This essay is about reward systems, and how they quietly shape human behaviour – often in ways their designers never intended.
Experiments in Motivation
In the early 1970s, psychologist Edward Deci ran a series of experiments involving people solving simple puzzles. They were based on the Soma Cube – think 3D Tetris with a bit of Lego thrown in – where four-block pieces fit together in various ways. Participants were challenged to recreate a prebuilt configuration shown to them.
They were split into two groups: those that were rewarded financially for completing the puzzles, and those that weren't.
In the initial experimental run, things panned out much as you might expect: getting paid improved performance. The experimental group solved more puzzles.
But this wouldn't be a very interesting essay if that were the end of the story. Fortunately for me, it isn't. Because when the payments stopped... the group that had been paid lost interest in the puzzles more quickly than the group that had never been rewarded.
The reward had changed the nature of the activity. Once the payment disappeared, so did much of the motivation.
And this isn't a one-off; there are lots of studies that dig into similar sort of ground. Take Lepper, Greene & Nesbitt's 1973 study examining something called the Overjustification Effect, for instance.
In this one, kids that enjoyed drawing were split into two groups: one that got a reward for drawing, and a second that didn't. They found something similar to Deci: rewards improved performance, but once removed, the children enjoyed drawing less than they had before the experiment.
And a third: Gneezy and Rustichini's 2000 study, A Fine is a Price. By introducing fees when parents arrived late at a nursery to pick up their kids, it was hoped that the behaviour would be discouraged. But, it created the exact opposite behaviour: more parents pitched up late. By reframing the behaviour away from 'violating a social norm', and towards 'this is a market transaction with a price', parents just bought their way out of it.
Three very different studies. Three very different contexts. One underlying pattern.
The System Is The Instruction
In all these studies, reward changed behaviour – not simply by encouraging or discouraging it, but by redefining what the activity actually meant.
The lesson from these studies is not that rewards never work. It's that rewards change systems.
And because systems produce behaviour... you tend to get the behaviour your system rewards, even if it's not the behaviour you were thinking of.
What (good) Game Designers Know
Anyone who designs games either knows this instinctively, or learns it very quickly.
Players don't necessarily behave the way designers hope they will. They behave the way the system encourages them to behave. And Players are exceptionally good at discovering what the system actually rewards.
And then optimising it. Hard.
For example – if repeating one encounter yields the most experience, Players will repeat it. If a particular enemy has the best drop rate, Players will farm it. If a system rewards resource accumulation, Players will build elaborate loops for harvesting resources.
Fighting games provide some great examples. Designers doubtless imagine a rich interplay of character abilities, spacing, play, and counter-play. Certainly, my beloved Street Fighter 6 has wildly different characters supporting wildly different play styles, and it's been pretty well balanced since launch – but if a particular tactic produces the most reliable wins, players will converge on it very quickly, particularly when there's a competitive element. SF6's throw loops, for example, boil what's essentially high-speed chess down to high-pressure scissors-paper-stone, by allowing aggressive plays to get in, and in, and in again via spacing and frame timings that automatically favour them. The system rewarded – and, at the time of writing, still rewards – the tactic with consistent advantage, so Players use it.
From the outside this can look like Players abusing mechanics. In reality they're simply responding rationally to the incentives the system presents.
Nintendo's Mario Kart series offers another classic example. The item system is deliberately designed to help players further back in the race by increasing the likelihood of powerful items. The result is that positioning within the race becomes strategic in ways the designers may not have originally intended. Skilled players sometimes deliberately hang back to improve their chances of pulling stronger items, turning what appears to be a racing game into something closer to probability management.
Less racing game chaos, more... Excel.
None of this behaviour is irrational. Players are simply responding to the reward structure the system presents.
And so game designers either instinctively know, or learn very quickly, that reward systems are not neutral.
They are instructions.
Every point, badge, leaderboard position, unlockable, or progression bar quietly communicates the same message: this is what success looks like.
And Players respond accordingly.
Organisations Are Systems Too
And wouldn't you know it, the exact same principles appear in organisational life.
Many companies believe they are rewarding productivity, collaboration, or ownership. But the metrics they actually measure often reward something slightly different.
If promotions reward hitting quarterly targets, teams optimise the quarter – sometimes at the expense of longer-term outcomes.
If organisations reward appearing competent rather than actually being competent, people quickly learn to protect their image rather than take risks.
If performance reviews reward visibility, employees optimise visibility. At a primarily work-from-home company I once worked at, visibility on the internal forum system was often assumed to be a proxy for effectiveness and contribution. It wasn't uncommon to see people gaming the system in order to compensate for shortcomings elsewhere.
None of this requires malicious intent. It simply requires a system that rewards one behaviour while hoping for another.
And over time, the system wins.
The system always wins.
Rewards Change Motivation
Motivation is a rich field of study in psychology, and it offers some surprisingly practical lessons for anyone designing systems or organisations.
One of the most important is that external rewards can influence behaviour in very different ways depending on what they signal.
Rewards that feel controlling ("do this and you will get that"), or that praise fixed traits such as intelligence rather than effort, tend to reduce intrinsic motivation. The behaviour becomes something performed for the reward. And, even more impactful, the eventual removal of the reward can reduce intrinsic motivation to below its original baseline.
