The theory of innovation, part 2: Evolution
How innovation works, and how you should design your business around it
This is a weekly-ish newsletter from Al Cottrill and Class35. It’s mostly an informal outlet for the theory and esoterica that underpins our practice. We hope you find it interesting.
Welcome to the second edition of Growth Complex.
Last week I wrote about how there is a common pattern to successful innovation wherever we see it.
When we see consistent patterns, it suggests there are some underlying dynamics at work. Which in turn suggests two tenets:
If we want to understand innovation, we need to understand the underlying driver of these patterns
It’s best not to fight these dynamics if you’re trying to do innovation effectively
But to understand why innovation has such a common structure to it, we have to go a bit deeper before we come back out. So this is the second last of three key ideas, before we get into the implications for innovation in organisations. I think.
Part 2. Innovation and Evolution
“Evolution across a population is nature’s trick for mastering uncertainty” - Eric Beinhocker
The hint is in the title, but the reason that innovation exhibits such a common pattern across these domains is because markets are a class of evolutionary system. And what you’re looking at is the structure of evolution.
We tend to think of evolution in terms of ‘what happens to animals’, but really nature is only one class of evolutionary system. Capitalism happens to be another one.
If we think about how species evolved from a few simple bacteria to the sheer wealth of biological forms that exist today, it was evolution that did it. In the same way that if we think about how simple the economy was back in the 1800s, when perhaps there were a few types of each product in the local store, to the range of SKUs you can get on Amazon today, we have seen a similar massive proliferation of designs. And the underlying dynamic is the same: evolution.
Eric Beinhocker (in perhaps my favourite book, The Origin of Wealth: Evolution, Complexity, and the Radical Remaking of Economics), describes evolution as: “an iterative process of experimentation, selection, and then amplification of things that work”
We don’t need to go into the mechanics of complex systems and evolutionary theory to know why this is (that might come in some later notes). But for now we can just skip to his following explanation:
Evolution is an algorithm; it is an all-purpose formula for innovation, a formula that, through its special brand of trial and error, creates new designs and solves difficult problems. ... In short, evolution’s simple recipe of “differentiate, select, and amplify” is a type of computer program—a program for creating novelty, knowledge, and growth. Because evolution is a form of information processing, it can do its order-creating work in realms ranging from computer software to the mind, to human culture, and to the economy.
So evolution in the natural world works something like this:
Differentiation: Mutation of genes, creating ‘design diversity’ of animals with slightly different characteristics
Selection of fit genes under competition for resources
Amplification of fit genes, via reproduction
But in the capitalism examples from the previous post, we can see evolution working the same way:
Differentiation through a wide variety of different designs
Selection of fit business models through competition for customers
Amplification through growth and reinvestment of profits
So less fit designs fail, and over time (hopefully) viable designs emerge.
Again, apply this back to the start-ups example. The market incentivises massive experimentation through the creation of different business model designs, and then through selection against the fitness function of the market (customer choice) within a competitive environment. Most fail and a few viable businesses emerge. It’s this process of evolutionary search – of experimentation, selection, amplification – that is an (or the only) effective process for consistently finding working business models.
If we want another example though, we can look at the way a child’s brain develops. Toddlers may not be as smart as adults, but they are far better learners. And while you might assume the brain gets more complex as it grows, in some ways, the opposite is true. As investor Michael Mauboussin writes in his (excellent) book More Than You Know: Finding Financial Wisdom in Unconventional Places:
“From a child's birth to the age of three, there is a huge increase in the number of synapses - connections between neurons - in the brain. In fact, a toddler has roughly 1 quadrillion synaptic connections, twice as many as an adult. Children have brains that are more active, more connected, and more flexible than those of grownups.'
But following this synaptic proliferation is a significant pruning process. Through experience, useful synaptic connections are strengthened, and those that aren't used get pruned (known as a Hebbian process after psychologist Donald Hebbs). Estimates suggest that young children lose approximately 20 billion synaptic connections each day." This process fine-tunes the brain to survive in its particular environment.”
You can see the similarity with the pattern of innovation from last week: proliferation of options and pruning back to a ‘fit’, or viable, design.
But Mauboussin asks probably the most important question, given that this process of synaptic overproduction and pruning is incredibly expensive in terms of wasted neural components and energy cost: “Why has evolution allowed this wasteful process to persist?”
Nature is pretty smart. Models of neural networks show that the overproduction/pruning approach is very flexible and more reliable at preserving information than a feed-forward network. Starting with lots of alternatives and winnowing down to the most useful ones proves to be a robust process, even though it appears quite inefficient.'
In essence, If you are in an environment of relatively high certainty, as in you are pretty sure the thing is going to work, you can just plan it/design it and do it/build it. But when you can’t predict what will work, the most reliable way forward is to adopt the evolutionary algorithm: start with a lot of alternatives and, through trial and error, prune the options down to the most useful. That is: why predict, when you can evolve?
It’s remarkable when you think about it. Why should capitalism work the same way as natural selection and work the same way as a child’s brain development? The answer is because they’re all classes of complex systems, and if you’d like to go deeper I’d recommend the Santa Fe Institute’s research, but we can leave it there for now.
What I care more about is the implications for all of this for helping organisations build successful innovation capabilities.
Where to from here
These dynamics are one of the main reasons capitalism outperformed socialism as an economic system for the creation of prosperity (we won’t get into the critiques here for now).
In essence, a major reason capitalism outperformed socialism was because it harnessed the evolutionary forces of the market. It didn’t try to plan centrally and predict what would work, instead it used the incentive system of the profit motive, plus the evolutionary system of the market for experimentation, selection and amplification to create substantially more innovation. And in doing so, it creates more wealth in terms of solutions to human problems and associated quality of life.
London Business School professor, Costas Markides’s work on successful strategic innovators identified a similar insight, and its implications for successful innovation within organisations:
In the socialist system, somebody tries to decide beforehand what is a good idea and then allocate resources accordingly. In the capitalist system, on the other hand, no central coordinating mechanism exists. Nobody tries to outsmart the market. Instead, multiple bets (i.e., initiatives) are made, and through some selection process (which is not necessarily efficient), winners and losers emerge. The capitalist system is certainly wasteful, but it is the best engine of progress so far designed.
What characterises successful strategic innovators is their ability to incorporate the essential features of the capitalist system into their organisations. They have purposefully created internal variety (even at the expense of efficiency) and then allowed the outside market to decide the winners and losers. Thus, within many strategic innovators is the harmonious coexistence of often conflicting features (i.e., variety) that are continuously tested in the market and, if found wanting, are eliminated without too much debate.
So we get a choice: an innovation system that operates more like ‘central planning’ and tries to pick winners. Or an innovation system that harnesses evolutionary dynamics, and recreate them inside the organisation (though this can’t be a straight lift-and-shift mind you).
Obviously I’m going to suggest the latter is the better choice, based on all the evidence. As I mentioned at the start, it’s generally a bad idea to design systems that conflict what look like natural ‘laws’.
But, most organisations get this wrong at a fundamental level. And in the coming posts I’m going to explain how.
Next week I’m going to introduce the last of three concepts that underpin the main reason innovation efforts in organisations fail. Then we can move onto the insight as to why, and how we can potentially build them more effectively.
If you have any questions, comments, corrections, disagreements, or just want me to expand on something, let me know. Feel free to jump into the comments section below, or you can find me at firstname.lastname@example.org. And if you like it, please share it.
Thanks, and see you next week. I hope…
And read part 3 here.