This might not make much sense to all the readers here in Australia, but I’ll give it a go – an interesting thing happened in the Monday Night Football this week – the Patriots lost after failing to convert a 4th and 2 on their own 28 yard line with a couple of minutes to go. Going for it was unusual – the common knowledge is that you always punt the ball in that situation. And as Matt Yglesias has been pointing out, in this case the common knowledge is wrong. Here’s his summary:

The point of punting is that you’re trading possession of the ball for field position. Whether that’s a good trade depends in part on how likely your offense is to secure a first down if you don’t kick the ball to your opponents, and in part on how good the opposing offense is. But the better their offense is, the worse kicking the ball over to them looks. The only reliable way to stop a really good offense is to be extremely reluctant to surrender the ball to them. Against a poor defense, field position is extremely valuable since they’re unlikely to score unless they get the ball close to your end zone. But what it means to be a great offense is that you’re a legitimate threat to score from any position—you really, really don’t want an offense like that to have the ball.

The problem is that conventional thinking in the NFL is that after three downs the default should be to give up possession of the ball unless it’s a desperation situation or something else special. But most of the time teams should be extremely reluctant to give the ball up. Fourth-and-shorts aren’t that hard to convert, and field position is a lot less valuable than possession of the ball. There’s just a convention of labeling any decision to run an offensive play as “risky” that’s completely independent of any actual assessment of the risks.

This point has been supported by analysis reported by the New York Times as well – it was the correct decision, it’s just that in this case it didn’t work.

What does this have to do with innovation? A couple of things. One is that many firms do things that aren’t what they should be doing simply because everyone else does them. Like NFL coaches, they aren’t even playing the odds, they’re just following routines that may or may not be the best ones. If your firm does this, it will become a serious obstacle to innovation.

The second point is that playing probabilities definitely does not mean that you will win every time. Every now and again you will screw up. Probably not as publicly as the Patriots just did, but it will happen. Being innovative by definition means that some of your ideas won’t work. You have to come to grips with this when managing innovation. But it also means that you can innovate more effectively if you have a regular system in place that allows you to try out lots of ideas and then scale up the ones that work. Previously, I’ve called this having an innovation algorithm.

Stephen Shapiro has a really nice post on this topic too – where he distinguishes between playing safe using statistics and going for it using probability.

If a statistics-driven innovation model does not work, what would a probability-based model look? Probability tells me that if everything is equal, the more bets I have, the more likely one will be successful. The odds of 1 success out of 200 are greater than 1 success out of 20.

But how can you have more bets without diluting your effort and potential returns? The key is to learn as you go.

He then outlines an approach where you place a large number of smaller bets, then scale up the ideas that work. I think this is exactly correct. This approach allows for failure, but it lowers the cost of failing. And it increases the probability of hitting big. It’s a good example of what an innovation algorithm might look like.