Motivation is not a single fuel you top up with rewards. It comes in kinds, the wanting can split from the liking, and piling on incentives can drain the very drive you meant to boost.
Let's start with the most surprising finding, because it breaks the classic common idea of "motivation". The brain runs two systems that usually travel together but can come apart: wanting (the pull to pursue something) and liking (the pleasure you actually get from it). Wanting is driven heavily by dopamine and works like a magnet toward a goal; liking is a separate, smaller pleasure signal. Because they are separable, you can intensely want something you no longer much enjoy. That gap is the engine of compulsive behaviour, from gambling to the endless scroll: the pursuit stays strong while the payoff fades (Berridge & Robinson, 1998). In other words, motivation is not the same as satisfaction.
The second pillar is the split between intrinsic and extrinsic motivation. Some activities are their own reward, you do them because they are interesting or satisfying in themselves. Others you do to get something else, money, marks, approval. The counterintuitive result is that bolting an extrinsic reward onto an intrinsic activity can undermine it. This is the overjustification effect: once you are paid to do something you used to do for fun, you start to read your own behaviour as done for the reward, and your spontaneous interest drops when the reward goes away. In the classic study, preschoolers who loved drawing were promised a certificate for doing it; afterwards they drew less on their own than children who got nothing or an unexpected reward (Lepper, Greene & Nisbett, 1973). Deci (1971) found the same with money and adults.
It pays to be precise about when this happens, because the headline "rewards kill motivation" is too broad. The undermining shows up for rewards that are expected, tangible, and handed over simply for doing the task. Rewards that arrive as a surprise do not do the damage, and verbal praise or useful feedback actually tends to raise intrinsic motivation rather than lower it (Deci, Koestner & Ryan, 1999). So the danger is narrow and specific: paying people to do the thing they already wanted to do. Which puts a scientific dent into the old adage "find a job you love and you'll never work a day in your life". Romantic, but loving the work is no guarantee it stays effortless, and being paid and measured for it could quietly dim the very enjoyment that made it feel that way.
The third pillar is what does sustain motivation. Self-Determination Theory points to three basic needs (Ryan & Deci, 2000). Autonomy is the sense that you are acting from your own choice rather than under control. Competence is the feeling of being effective and growing. Relatedness is feeling connected to other people. When a job, a habit, or a cause meets these, motivation tends to be self-sustaining and to come with better wellbeing; when they are thwarted, you get the brittle, pressured kind of motivation that collapses the moment the pressure lifts. Across 123 countries, having these needs met predicted wellbeing (Tay & Diener, 2011).
The reward debate was a genuine fight, not a footnote. One large meta-analysis argued that rewards, on the whole, do not reduce intrinsic motivation, and that only expected tangible rewards for an interesting task do, by a small amount, while praise reliably helps (Cameron & Pierce, 1994). A second, larger one argued the undermining effect is real and robust across several kinds of contingent reward (Deci, Koestner & Ryan, 1999). The useful way to read this is that both camps agree on the narrow case (an expected, tangible payment for an already-interesting task is the risky one) and disagree mostly about how big and how important that effect is in the wild.
Two motivational staples fare worse. Maslow's hierarchy of needs, the pyramid from food and safety up to self-actualisation, is everywhere in business training, but the strict order, the claim that you climb one level only after satisfying the one below, has never really replicated. The broader idea that humans have several kinds of need survives; the staircase does not (Tay & Diener, 2011). In other words: we do seem to have a series of needs at different levels, but we don't go about satisfying them in a specific "upward" order. In fact, we pursue them more or less simultaneously, and a higher need can pay off even while a lower one is unmet.
And ego depletion, the popular idea that willpower is a single fuel that drains as you use it, took a serious hit: a preregistered replication across 23 labs and more than two thousand people found an effect statistically indistinguishable from zero (Hagger et al., 2016). Self-control failure is real, but "you ran out of willpower fuel" is not a safe explanation for it.
Much of the reward research is short-term and run in the lab, with intrinsic motivation measured by how long people keep doing an activity when no one is making them. Real-world incentives are messier: a clear reward can genuinely help with a boring task or with getting a habit started, and the long-run effects of pay on meaningful work are hard to isolate. Self-Determination Theory's three needs are also measured partly through the outcomes they are meant to explain, which leaves some room for circular reasoning.
When exactly does an incentive help and when does it crowd out the drive that was already there? Are autonomy, competence, and relatedness truly universal, or weighted differently across cultures? And if willpower is not a tank that empties, what actually explains why self-control fails when it does? That mechanism is now genuinely open.
The practical core is one question asked before you reach for a reward: does this person already find the thing worthwhile? If yes, an external carrot is a risk, not a safe bet.
Be careful where you point money. For work people already find meaningful, cash bonuses tied to simply doing it can crowd out the internal drive you were relying on; you are better off feeding autonomy (real choice over how the work is done), competence (mastery, progress, useful feedback), and relatedness (a team worth showing up for). Save expected, tangible, contingent rewards for genuinely dull tasks where there is little intrinsic motivation to damage. And note that praise and good feedback lift intrinsic motivation, so they cost almost nothing and only help. In marketing, separate wanting from liking: you can manufacture wanting with cues, anticipation, and variable rewards without the product being any more liked, which may drive the first purchase but not the loyalty that comes from genuine satisfaction. Finally, drop Maslow-pyramid segmentations, which do not predict; map your category to autonomy, competence, or relatedness payoffs instead.
Activists and volunteers run almost entirely on intrinsic motivation and relatedness, so over-managing them or replacing meaning with transactional rewards can quietly demotivate the people you most need. It also helps to separate wanting from liking in the electorate: enthusiasm and turnout (wanting) can run high even when actual satisfaction with the candidate (liking) is low, and the two respond to different levers.
Incentives and penalties for civic and health behaviour can backfire when they replace a sense of duty with a transaction, the crowding-out problem in policy form, so design them to support people's sense of autonomy and competence rather than to simply pay for compliance. And be wary of policy built on the willpower-as-fuel story ("people just need more self-control"). Structure, defaults, and removing friction do more durable work than exhortation, because the fuel-tank model they lean on does not hold up.
Before adding a reward, check whether the person already cares. If they do, support autonomy, competence, and relatedness, and use feedback and praise rather than a contingent payment. If the task is genuinely dull, a clear reward is fine and often necessary. Separately, learn to tell your own wanting from your liking: ask whether a strong pull is a sign you will actually enjoy something, or just a craving the cues have built. And stop treating willpower as a tank to be rationed.
.png)