Friday, October 09, 2015

What Christensen gets wrong on disruptive innovation.

tl;dr Not a lot.

Once again we have one of those debates raging that is taking a pop at the whole idea of disruptive innovation and once again it questions the whole predictability of it. We have the usual claims of "Is disruptive innovation dead wrong?" and shades of the Jill Lapore debate of 2014.

Alas, this won't stop because there are (at least) two different forms of disruption - one which can be anticipated and one which can't. The problem with the current debates on both sides is they treat disruptive innovation as if there is only one form.

Yes, business builds inertia. Yes, substitution occurs. Yes, the principles of disruptive innovation appear sound and Christensen looks right ... BUT ... and here's the catcha ... when we talk about disruption through product vs product substitution then we can't anticipate this change but when we talk about disruption through product to commodity (+utility) substitution then we can! We can do this because we have a pretty good understanding of how stuff evolves through demand and supply competition.

So, for example with the evolution of smartphones and a product to product substitution (i.e. involving a discontinuity with the past) such as Blackberry or Nokia vs the iPhone then we have no idea which way this would go. Any analysis is post event. It could have been Nokia or the iPhone that succeeded. No-one knew. Christensen didn't know. Just because he called the wrong side doesn't mean the concept of disruptive innovation is wrong. It isn't. It's just almost impossible to anticipate.

However, the smartphone is a value chain (which can be mapped) and it contains a component known as the operating system. The application of competition (supply and demand) means that this component will eventually evolve to a commodity, especially if someone drives it there with an open approach. Be wary of Geeks bearing gifts.

Hence with Android vs iOS then we already know the OS is becoming commodity, that Android is driving it there and we know that Android will dominate. In fact this has already happened but we knew this ages ago. We could anticipate this change and its effects and yes, eventually it means disruption of the past models and systems based upon it. 

This will happen no matter how well Cook plays the supply chain, no matter how much people cheer for Apple. Competitive forces and the Red Queen effect are against Apple on this. They can hold out but eventually they'll end up adapting (i.e. Apple on Android) or they'll corner a niche space.

The principles behind disruptive innovation are sound in my view and I can't see anything that Christensen got wrong bar one thing. The reason why we seem to be having these endless debates is this assumption that there is only one form of disruption. It's either got to be anticipatable (i.e. predictable to a certain degree) or not. Alas there's not one form, there never has been, there never will be. There are two forms - one you can anticipate and one you can't - and the ways in which you need to deal with both forms are different.

For example.

Take an act, A, that is evolving through competition forming many different instances of that act. Let us look at A1 to A2, a product to product substitution (I've shown this in mapping form in figure 1 below). Then yes, past vendors will have inertia to this change due to existing business models, they cannot know that this new instance will disrupt their existing business model, they cannot anticipate that it will happen. This is Apple vs Blackberry. 

However, in the case of something evolving from product to a commodity or more utility (shown as A2 to A3), then yes past vendors will have inertia to this change due to existing business models BUT we can anticipate such a change well in advance and what its impacts will be. It doesn't matter whether it's iOS vs Android or Utility computing (e.g. AWS) vs old product vendors of servers.

Figure 1 - Evolution

These two forms of disruption are different,  they have subtle differences in characteristics (one of which is our ability to anticipate them) and we therefore should react in different ways. I've summarised some of this in figure 2.

Figure 2 - Differences in forms of disruption

Because one form is anticipatable (the change from product to commodity / utility forms is known as a 'war' state)  then I can use a variety of weak signals (related to publication types and changes in behaviour) to identify when these anticipatable forms of disruption are likely to occur. This doesn't mean I can get the date exactly right but I can prepare well in advance (see figure 3).

Figure 3 - Future points of War (from 2014)

[For reference, if you're trying to do this form of analysis with diffusion curves or hype cycles, then you've no hope. Diffusion curves measure diffusion, hype cycles approximate hype - neither can be used to determine evolution. You'll be running down endless dead ends.]

So, let us see this in action. Let us take two things - Big Data and Immersion

Big Data.

From the weak signals we already know that the "war" is upon us (actually, we knew this war was heading our way many years ago). This means new entrants will move into the market providing more commodity / utility forms. This has already happened with the likes of Google, Amazon and Microsoft entering the field. We know the war takes about 10-15 years to work its way through, that it's not a linear change but a punctuated equilibrium, that it involves co-evolution of practice and if those changes are large enough then new forms of organisations will emerge. We also know that existing players will have inertia (there are 16 different forms) and inertia is a killer. There are also a host of repeatable patterns and gameplay including ecosystem models that can be used to spice the fight up. We also know most companies have little to no situational awareness and therefore suck at this.

So, we can say that the new entrants will start to grow exponentially. The existing players and many of their existing customers will dismiss them as not being capable. Within 5-8 years the new players will still represent a fraction of the market (e.g. 3%-5%), the existing players will still dismiss pointing to past data but may graciously state that the systems are ok for test and development (of course, in production they'll claim you'll need the vendor's product ). During this time new methods of working will have co-evolved with the more utility forms. By around 10-15 years panic will have set into the product vendors as they see their entire business dismantled as they have no time to adapt. They will be disrupted by a change which could be anticipated in terms of what and roughly when. See figure 4

Figure 4 - Big Data


Now let us turn our attention to immersive tech (including both virtual and additive reality). Let us consider Oculus Rift and the HoloLens. Will the Hololens disrupt the Oculus Rift and related VR equipment?

Well, it might do but we have no blinking idea. This is a product vs product substitution. It could go either way. It might not even cause disruption. There is nowt we can anticipate about this (see figure 5).

Figure 5 - Oculus Rift vs HoloLens

We have no idea which way this will go - it's like Apple vs Blackberry. Any claims are guesswork or post event analysis. The best you can do (see figure 2) is have an adaptable culture and lots of very tight looped horizon scanning. There's some gameplay available but nothing like what you can do with more anticipatable change.

Of course, let us say HoloLens does disrupt the market. I can say that by 2025-2030 then we're going to see more commodity forms of immersive tech and this will change the market again most likely led by new entrants as the winners of the product wars will have built inertia due to past success.

This is the point, there are two very distinct forms of disruption. It's not all the same despite everyone treating is as such. Alas, people ignore this.

So in the meantime, we will continue to hear this fairly pointless debate rage on with both sides giving examples and counters until someone starts to consider that the reason why disruptive innovation seems to show polar opposite characteristics of predictability is there's more than one form. 

We've covered this many times before over the years, I sound like a stuck record and so I'll stop there.