Have you heard of the Innovator’s Dilemma? It’s a form of Game Theory in action that helps explain why early-stage startups can go from 0 to $100mn in such a short amount of time. David vs Goliath stories don’t all end with glossy IPOs and founders with their faces on Forbes. Many times they end in a small, but lucrative acquisition at an early stage. This knowledge is a useful arrow to have in the quiver when shooting for the stars as an early-stage startup.
Have you heard of the Innovator’s Dilemma?
On one hand, you have the incumbents – the market giants who hold the majority of market share. On the other hand, you have the disruptors – the companies who are creating new products, services, and categories that want to take the incumbents down. History is littered with successful disruptors, but the question is always, how do you know when the market will change?
For every Kodak there is an incumbent that sustained its dominance and profitability for the last decade through incremental innovations, strategy shifts or M&A Activity. For every Uber there are hundreds of startups that lost billions of dollars in investment capital, all thinking they were the next big thing.
Bayes Theorem – Calculating Probabilities
In the opening paragraph, we mentioned the application of Game Theory, which is a subject for another post. But at the heart of Game Theory is the need to estimate probabilities, and that’s why we will introduce Bayes Theorem as a way to estimate probabilities as it relates to markets. That way, we can put the Innovators Dilemma into context and try and create a framework for decision-making, from both the innovators’ and incumbents’ perspectives.
Bayes Theorem is based on conditional probabilities ie. if this, then that.
>P(A|B) is called the posterior; this is what we are trying to estimate. In the above example, this would be the “probability of X given that Y factor is true”.
>P(B|A) is called the likelihood; this is the probability of observing the new evidence, given our initial hypothesis. In the above example, this would be the “probability of Y factor given that X is true”.
>P(A) is called the prior; this is the probability of our hypothesis without any additional prior information. In the above example, this would be the “probability of X”.
>P(B) is called the marginal likelihood; this is the total probability of observing the evidence. In the above example, this would be the “probability of Y factor”. In many applications of Bayes Rule, this is ignored, as it mainly serves as normalization.
Bayes Theorem is a very statistical, math-heavy type of work that can involve large amounts of data and require extensive experimentation. For this reason, we see a large application of data science, where computers are able to parse through large data sets to test inferences and build forecasts in real-time. In an innovation context, this could be trying to infer whether a new product will catch fire in the market using a series of existing data points. It’s not a scientific tool used to predict exact outcomes, it’s an inference tool used to build more accurate estimates.
Detecting a Shift in the Market
How many times have we seen experts and executives make ‘bold calls’ that turned out to be disastrously wrong?
The Internet. The iPhone. Bitcoin.
Instead of hostile predictions against change, what if entrepreneurs and innovators simply tried to assess the probability that one thing or another would succeed.
In the cases where the odds seemed awful of a certain technology succeeding, the payoff to success would have been tremendous to the upside. On the flip side, a low probability of success may have given some executives a momentary feeling of comfort, only to see billions wiped off their market cap and their company itself disrupted.
This is where the theory of the hedge comes into play, and why sometimes brand-new startups and their teams get acquired for hundreds of millions of dollars. It’s not the actual value of the technology or service in today’s terms, it is the future value of said offering that may save the whole company’s future.
As described in the video above, even with seemingly endless resources, mid-to-large sized companies cannot build these technologies in-house. It is a matter of culture and incentives. Nor could any startup sell their new technologies or services to the mass market in the early stages. Instead, they sell an idea or nascent product to early adopters and hope that they are able to remain liquid for long enough to reach the mainstream market.
Innovation Strategy at an Early Stage
Going back to Bayes Theorem, you will probably be wondering how on earth one could be expected to put this funny-looking formula into practice.
The idea is not to sit there and crunch numbers but to instead understand the practical nature of it from the incumbent’s perspective. Here are some examples:
>If, for example, your thesis is right on X market, what is the probability that Y will buy you out?
ex. You can see many Fintechs entering the market, enduring heavy losses for years on end. Will Banks buy them at some point because of their core innovations? (A) or will they not be profitable enough? (B)
>If you have data to show that X Customer segment is willing to pay big money for your new product, what is the probability that Y Incumbent will be able to respond quickly enough to prevent you from stealing their customers?
ex. When you look at a company like Farfetch that created a hyper-innovative flex-inventory model, you can see that many high-fashion department stores were not able to respond to this eCommerce/New Retail trend (A) and have lost a large % of their customer base. (B)
To actually begin calculating the probability of these events listed above would take large data sets looking at previous events in relation to new markets, M&A, and how sales are impacted in old categories when new categories emerge. It is a complex array of math and data science to get real answers, but the basic premise of this type of thinking can get you out of the trap of fixating on one outcome or another without any flexibility.
As an innovator, the problem is always the race against the clock. Capital is finite for new businesses, which is why the company must reach cash positive or be on the endless treadmill of raising VC money. Many new founders include ‘acquisition by Z’ in their pitch decks as their exit strategy, but the probability of this actually occurring are very slim for even seasoned entrepreneurs. That’s why playing around with probabilities early-on, and building a deeper understanding of the market and consumer behavior can help shape strategies over the mid-term and buy precious time to reach a stage where a new business can get over the hump and sustain itself for the long-term.