Voice is the next big thing in mobile

One of the very big and obvious trends in mobile in the last couple of years has been the explosion of new messaging forms, enabled by the permissionless innovation of the internet and by the emergence of the smartphone as a social platform. So far, voice has been relatively unaffected, and the VOIP players from the desktop era, most obviously Skype, have had pretty little impact on how we speak to people. With the shift of networks (slowly) to LTE and the increased sophistication of smartphone platforms this is now starting to change.  

There is not really a ‘free’ story here in the way that there was for SMS: there is not the same vast arbitrage opportunity that there was between SMS pricing and data pricing. Generally price arbitrages like that arise when there is a big gap between the price charged and the underlying cost structure, but this is not the case of voice as it was for SMS. Unlike SMS, voice does actually have significant marginal cost to provide (though provisioning capacity is pretty lumpy), and VOIP does not make that cheaper - in fact on 3G networks VOIP tends to use several times more radio network capacity since it’s not optimized for the way the network works. On LTE (eventually, and to generalize hugely) all voice will be VOIP, but that doesn’t mean the network will no longer cost money to run. And operators are already (painfully) moving their pricing towards integrated bundles anyway, so the cope for arbitrage from voice to data pricing is pretty limited, I think (I discussed these issues in more detail here). 

On the other hand, there is a lot of scope to innovate around the actual voice experience in much the same way that we have seen around messaging apps. 

In messaging we have seen two levels of innovation in experience. WhatsApp delivers 'SMS 2.0' - it does the things that telco technologies like IMS were supposed to add close to a decade ago, but not much more, so far. On the other, we have things like Line, Wechat, Kik and Snapchat that actually change what messaging is, even before you look at the platform elements of their offerings. I suspect (but no more) that it is harder fundamentally to change what voice is than to create alternatives to a snippet of text, and so the basic voice experience might change less - the innovation, like WhatsApp, may be more about the handling and routing and setup. That is, it kills the dialer. That's one way to look at Talko - it rethinks what ‘dialer' mean when there is no DTMF (or pulse), and builds value around that.

Most of the old generation services in this space have been focused on international calling and perhaps roaming, either using gateways (that is, supplanting calling cards) or complicated SIM/MVNO setups that never went mainstream. I always felt that that approach focused on too small a problem - we tend to over-estimate how many normal people know enough people (or any people) in another country, outside very specific groups such as immigrants and MBAs. And even where people do know someone abroad, that's generally not of enough of their calling to switch all the rest of their calls to a new platform. Hence Skype grew to a very large proportion of international calls and created a whole new market for free international voice, but remains a pretty small proportion of actual, well, calls. This coming new generation might find more universal use cases.  

There are a bunch of interesting second order questions from this. Most obviously, how does it interact with the current status of the PSTN numbering system as the lowest common denominator? 

One of the interesting changes between old and new generations of social comms app is that the older generation (Skype etc) was all about interconnect with the PSTN (at least, that's where the money came from) but tended to ignore the actual number itself. Conversely WhatsApp et al rely on the PSTN number as an identify platform (a 'social graph') but have no interest in connecting to the actual PSTN itself. 

If, now, you move your voice to a social app, what happens to your phone number? Do you still use it in the WhatsApp sense, as an ID number rather like an ICQ ID? Does it still matter if it changes? Or does it linger on as something for old people and as a mailbox for boring official things (like email?). 

Next, do we think differently about fragmentation of voice than of messaging? It seems clear that we are not going to have one single point of contact for messaging systems. Everyone will have several of Facebook, LinkedIn, WhatsApp, Line and so on, with new apps coming and going, some quickly like fireworks, some finding a way to stick. But are we happy having multiple paths for voice in the same way? I can’t think of an a priori reason why it would be different, other than perhaps urgency, but who knows?

Finally, if the phone number goes then commoditising the cellular network becomes that much easier, especially in countries where strong competition means that the networks themselves are roughly comparable. We obviously have number portability already, though how fast and frictionless it is varies hugely by country, but if you could take your comms from network to network just as you can take your SIM from phone to phone then the commoditisation of the network layer would be much easier. The sticking point here is the interlocking effect of handset subsidies and contracts, but over half of Europe and a growing part of the USA is on prepay and there's clearly a move to shift contract users from direct subsides to instalment plans as well. Soft SIMs (despite all the very obvious problems) would accelerate this, if operators ever decided to adopt them, for whatever reason. 

Hence, one could propose a nightmare scenario for mobile operators in which your voice contact points become as free-form as your messaging contact points have become, with lots of different possible channels, the PSTN number becoming as vestigial as email is (apparently) for today's teenagers, and SIMs (or connection plans on a soft SIM) sold from week to week on a pure price /gig basis. I'm far from sure that that will happen, certainly any time soon, but it is certainly a possibility.