to mind that Twitter or Facebook are run by individual companies, and they have plenty of choice if they do mind. 

(Edit: It's also clear that users don't always mind about variable quality or reliability either. But then you already knew that, if you'd witnessed the original uptake of patchy/drop-ridden cellular telephony. I will address the wider issue of QoS, QoE and Net Neutrality in another upcoming post)

No, you can't reach me

This supposed issue of “ubiquitous reachability” is also a chimera. Increasingly, people don’t want it. They want something much more nuanced – easy reach by some people (friends, clients, colleagues), slightly more difficult reach by others (loose contacts, who ought to make a bit of effort, as a filter), and no reach at all by others (eg telesales). Facebook, LinkedIn and various other social networks get this – they build in ideas like “mutual contacts”, contact requests, “how do you know X” functions and so on.

In a busy, networked, multitasking world, we simply don’t want ubiquitous reach.

The problem is that IMS proponents, most vendors & operators, and industry bodies, never bother to think about behavioural psychology, or social anthropology. They develop technical standards based purely on engineering principles, not human ones.

We now have enough advanced technology that we can engineer pretty much any form of communication that we want. So telecoms companies need to start with “want” not “engineer”. It is conspicuous how few IMS, RCS and “ubiquity” advocates mention end-users, or talk about actual behaviour and preferences, rather than how they’d like the world to work.

Ubiquity might occur again in telecoms, but it will be earned, not assumed or mandated.

A federal imperative?

Once you understand that, you understand why the federation approach to telecoms also fails. Not only does it take far too long to evolve – and with too many compromises based on committees – but the underlying economics are bunk as well. Federation of services means that each operator produces, distributes and sells the same commodity product. You can call these “dumb services”. No other industry has 1000+ manufacturers of an undifferentiated commodity with falling prices and zero shipping costs.

Federation may occur after services are successful, and the owners/users think that there is a good rationale. They will use tools like SBCs and other gateways, which are getting ever-cheaper and more powerful. Federation is not a starting point. You can federate from a position of strength, not in anticipation of it, or else you risk creating a brittle, inflexible, slow-moving bureaucracy which is incapable of backtracking when it makes a mistake.

For all their size and power, companies like Facebook and Microsoft and Google have changed direction – often very rapidly – when faced with challenges from their users (or competition authorities). I can’t remember 3GPP or GSMA ever doing the same.

Again, the difference revolves around users. There is no mechanism for end-users to force a change in “ubiquitous” services, especially if they are somehow viewed as special. Regulators can play around with pricing and a few other issues, but cannot easily drive changes in the underlying technology or characteristics of the services themselves, especially in a short timespan.

For Facebook, every change it makes risks key people – ie social “hubs” with lots of friends – abandoning the service and switching allegiance elsewhere, potentially taking hundreds of people with them. It is those individuals, if anyone, who holds the “hegemony” – much like real life, where the most popular and connected people determine success or failure of restaurants or theatres or fashion brands.

There is no path for end-users to petition the 3GPP to change the nature of deep-packet inspection, or the role of SIM cards. For federated services, churning doesn’t help, because there is no competition at the basic layer of service features and capabilities. You have to take what you’re given.

Unsurprisingly, in a world of choice and crowd-sourced product direction, this is not popular any more. Users are rejecting federated services for better, more-tailored and often free/cheap alternatives, delivered via open Internet access and apps. The more egregious moves (eg on privacy and net neutrality) are sometimes ultimately tackled through the ballot box, although the obfuscating noise of politics and lobbying makes that a tricky path - and which of course get conflated with a hundred other non-telecoms issues.

Dial 911!

Emergency calling usually rears its head at this point in the argument, as an example of the “greater good” that customers are only aware of when they really need it. It is used as excuse for continuing the controlled, centralised, federated-telco model.

I think that is a non-sequitur.

I agree that good emergency communications is a must. It also needs a bottom-up rethink. Nobody sensible would suggest being able to call 911 from inside voice chat in World of Warcraft (“Police? My sword’s been stolen”). But nobody sensible would say it’s a bad idea to allow SMS’s to emergency services either, yet 20 years on it’s still not possible in most countries.

We need to start thinking about decoupling emergency comms from the telephone network and look – in the round – at better evolution paths. For $50-100bn, we could probably find a global 5MHz of spectrum, build dedicated networks (high range/low capacity is fine given the loads) and give every person on the planet a cheap cellular emergency keyfob or bracelet. Alternatively, banging together the heads of Microsoft, Google, Apple, Facebook AND some telcos could yield a rich and extensible “Emergency API” that far exceeds today’s voice-only 911.

It shouldn’t be the emergency tail wagging the broader communications dog. In fact, the tail shouldn’t be attached to the dog anyway, solely because of a 100-year old legacy technology and industry structure that is breaking up. Let’s use the ITU event next week to start thinking about imaginative alternatives for the next 100 years more seriously.

So, overall I think it is time to rethink this term “ubiquitous”. What ought to be ubiquitous is the right for the individual to be contacted primarily on their terms, not those of whoever is trying to contact them. If we decide as a group that we still want a lowest-common denominator telephony service in perpetuity, then optimise GSM and CS voice even more, for maximum efficency and lowest cost and power consumption. 

GSM, unlike VoLTE, RCS or IMS, has earned its ubiquity. IMS and VoLTE might succeed and become ubiquitous eventually (RCS certainly won't). But the industry shouldn't assume or pretend that it's inevitable, because it isn't.