Monday, January 11, 2016

The Rhyme in Generations of Disruptive Platforms

I like the very narrow definition of Platforms suggested by @chamath. It is hard to resist quoting him in full:

“I was in charge of Facebook platform. We trumpeted it out like it was some hot shit big deal. And… [Bill] Gates said something along the lines of, ‘That’s a crock of shit. This isn’t a platform. A platform is when the economic value of everybody that uses it, exceeds the value of the company that creates it. Then it’s a platform.’ If you apply that simple methodology to any company that says they’re a platform, there are only 3 platforms in the world! You know? Windows is a quasi-platform, decaying, but still, iOS is platform, and Android is platform… these [other services] aren’t “platforms”, these are APIs, these are developer tools, which is good, and it’s a bridge to something, but it’s not a platform, so let’s stop calling it a platform, let’s call it what it is, which is a bunch of endpoints.”

I would only add that besides Windows, or rather “Wintel”, which is a Platform for the “PC”, Lintel (or Linux plus Intel) is also a Platform – albeit one that powered the server side, and which isn’t dominated by a single Company. In fact, this commodity hardware Platform built with Linux and other open source software (from LAMP to Hadoop) powered the rise of all the companies offering Cloud services, whether they built their own data centers like Google and Facebook, or outsourced them to AWS or Google’s App Engine.

Like @chamath, Benedict Evans, who discusses the move from a PC ecosystem to a Mobile ecosystem, didn’t see the need to separate out Lintel from Wintel either, and he lumps them together as the “PC.”  That may make sense from a hardware ecosystem perspective, but thinking of Wintel and Lintel as two different Platforms can help us listen better to the rhyme of history  (for, as Mark Twain has indicated, history may not repeat itself, but it certainly does rhyme.)

And with this perspective, as illustrated in the table below, there does seem to be a certain rhyme in the rise of new Platforms. It is most striking that new platforms don’t seem to disrupt the previous generation Platform. Quite the opposite in fact: they depend heavily on the previous generation, without which, they could not thrive. But they do tend to disrupt or replace the generation before last (in an “alternate rhyme scheme”, ABAB).

For example, Cloud services would not be possible without the dominance of the Wintel PC and the browser, but the Cloud didn’t replace the PC – it replaced proprietary server stacks (eg mainframes) and disrupted their business model.  Similarly, the current Mobile ecosystem relies heavily on the Cloud and its services (forming the foundation of the app ecosystem). But as Ben Evans points out, today, “Mobile” is effectively on its way to replace the PC. Again, it is not replacing the preceding generation of Platform, but the penultimate one.

Hardware Emblematic Company Client / Server Dis / Aggregators Disrupts /
Main frames / Minicomputers Proprietary IBM /
Server N/A Typewriters etc. etc.
PC (Wintel) PC Microsoft /
Client Dis-aggregator Terminals etc. etc.
Cloud (Lintel) PC Google Server Both Minicomputers/ Main frames
Mobile Mobile Apple (IOS) /
Client Aggregator PC (Wintel)
The Next New Platform Mobile? ? ? Server? Both? Cloud?

In the table above, I have also indicated a Chirstensenian accordion of aggregation and disaggregation performed by each generation of Platform. But I would note that Cloud services are both aggregators and disaggregators. They are built on an aggregated (or integrated) stack of software and hardware to deliver their services. Yet, each of those services can also be seen as a specific disaggregated service offering a sliver of functionality “as a Service.”

It is also striking that, seen this way, Platform generations swing back and forth between client side and server side dominance, like a metronome.

So if we let this music play a little longer, we could surmise that the next great Platform will emerge on the server side, based on an integrated Mobile hardware stack that offers slivers of disaggregated services which will disrupt the Cloud!

Simple enough!

This wouldn’t mean that this Mobile hardware stack will just be used in Google’s data centres to do the same thing Google’s servers are doing now — that may happen, but that won’t constitute a new Platform. New Platforms inevitably involve new services, use-cases, and value propositions. Arguably, the same way Google’s servers disrupted mainframes by starting out as less serious “toys” delivering slightly different value propositions, this new Platform will not substitute Cloud services directly, at least to begin with.

Yet if we indulge in the outrageous results of such an interpolation, we could well predict that these little devices will eventually replace Cloud Services as we know them today.

 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
(cross posted on medium)