What happens when you take a reward that people have learned to treat as valuable and replace it with one that is merely symbolic? That is the question behind a striking workplace finding: swapping points for badges lifted repeats of a 90-second task by 26% by 4pm. The number is specific enough to be interesting and odd enough to demand explanation. Why would a token with no cash value outperform a currency that can be exchanged for something real — and why does the effect appear to strengthen as the working day wears on?

The counterintuitive finding: symbolic rewards beating tangible ones

The result runs against a straightforward economic model. If points can be redeemed and badges cannot, rational choice should favour points. Yet the observed pattern is the opposite, and the timing matters as much as the direction. A 26% increase in task repeats by late afternoon suggests the badge is doing something that points were failing to do during the hours when attention and motivation typically sag.

Behavioural research offers a plausible account. Points are a means to an end; their value is deferred and comparative. You accumulate them, you check your balance, you wonder how you rank. Badges, by contrast, are immediate, categorical and identity-bearing. They say something about what you are, not what you will eventually be able to buy. That distinction — between a medium of exchange and a marker of status — turns out to be psychologically load-bearing.

There is also a measurement point worth flagging. A 26% lift in repeats of a 90-second task is a measure of persistence, not quality or enjoyment. The finding tells us that people kept going back to a short, discrete activity more often. Whether they did it better, or whether the extra repetitions reflected genuine engagement rather than a compulsion to collect, is a separate question — and one worth holding in mind throughout.

Why badges speak to identity and points speak to arithmetic

Categorical versus continuous feedback

Points are continuous. You can always have more, and the gap between your total and the next threshold is a permanent, low-grade irritant. Badges are categorical: you have it or you do not. Once earned, a badge stops generating arithmetic anxiety and starts functioning as a completed fact about you. That closure is rewarding in a way an ever-rising counter can never be.

This connects to a well-established idea in motivation research. Deci and Ryan's self-determination theory distinguishes between extrinsic rewards that feel controlling and those that support competence and autonomy. A badge framed as recognition of a specific achievement feeds competence. A points balance framed as a running tally of what you still owe feeds a sense of obligation. The same underlying activity can feel like either, depending entirely on how the feedback is dressed.

The progress principle and visible milestones

Teresa Amabile and Steven Kramer's work on the "progress principle" found that small, visible wins are among the strongest drivers of positive inner work life. Badges are essentially pre-packaged visible wins — milestones that have been named, designed and made legible. Points require you to do the interpretive work yourself: am I doing well? Is 4,200 a lot? Badges remove that cognitive step. For a 90-second task repeated many times across a day, removing a moment of interpretation on every cycle is not a trivial saving.

The 4pm effect: reward salience when self-regulation is depleted

The timing of the 26% figure is the most interesting part of the result. Why would a badge's advantage grow as the day progresses?

The most cited framework here is ego depletion — the idea, associated with Roy Baumeister's research, that self-control draws on a limited resource that depletes with use. The replication record for ego depletion is genuinely contested, and it would be overreach to lean on it heavily. But the broader observation stands on firmer ground: late in the day, people default to whatever requires least deliberation.

Points demand deliberation. They invite comparison, calculation and a small act of deferred gratification each time you consider whether to continue. Badges demand almost none. Under cognitive load, the low-deliberation option wins. That is a parsimonious explanation for a late-afternoon divergence, and it does not require any strong claim about willpower as a finite fuel.

There is a second mechanism worth considering: the end-of-day accounting habit. Many people mentally close the loop in the late afternoon, reviewing what they have and have not done. A points balance at 4pm is a number that says not enough yet. A badge collection at 4pm is a set of things you have already achieved. The same objective day reads as deficit under one system and as accumulated evidence under the other.

What this means for designing repeated tasks

Anyone building systems that ask people to repeat short tasks — training modules, quality checks, micro-learning, compliance routines — should take three things from this.

First, name the milestones. An unnamed threshold is just a number. A named one becomes a thing people can hold. The badge's advantage appears to come substantially from the act of naming and bounding an achievement, not from the graphic itself.

Second, consider what your feedback invites people to calculate. If every glance at the system prompts a comparison or a subtraction, you are charging a small cognitive tax on every repetition. Over a day of 90-second cycles, that tax compounds.

Third, respect the time-of-day pattern. If your task is repeated through the afternoon, the design of late-day feedback deserves separate attention. A system that motivates well at 10am may be quietly working against you at 4pm, and the fix may be as simple as surfacing completed milestones rather than outstanding balances in the interface people see most.

There is a caution to add. Badges can tip into hollow gamification — decoration that signals nothing and recognises everything. The distinction that matters is whether a badge marks something genuinely achieved and specific. If it does, it carries information about competence. If it does not, it is a sticker, and the effect will not survive contact with a sceptical user for long.

Where to look next

The obvious next question is durability. A 26% lift measured by 4pm says nothing about week four. Novelty is a well-documented confound in reward research — new signals get attention, and attention fades. The interesting test is whether badge-driven persistence holds once the badges stop being new, and whether it holds better or worse than points did over a comparable period.

The second question is whether the effect is about badges specifically or about categorical feedback generally. A rank, a title, a completed checklist item — these share the categorical structure without being badges. If the mechanism is closure and identity rather than collectibility, the design space is much wider than a row of icons.

The third is the honest one. Persistence is not the same as value. If you are running a system that asks for repeated effort, the metric to watch is not repeats alone but repeats weighted by whether the work was worth doing. A reward structure that lifts repetition by a quarter is only good news if the repetitions are the ones you actually wanted.