But rewards that reinforce competence – feedback that signals progress or improvement – can increase intrinsic motivation. They support the feeling that someone is getting better at something meaningful.
Mueller & Dweck's 1998 paper, Praise For Intelligence Can Undermine Children's Motivation And Performance, illustrates this distinction beautifully.
In one set of experiments, children praised for intelligence became more risk-averse. When tasks became difficult, they were more likely to give up. The praise had quietly changed the goal: appearing smart became more important than learning.
Children praised for effort behaved very differently. They persisted longer, enjoyed challenges more, and were more likely to improve their performance.
Again, we're seeing that the structure of a system is defining what's coming out of the other end. A system that rewards success rewards success, whereas a system that rewards effort rewards ability.
Once you start seeing this pattern, it becomes difficult to unsee, because reward systems are everywhere.
Social media platforms reward engagement, which often means that outrage spreads more easily than nuance.
Corporate metrics reward measurable outcomes, which can quietly transform work into an exercise in managing numbers rather than improving reality. Performance at work stops being about performance at work, and starts being a performance of work.
And something we've all been through, but that pretty much everyone agrees aren't a great test of intelligence or ability: exams. Teaching and learning specifically to pass a test isn't how we create adaptable and grounded humans.
None of these systems necessarily set out to produce these behaviours.
But the systems we're talking about don't care about intentions. They provide incentives, and those moving through the system optimise their way toward them.
Why Gamification Often Fails
Given that systems can produce desirable outputs, it's perhaps no surprise that the idea of gamification emerged.
The thinking seems straightforward enough. Games are engaging. Games use clear reward systems. So if we borrow the language of games – points, badges, leaderboards, streaks – we should be able to encourage behaviour in the real world.
Right?
Right.
But also wrong.
Because gamification systems tend to simplify things, they often end up rewarding the easiest measurable proxy for a behaviour rather than the behaviour itself.
The metrics look good. Engagement goes up. The graphs move in the right direction.
But that doesn't necessarily mean the underlying behaviour has changed in any meaningful way.
A fitness app that rewards streaks may produce streak preservation rather than genuine engagement with exercise. A system that rewards steps may produce step counts rather than meaningful movement.
And fitness apps that reward behaviour with money, coupons, or vouchers may be quietly reframing the activity altogether. Exercise stops being something done for health, enjoyment, or progress, and becomes a transaction.
"Do this, get that."
Which – as the research earlier in this essay suggests – can actually undermine the very motivation the system was hoping to encourage, even while engagement teams point at dashboards and celebrate the fact that NUMBER GO UP.
None of this means games are the problem.
The problem is misunderstanding why games work in the first place.
Games don't motivate people simply through rewards. They motivate people through systems that support autonomy, competence, and meaningful progress.
Designing Behaviour, Not Rewards
And that's why I went in a different direction when designing my own fitness app, Threkka.
Rather than layering rewards on top of behaviour, the goal was to design a system where the behaviour itself becomes the reward.
In Threkka, you're introducing characters to – and helping them enjoy – the very act of movement. You're helping them ritualise it – to exorcise their demons whilst you exercise yours.
Everything in the game world is designed so that the journey you're on echoes the journey they're on. You're in it together.
Literally so, because the only way to earn Sweat – the core currency that underpins the Threkkian economy – is by taking in-game characters out into the real world as you move, exercise, and train.
Threkka deliberately deprioritises simple gamification. Instead it tries to support the kinds of motivations that psychology tells us are more durable: autonomy, competence, and progress.
You choose when and how you move. Threkka reflects that movement back to you in ways that make progress visible, tangible, and quietly world-changing. In other words, the reward isn't layered on top of the behaviour.
The behaviour is the reward.
And the reward is the behaviour.
This also means the system avoids optimising for a single measurable metric. Steps alone don't matter. Distance alone doesn't matter. Speed alone doesn't matter.
What matters is that movement is happening.
What matters is making movement a daily ritual.
Threkka's goal was never to optimise numbers on a dashboard. It was to create a system that makes movement feel meaningful enough that people want to keep doing it.
Systems Produce What They Reward
If there's a lesson running through all of this, it's a simple one.
It's that systems don't just shape behaviour; they shape the behaviour they reward.
And that behaviour isn't always the behaviour the system was originally designed to encourage.
- Games reward points, and Players optimise for points.
- Companies reward metrics, and employees optimise for metrics.
- Gamified systems reward streaks or engagement, and people optimise for streaks or engagement.
None of this requires bad intentions; rather, it simply requires a system whose incentives point in a slightly different direction from the outcome its designers were hoping for.
And once those incentives exist, people will find them, and they will optimise for them. Before long, the behaviour the system actually produces can look very different from the behaviour it was supposed to encourage.
That's why the design of systems matters so much: they quietly redefine what success looks like.
And once that (re)definition exists, people will move towards it – often far more efficiently than the designers of the system ever expected.
Which means that whenever we design systems – whether in games, organisations, schools, or apps – we're not just designing processes or incentives.
We're defining what success looks like.
And people will optimise for it.