Many years ago, in an earlier role, I saw two teams in the same company work against each other for almost a year. Both teams had good people. Nobody was doing it on purpose. Each team was chasing its own targets, and the quickest way to hit them made life harder for the other team. Everyone knew the company was losing because of it. But nobody wanted to be the first one to stop.
I did not have a name for it then. Much later, I came across the Prisoner’s Dilemma from game theory and realised that is exactly what I had seen. Since then I have noticed it in almost every organisation I have worked with.
The setup is simple. Two prisoners are questioned in separate rooms. Each is offered a lighter sentence for testifying against the other. If both stay quiet, both get a light sentence. If both testify, both get a heavy one. But whatever the other person does, testifying is always better for you. So two sensible people, each making the right choice for themselves, end up with the worst result for both. This stuck point is called a Nash equilibrium, after John Nash. Neither side can do better by changing only their own move, even though a better outcome was always available.
Good targets inside a team, bad behaviour between teams
When I look back, neither team was badly run. Each had a sensible scorecard. The problem was in the space between them. Nothing in either scorecard rewarded helping the other team.
This was my first lesson. You do not fix this by asking people to be more cooperative. You fix it by changing the game, which means changing what people are measured and rewarded on, so that helping each other no longer hurts their own numbers.
The second lesson came from another idea: signaling, from economist Michael Spence’s work on job markets. When anyone can claim anything, the claims that count are the ones that cost something to make. A long warranty is believable because only a company confident in its product can afford it. Big brand advertising works the same way. Only a company planning to stay around for years would spend like that.
I saw the same thing in how good leaders built trust with teams they did not manage. Saying “you can count on me” costs nothing, so it means little. What worked was helping the other team first, in a way that cost them something, before asking for anything in return. That is hard to fake.
What finally changed
The two teams did start working together in the end. Not because of a mail asking them to collaborate. It happened when leadership changed how each team was measured, so that helping the other team counted towards your own scorecard instead of against it.
I have carried this into every role since. You cannot talk people out of a bad equilibrium. They are behaving sensibly for the rules they are working under. If you want different behaviour, change the rules. Otherwise you are asking people to keep losing for the company’s sake, with nothing in their own targets to show for it.
For leaders:
- Look at where scorecards meet. When two good teams keep clashing, check the targets before you judge the people.
- Put shared goals on both scorecards. If cooperation is not measured, it will lose to whatever is.
- Help first. Do something visible for a peer team before you need something back. That is how trust gets built.
- If the problem repeats with different people, fix the system. The game is the problem, not the players.
It still surprises me that an old maths problem about two prisoners explains so well what I saw between two departments in one building. Once I could name it, I knew I was not dealing with bad people. I was dealing with a bad game. Those are two very different problems.
