96 of every 100 people
are gone in 30 days

The short answer

People do not quit fitness apps because the workouts were wrong. They quit because the app asked for more decisions, more setup and more attention than the results were worth, and most of that happens before the first session.

The numbers

Fitness subscriptions churn at 9.2% a month, the steepest curve of any consumer subscription category. The reasons people give when they cancel are the interesting part. Lost motivation or an abandoned goal accounts for 38%. Free alternatives account for 25%. Price, on its own, is nowhere near the top.

Gyms show the same shape. Half of new members quit before month six, and members who attend fewer than four times in their first month have an 80% chance of cancelling. The first ninety days decide almost everything, and the pattern repeats every January: a huge intake, then 40 to 60% gone by February.

Failure one: the plan was written for nobody

The most repeated complaint across trainers, coaches and app reviews is that generic programming is built for a standard human who does not exist. It ignores injury history, the equipment you actually own, and the time you actually have, so it produces plateaus for some people and overuse for others.

78% of people who train at home say they prefer plans tailored to their goals and abilities. More interestingly, research through 2025 and 2026 has been reframing exercise "non-responders": a large share of people who appeared not to respond to training simply had the wrong programme, not the wrong genetics.

The word "personalised" has been used so heavily that it now carries almost no information. What distinguishes a plan is which inputs it actually takes. Ask any app what it knows about your injuries, your equipment and your Tuesday.

Failure two: friction, not motivation

This one is underrated relative to how often it causes the missed session. Research into fitness app abandonment names setup burden and notification fatigue as primary drivers of people who never start the programme they downloaded.

Be careful with the popular version of this argument, though. The widely repeated claim that people make "over 200 food decisions a day" does not survive checking: nothing was ever counted, the figure came from multiplying two sets of self-reported guesses, and a 2025 analysis showed it is an artifact of asking people to estimate subcategories separately. We used it here before we checked it, and we have taken it out. The full version of this argument is worth reading, because the honest evidence points somewhere more useful than decision fatigue.

The cleanest proof that friction beats features comes from gym check-in data. Wallet-pass check-in runs at 28 to 38% adoption. The equivalent app-based check-in runs at 8 to 12%. Same members, same building, same benefit. The difference is a download and a login.

Failure three: the app that nags

There is a genuine backlash forming against over-personalised nudging. The documented pattern is that users opt out of push notifications entirely or revoke health permissions, because the app has started to feel like a nagging parent. The fix being recommended across the industry is a quiet mode, or letting people choose nudge intensity during onboarding.

Streaks belong in this category too. A streak is a counter whose entire emotional weight comes from breaking, and the day it breaks is the day people stop opening the app. It is worth knowing that habit formation takes a median of 66 days rather than 21, and that missing one day has been shown not to materially impair habit formation, even though almost everyone believes it resets them.

Failure four: nobody explained why

Confusion about goals, results and expectations is one of the most common reasons people leave a personal trainer, and the same analysis finds clients disengage when they do not understand why their plan is built the way it is. Drop-off appears in the data long before anybody admits they are drifting.

This is the cheapest failure to fix and the most commonly ignored. A single line explaining why today is a lower-body day, or why this week is lighter, costs nothing and closes the gap.

What actually helps

The best-supported lever is reducing what the next step costs, rather than trying to make people want it more. In the largest meta-analysis of eating interventions, changes that made the healthy option easier outperformed labels and information by more than three to one.

Planning helps too, though less than the popular version claims: for physical activity specifically the effect is small, and two field experiments on gym attendance found planning prompts produced nothing. More detail, including the finding that fixing your workout to a set daily time made attendance worse, is in the piece on deciding.

Beyond that: make the programme reflect your real constraints rather than an idealised week, and be quiet when there is nothing to say.

← Back to the library