Goodbye retrovoice; Hello neovoice

I’d like to share some initial thoughts on the “great transition” in the voice market from a telephony-centric world into a post-telephony world.

Let’s first take a moment to celebrate the success of telecoms voice services. Telephony is the original and best distributed computing application. It has successfully funded the development of the access infrastructure on which the Internet now rests. Telephony has transformed society and commerce in a positive way. There is nothing here to be ashamed of, even as we transcend the limits of the past. Go hug a phone company today, while you still can.The danger is that we unconsciously perpetuate the constraints of this retrovoice world as we head into neovoice. The unspoken assumptions of the retrovoice past are:

  • It starts and ends with telephones, or devices that look and behave like telephones.
  • Being able to reach everyone, all the time, is how you create value.
  • The caller is the powerful party, since they are paying. They choose the timing and topic.
  • What matters is creating federated and interoperable services that adhere to standards.
  • There is a one-size-fits-all solution to voice, and it is delivered as a service.
  • Call quality is everything: blocks, drops, buzzes and fuzzes are no-nos.
  • Nothing is recorded by default; retaining the conversation is an exception.
  • Voice is delivered over special-purpose networks.
  • Money passes between network operators based on call volume and origination patterns, and cascades from the application to the infrastructure.
The world we are heading towards is pretty much the antithesis of this:
  • It starts and ends with software applications running on devices that are computers of all sizes; mobile apps and browsers will rival dialler applications as points of origination and termination.
  • Being reachable and interruptible by everyone, all the time, is a nuisance and has negative value. Communication is partitioned, contextual and consensual.
  • The callee is the powerful party, since their attention is valuable. They choose the timing and topic. “Calls” are replaced by “invitations and offers to talk”.
  • What matters is creating innovative services that solve a problem, standards be damned if necessary.
  • Voice has to fit the needs of its context, and it is delivered as a feature or function of some other application.
  • Call quality still matters enormously, but not at the price of lacking integration with our software, or at the cost of massive infrastructure duplication.
  • Everything is recordable by default, and one or both parties will choose to do so a significant proportion of the time.
  • Voice is delivered over general-purpose networks (and that generality will evolve to be more voice-friendly).
  • Interconnect is based on an Internet-style free-for-all. What happens between the user and the backbone is between consenting adults, and is only regulated at the extremes.

Both ways of seeing the world have a common theme of human social interaction, but the nature of that human discourse is fundamentally changing due to smartphones, broadband and social media. Rather than look for winners and losers within the current framing for voice, we should expect a great deal of market entry and exit over the next decade — even in markets that have been stable for a very long time.Ten years ago, I started a blog called Telepocalypse, which forecast exactly what is happening now due to ‘over the top’ players displacing core telecoms services revenues. I shut down that blog last week. We’re at the end of the initial “destructive” phase of the creative destruction — and it is time for the “creative” part to take over.  To thrive in this neovoice world, you will need good maps and guides. Navigating by your retrovoice map will result in you getting very lost.

To keep up to date with the latest fresh thinking on telecommunication, please sign up for the Geddes newsletter

Join the Newsletter

Get your fresh thinking with free Geddes newsletter
  • This field is for validation purposes and should be left unchanged.