Building corporate incubation with Monty Hall

A few days ago, I was discussing how wrapping your head around the Monty Hall problem could help you better understand the importance of failure in innovation. If you take it a step further today, we could argue that most corporate incubation programs should be build around this probabilistic calculation.

If you remember, the Monty Hall problem explains the very counter-intuitive probabilistic calculation of a game where Monty offers to pick one of three doors (1, 2, 3) to have a chance to open the only one that hides a brand new car.

As soon as you choose door 1, Monty who knows where is the car, will open up one of the two other doors where he knows the car isn’t (2 in our case). You’re now left with doors 1 and 3 unopened and Monty asks if you’d like to switch your choice to door 3…

Which you should, because you still have 1/3 chance to have the car on door 1, but since Monty knows where is the car and reveals that 2 is empty, then door 3 has now 2/3 chance of hiding the car.

Yes, I know you’re still confused by that.

It’s OK, the maths are quite exact and that was my point: understanding why you are confused by these maths could help you adjust to what is at play in innovation, and why failure is so important.

Imagine now that you would apply this Monty Hall metaphor to how you design and run a corporate incubation program.

Instead of just 3 doors, you’d have 10 of them. But they aren’t doors anymore, they are now innovation projects, or as I’d rather say, business prototypes.

The project 1 is the one that is chosen by your vanilla innovation pipeline as the best candidate and it will get most of the company’s resources for the next three years. And your corporate incubation program will be Monty Hall: it will explore alternative projects (other doors) to see if they hold any promise (is there a car behind?)

But of course innovation, like many things in the universe, runs on money. So you can’t explore all possibilities, you just have 7 alternative projects that you can run in a fast and dirty way (that would mean ‘in startup mode’), and two others possibilities will remain unexplored.

This is not a perfect Monty Hall setup now, but close enough.

If you understand how it works, the project number 1 as being the project that makes the more sense for your company will be hard to change or worse, to kill. Most of the teams are committed to it, you have business plans, sunk costs, and KPIs connected to it. You might even have broadcasted it to your shareholders and the press.

Given your approach to innovation is way too deterministic, your underfunded corporate incubation program can totally be designed as a scout in the future of the market, that just like Monty Hall, would open alternate doors and check for “cars”.

In which case, every project failed in your corporate innovation program is a tremendous news because you accumulate asymmetric odds to spot “where the car is” and build unfair advantage.

That requires two critical things though:

1. Understanding that this is a Monty Hall situation, and every project failed in the incubation program DO NOT REINFORCE YOUR INITIAL CHOICE, but rather should push you to invest more in the possibilities you didn’t touch yet.

2. If you do find “a car” in your incubation program, well, you found a car and should UNDERSTAND HOW TO LET GO OF YOUR INITIAL PROJECT!

These two things are terribly difficult for one single reason: they play on your innovation culture, or lack thereof. Innovation is everything but business as usual for most of us.