If exploratory projects take twice as long, you get half as many of them, which is half as good, right? It’s actually much worse than that and here’s why. This is a story of innovation and incentives and unintended consequences, but first it’s a story about poker.
In a poker tournament, every player puts in a dollar and gets some chips. When one player has all the chips they get all the dollars. (The real story is more complicated, but this will do for now.)
If these are the only rules we have a very boring, very long tournament. The optimal strategy is to wait for the best hand (aces) and then bet all your chips. Everybody else knows that’s your strategy, though, because that’s their strategy too, so they fold.
We need a way to incent risk taking, so that the tournament is done before all but one player dies of old age. Enter the forced or “blind” bet. Each player takes turns putting out a one-chip bet.
Let’s say I was just forced to bet blind. Now you can play good-but-not-the-best cards because you might win the chip I just put out. If I don’t play my cards, eventually I’m going to run out of chips, so I am going to play more of my not-the-best cards. Other players know this, so they will risk some of their chips too. Now we have a game!
As the tournament goes on though, things start to slow down. Players who have accumulated a lot of chips are less likely to play, since they can survive more blind bets. It’s still going to take too long for one player to win all the chips. We can’t have a tournament that gets more boring the longer it lasts.
To keep things moving, we need another incentive for risk taking. Every hour we are going to take away half of every player’s chips. If a player only has one chip left when this happens they are out.
Now even the players with lots of chips are forced to play (take risks with) more hands. What looks like a big stack of chips now is going to be much smaller in a couple of hours. The winning player is going to have to double their chips every hour on average.
Now we have a tournament that starts interesting and stays interesting. Mission accomplished! What does this have to do with innovation? Let’s think about strategy.
Small, Medium, Large
There are (roughly speaking) three flavors of player strategy depending on how many chips they have stacked in front of them.
- Large. If I have lots of chips, then I have maximum freedom. I can play hands that have small chance of winning a lot of chips (say 1 time in 10 I win 20 times as many chips as I bet). If I lose this particular hand, I still have lots of chips.
- Medium. Now I have to be more conservative. If 1 time in 10 I only win 5 times as many chips, then that’s a losing play. Now I face a dilemma — I need to wait for a good hand that is likely to win but I have to play something because those blind bets are eating my stack and the hour is almost up.
- Small. If I don’t have many chips I need to take whatever risk is necessary to double my stack. If I don’t then I’m out for sure. If I get a decent hand I’ll just bet all my chips and see what happens. The smaller my stack, the lower my standards become for “decent”.
Summary (So Far)
We wanted a poker tournament that would be exciting start to finish. Blind bets and escalation led players to take risks, but the number and quality of those risks changed depending on how many chips they held.
Innovation, right? We’re going to talk about why delay is so very very damaging to innovation, right? Right. Here is the analogy.
Let’s say our company has a responsible performance review practice. Every six months every programmer is going to be judged by what they accomplished in the preceding half year. Makes sense, right? We don’t want people just loafing, or even working hard on stuff that makes no difference.
The outcome of programming projects is highly variable and unpredictable. Little changes can make big differences. We would like to incent risk taking so we capture the value of those little changes.
At first projects are short. A programmer can try something for a week. If it works, great, if not, move on. At the end of six months, they have a portfolio of experiments, some of which have worked out. All is well.
The company grows based on the fruits of those experiments. Sensible sources of delay creep in. Legal review. Content review. Architecture review. Marketing review. Design. Research. Now the experiment that would have taken a week takes two months.
The incentives have changed. Rather than be excited about all the possibilities, the programmer has to worry about the chance that none of the three experiments they have time for in this review period will work out. Risk taking drops. Programmers stick with low-risk/low-reward tasks longer because the alternative is high-risk/no-reward.
The company grows further and more sensible sources of delay creep in. Now programmers have time for one experiment per review cycle, maybe even one experiment for two review cycles. The incentives call for even fewer risks to be taken. Programmers stick with safety. If there is no safe project, they go with whatever looks reasonable.
These three risk-management strategies are analogous to the tournament strategies. Short projects are like a big stack of chips. Quick, playful experiments make sense because something is bound to pay off. Longer projects are like a medium stack. Any project undertaken needs to be pretty certain to pay off. Projects longer than the review cycle are a short stack — the first reasonable project has to be made to pay off (or be seen to pay off).
It feels cynical to say “or be seen to pay off”, but it matches my experience. As delay grows, reality becomes less important and appearance becomes more important. I have come to have compassion for myself and others in such situations, though. We aren’t bad people, we are responding to our incentives.
The cost of delay in choking innovation is insidious. From outside the company looks to be doing fine. Smaller wins times a larger market equals continued growth. Invisible are all the little roads not taken that could have turned out to be superhighways.
People say, “We’ve picked all the low-hanging fruit,” as if innovation used to be easy because there were so many opportunities and now there aren’t. Maybe that’s wrong. Maybe the opportunities are there but the incentives aren’t.