When Rewards Backfire: What 128 Experiments Say About Points, Streaks and XP

The short answer

Adding points to something does not reliably make people want to do it more. Under specific and very common conditions, it makes them want to do it less once the points stop.

Deci, Koestner and Ryan meta-analysed 128 experiments in Psychological Bulletin and found that expected tangible rewards, delivered contingent on doing a task, reliably undermined intrinsic motivation for that task. The effect held whether the reward was contingent on engaging with the task, completing it, or performing it well.

This is the single most important finding for anyone shipping a gamified product, and it is almost never cited by anyone shipping one.

What "undermining" actually means

The standard measure is the free choice period. Participants do a task, some are rewarded, then the reward is removed and everyone gets unstructured time with the task freely available. The question is what people do when nobody is paying attention and nothing is at stake.

Rewarded participants consistently chose the task less than unrewarded ones. They also tended to report lower interest in it.

The mechanism, in Self-Determination Theory terms, is perceived locus of causality. Intrinsically motivated behaviour is experienced as self-chosen. A contingent tangible reward shifts the perceived reason for acting from "I want to" to "I am doing this to get the thing." When the thing goes away, so does the reason.

This is sometimes called the overjustification effect: you supply an external justification for behaviour that already had an internal one, and the internal one atrophies.

The exceptions, which are the useful part

The finding is not "all rewards are bad." The moderators are what make it actionable.

Two rules fall straight out of that table.

Informational feedback beats tangible reward. Feedback that tells you how you are doing supports competence without shifting the locus of causality. Feedback that is used to pressure you ("you should keep this up") behaves like a contingent reward and undermines.

The undermining applies where intrinsic motivation already exists. For genuinely uninteresting tasks there is less intrinsic motivation to displace. The risk concentrates on exactly the work people already care about, which in a productivity product is the work that matters most.

What this means for gamification

Most gamified products are built in the dangerous quadrant. Expected, tangible, contingent, and attached to the core behaviour. The typical result is a system that produces strong engagement while the rewards are salient and a sharp collapse when the novelty ends, with the user now less interested in the underlying behaviour than before they installed it.

Design rules that follow:

1. Never make the reward the reason for the core behaviour. Points can decorate progress. They must not become the answer to "why am I doing this."
2. Keep the reward layer separable from the outcome measure. If points inflate your primary metric, the metric now measures reward-seeking rather than progress.
3. Weight informational feedback over tangible reward. Praise the work, report the numbers, and let the reward layer be secondary.
4. Watch what happens when the rewards are removed. That, not daily engagement, is the diagnostic.

Why UpTicker keeps XP away from your productivity score

UpTicker has a full progression system: 30 levels, six tiers, an XP curve, streaks. That system sits deliberately outside the measurement layer.

The productivity score uses base points only, never the gamified multiplier. Streak bonuses and level multipliers do not inflate it. This means the score reaches 100% only when the work is genuinely complete, and it means your primary signal about your own progress is not contaminated by reward accumulation. If XP could raise the score, the score would stop being feedback about the work and start being feedback about playing the game. The literature above says that substitution is exactly what damages long term motivation for the underlying behaviour.

The coach reports on work, not on points. Per Kluger and DeNisi, feedback needs to point at the task. Per Deci, Koestner and Ryan, informational feedback supports intrinsic motivation while controlling delivery undermines it. Both point the same direction: describe the work, attach a next action, do not use the reward layer as leverage.

XP accrual is silent. Level ups happen, but the XP strip was removed from the main header. Progression is available if you want it and it is not the frame through which you see your day. That is the difference between a reward that decorates the behaviour and a reward that replaces the reason for it.

Frequently asked questions

Do rewards always reduce motivation? No. The meta-analysis found the undermining effect concentrated in expected, tangible rewards delivered contingent on the task. Unexpected rewards and rewards not tied to task performance did not show the effect.

Does praise undermine motivation too? Generally the opposite. Verbal praise and positive feedback tended to enhance intrinsic motivation, unless delivered in a controlling manner.

What is the overjustification effect? Supplying an external reason for behaviour that already had an internal one, which weakens the internal reason. Once the external reward stops, engagement can fall below its original level.

Should productivity apps avoid points entirely? Not necessarily, but points should not become the reason for the core behaviour, and they should not inflate the measure you use to judge your own progress.

How do I tell if a reward system is harming me? Look at what you do when the reward is unavailable. If the behaviour stops when the streak is already broken or the points do not count, the reward has become the reason.

Reference

Deci, E. L., Koestner, R., and Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627 to 668.