The mechanism nobody tells you when you build your first bonus structure

When an entrepreneur first hires a team and starts thinking about incentives, the logic seems straightforward. Pay people more for doing more of the right things, and they will do more of the right things. The psychology research has spent fifty years establishing why this logic is wrong in a specific and important way.

Lepper, Greene and Nisbett demonstrated the foundational mechanism in 1973. When you reward people for something they already find meaningful, their internal explanation for doing it shifts from “I do this because I care about it” to “I do this for the reward.” The internal motivation doesn’t add to the external one — it gets replaced by it. When the reward is withdrawn or unmet, the motivation goes with it. The external incentive has crowded out the internal reason.

The most consequential word in this research is expected. It is the anticipation of the reward, not just its receipt, that triggers the process. A bonus structure, by definition, is an expected reward — which means it activates the crowding-out effect the moment it is introduced, before a single payment has been made.

Why the economics counterpart reaches the same conclusion

Bruno Frey’s motivation crowding theory, developed independently of the psychology research, arrived at the same finding through an economic lens. External interventions perceived as controlling undermine intrinsic motivation; those perceived as supportive can enhance it. The mechanism operates through perceived locus of control: when a financial incentive signals that behaviour isn’t trusted unless it is measured and paid for, it communicates something about the relationship that the payment itself cannot undo.

There is a specific nuance the research consistently finds that most bonus structures miss entirely. The crowding-out effect is not uniform — it is strongest where the work was already meaningful, and weakest where it wasn’t. A performance bonus for genuinely dull, repetitive work produces little psychological damage. A performance bonus for creative, relational, or complex work — the kind an entrepreneur most wants their team doing well — is precisely where the damage is greatest. The incentive structure tends to be most harmful exactly where the stakes are highest.

The metric problem that compounds the motivation problem

Goodhart’s Law provides the third mechanism: when a measure becomes a target, it ceases to be a good measure. The original observation came from monetary policy in 1975, but the principle has been replicated across healthcare, education, policing, and corporate performance management with depressing consistency.

The NHS wait time case is the clearest available illustration. The government introduced a target — no patient should wait more than eighteen weeks from referral to treatment. Wait times measured against the target improved dramatically. Doctors began delaying official referrals to avoid starting the clock. Easy cases were prioritised to hit targets while complex ones were deferred. The metric improved. Patient outcomes did not reliably improve and in some cases worsened.

Applied to a bonus structure: the moment a metric is attached to a payment, it stops measuring what it was meant to measure and starts measuring how motivated people are to appear to perform. Gaming a metric is almost always cheaper and faster than genuinely improving the underlying thing the metric was designed to capture. The incentive structure optimises for the proxy, not the goal.

Why loss framing doesn’t rescue it

Behavioural economists have proposed that loss-framed bonuses — prepaid and clawed back if targets are missed — should outperform conventional bonuses by harnessing loss aversion. The theory is coherent. A nationwide field experiment among 294 car dealers tested it and found that dealers in loss-framed contracts sold 5% fewer vehicles than control dealers over four months, generating a revenue loss of approximately $45 million. Loss framing raised incentives for risk mitigation and inefficient multitasking in ways that damaged rather than enhanced performance.

The research on bonus versus penalty framing reveals a genuine dilemma. Penalty contracts do elicit more effort in controlled conditions — loss aversion is real. But employees consistently experience them as less fair, and the resentment they generate erodes the trust, engagement, and long-term commitment that effective teams require. Organisations face a real trade-off between short-term effort extraction and the relational infrastructure that makes sustained performance possible.

What the research actually prescribes

Self-determination theory’s three universal psychological needs — autonomy, competence, and relatedness — are what sustain genuine intrinsic motivation. The design question for any incentive structure is not “how much?” but “does this signal trust or control?” Only the former preserves the motivational infrastructure that made the person effective before the bonus existed.

Google’s 20% time — the policy allowing employees to spend a fifth of their time on self-directed projects — produced Gmail and Google News without a bonus target attached to either. The structure worked not because it paid people more, but because it preserved the autonomy that external reward structures tend to erode. The return on not destroying motivation is difficult to quantify. It is also concretely real.

The practical implication for an entrepreneur building their first incentive structure: the most important question is not what to pay for, but what you are communicating about trust, control, and the relationship between the work and the person doing it. Financial incentives are not neutral additions to an existing motivational environment. They change the environment itself.

A book worth reading alongside this

The Tyranny of Metrics by Jerry Muller is the most comprehensive available treatment of what happens when organisations optimise for what is measured rather than what matters. His case analyses across healthcare, education, policing, and corporate performance management provide the applied evidence base for Goodhart’s Law at institutional scale. For any entrepreneur who has ever noticed that a metric they introduced seemed to produce the right numbers while something less visible got worse, this book explains the mechanism with unusual clarity and rigour.

Have questions about this article?

If any part of this article raised questions you want to explore further, courbot.co is built for exactly that. It is courben.co’s AI assistant, designed around the psychology of entrepreneurship. Ask it anything from this article.

This article is for educational and informational purposes only. Sources: Lepper, M.R., Greene, D. & Nisbett, R.E. (1973), Journal of Personality and Social Psychology. Deci, E.L., Koestner, R. & Ryan, R.M. (1999), Psychological Bulletin. Frey, B.S. & Jegen, R. (2001), Journal of Economic Surveys. Goodhart, C. (1975), Problems of Monetary Management.