A trial with 602 people, not a lab test
The most-cited study in this debate is STEP UP, run by Patel and colleagues (2019) and published in JAMA Internal Medicine. The team recruited 602 overweight or obese adults (BMI of 25 or higher) across 40 U.S. states, gave each one a wearable to count daily steps, and randomized them into four groups: a control that just wore the tracker, and three gamification groups with social incentives — support, collaboration and competition — each running the same points mechanic (70 points at the start of the week, losing 10 for every day the step goal was missed, with a chance to move up or down a level) on top of a different way of comparing each person with others. The active phase ran 24 weeks.
The result from that phase is straightforward: all three gamification groups walked significantly more steps per day than the control across the 24 weeks, and competition — putting each person's performance up against other people's, rather than just tracking their own goal — was the most effective of the three.
When the points stopped counting
The part most summaries of gamification leave out is what STEP UP measured next: 12 weeks of follow-up, with no active incentive at all, just the tracker still counting steps. In that phase, activity dropped in all three gamification groups. Of the three, only competition kept a statistically significant difference over control — 569 more steps a day (95% CI, 142 to 996). Support and collaboration, the two formats that never put anyone in direct contest with another person, lost their advantage as soon as the game stopped running.
A meta-analysis of 16 studies shows the same curve
STEP UP is a single trial, however large. The question about the pooled effect size was answered by Mazéas, Duclos, Pereira and Chalabaev (2022), in a systematic review with meta-analysis published in the Journal of Medical Internet Research, pooling 16 randomized controlled trials and 2,407 people. During the intervention, gamification's effect on physical activity was small to moderate — a Hedges' g of 0.42 (95% CI, 0.14 to 0.69).
The number that matters for this article's title is a different one: in the studies that also measured the period after the intervention ended — an average of 14.4 weeks later, in a window of 12 to 24 weeks — the effect fell to g = 0.15 (95% CI, 0.07 to 0.23). Still greater than zero — the gain does not vanish completely — but much smaller than during the game. It is nearly the same curve STEP UP showed person by person: the effect is real while the mechanism is switched on, and shrinks a great deal once it is switched off.
What that drop means — and what it does not
That drop is not a reason to write gamification off as "just a placebo" or marketing. An effect that goes from moderate to small, but stays statistically greater than zero 14 weeks after the game ends, is different from an effect that resets to zero the moment the incentive disappears. And not every format dropped the same way: in STEP UP, competition survived where support and collaboration did not — a sign that the ingredient that lasts is ongoing comparison with other people, not just the act of earning points.
What the drop shows, honestly, is that gamification works as a push, not a permanent transformation. It helps someone get off zero or get through a stretch of low motivation — it does not replace the personal reason to keep training once the scoreboard stops counting.
Where this shows up in the crew challenge
PulsoGym's crew challenge uses the ingredient STEP UP showed survives — a shared goal and a scoreboard everyone can follow, not just the coincidence of training in the same place. But the honesty these two studies call for applies here too: the challenge is the push, not the reason. What is left once it ends is the log of load and sessions in the app — that record, unlike the scoreboard, does not switch off when the challenge does.