People, companies and groups — who is really deciding, and what are they competing for?

Individuals have motives, and they are mixed. Economists once modelled people as calculating machines maximising their own gain. Real people are more interesting: they want money, yes, but also respect, an easy afternoon, not looking foolish in front of colleagues, and — often genuinely — the job done well.

The critical point is not that people are selfish. It's that people respond to what they are measured on. This has a name, Goodhart's law: when a measure becomes a target, it stops being a good measure. Tell a call centre that average handling time is the number that matters, and calls get shorter — including the ones that needed to be longer.

A company, a charity, an NHS trust or a partnership is a "legal person": it can own things, sign contracts and be sued. That's a useful legal fiction. It becomes a dangerous one when we forget it's a fiction. A company cannot want, fear, or intend anything. Only the people inside it can.

So when someone says "the organisation wants to improve quality", translate it: which people, rewarded how? Economists describe the relationship between an organisation and the person acting for it as the principal–agent problem. In its simplest form, the person doing the work chooses how much effort to put in, $e$, to make their own position as good as possible:

In plain words: they weigh what they'll be paid or credited for the effort, $w(e)$, against what it costs them, $c(e)$ — in time, stress, and risk. Whatever the mission statement says, that little sum is what the organisation "wants" in practice, because it is what its people will actually do. Change the reward and you change the organisation. Change the mission statement and you change the wall.

Put people and entities together — a project team, a supply chain, a partnership of companies — and a new problem appears. When a group works towards a shared benefit, each member gets a slice of the benefit but pays their own full cost of contributing. The economist Mancur Olson showed what follows. Roughly, a member contributes only when their share of the benefit beats their personal cost:

In plain words: $B$ is the total benefit to the group, $n$ is how many share it, $c$ is what it costs one member to pitch in. As the group grows, each person's slice shrinks — and at some size, sitting back and letting others do the work becomes the sensible individual choice. That's free-riding, and it's not a character flaw; it's arithmetic. Big improvement programmes with dozens of "stakeholders" and no one whose own slice justifies the effort stall for exactly this reason.

The fix isn't exhortation. It's making contributions visible, making the benefit personal, or keeping the group small enough that everyone's slice matters.

Every organisation claims one. Here is a simple test for whether it's real: do the rewards point the same way as the purpose? If the purpose is patient safety but the promotion criteria are throughput, the purpose is a slogan and throughput is the real one. If sales are rewarded on volume but the purpose is customer retention, look at the churn figures and you'll see which one is winning.

Shared purpose is genuine when three things line up: what people are told matters, what they are measured on, and what they see rewarded. Two out of three produces cynicism. One out of three produces a poster.

Competition between organisations is often healthy — it forces improvement. Competition inside a group is far more double-edged. The classic illustration is the prisoner's dilemma, a situation where two parties each do better by cooperating, but each does even better by defecting if the other cooperates — so both defect, and both end up worse off.

A simplified payoff table makes it concrete (numbers are "value to each side"):

• Both cooperate: 3 and 3
• One defects, one cooperates: 5 to the defector, 0 to the cooperator
• Both defect: 1 and 1

Whatever the other side does, defecting pays more for you personally — yet if everyone follows that logic, everyone gets 1 instead of 3. That outcome, where nobody can improve by changing only their own choice, is called a Nash equilibrium, and it is a trap. Departments that hoard resources, teams that hide problems until someone else gets blamed, sites that "win" the internal league table by pushing work onto others — all of it is this table, played out in meetings.

The way out is well established: repetition and reputation. When the same parties will deal with each other again and again, cooperation becomes the winning strategy because defection gets punished next round. Organisations that rotate people constantly, or measure teams against each other on short cycles, quietly destroy that repetition — and then wonder why silos appear.

• People act on what they're measured and rewarded on, not on what they're told.
• A legal entity has no wishes; asking what "the company wants" is a category error. Ask who inside it decides, and what they gain.
• Groups leak effort as they grow unless each member's share of the benefit stays worth the cost.
• Shared purpose is real only when message, measure and reward all point the same way.
• Competition improves organisations and, mishandled, poisons groups; the antidote is repeated dealing and reputation, not another league table.

Before you redesign a process, redesign the question. Not "what should the organisation do?" but "who decides, what do they get, and what happens to them if it goes wrong?" Get that right and the process tends to fix itself. Get it wrong and no amount of mapping will save you.