An 18-Month Trial Kept 79% of Its Participants. That Is Not What App Retention Looks Like.
Against a category that loses most of its users inside a month, one trial reported 98.5% retention at six months, 89% at twelve, and 79% at eighteen. It is worth understanding both why that happened and why you should not expect it from anything you download.
The trial
Two hundred adults were randomised to either immediate access to a mobile behaviour-change support system or access after a six-month wait, then followed for eighteen months.
At six months the immediate-access group improved on every eating-behaviour measure taken: cognitive restraint up 8.4 points, uncontrolled eating down 8.2, emotional eating down 9.4, binge eating down 3.4, all at p<0.001. Then the delayed group produced the same improvements after their own six months of use (International Journal of Obesity, PMID 42538370, July 2026).
That internal replication is the strongest feature of the design. A waitlist arm reproducing the effect on its own schedule is far more convincing than a before-and-after chart, because it rules out the ordinary explanations a before-and-after chart leaves open.
The trial has real limits. Every outcome except body weight was a self-reported questionnaire, blinding was impossible, the population was people with clinical obesity rather than general adults, and the trial was run by the people who own the product. Two of the authors founded Onnikka Health Oy, which holds the app’s intellectual property, and they, the first author and the University of Oulu are all shareholders in it. That does not invalidate the result. It does mean it is not independent evidence.
What moved
The detail worth carrying is what the intervention changed. The trial never measured what people ate. What it measured, and what moved, was how people related to eating: how much they held back on purpose, how often eating felt out of control, how often they ate from emotion, and how often they binged. Weight change tracked those behavioural shifts rather than standing apart from them.
That is a different theory of how these products work than the one the category usually sells. The app bundled articles, reflection tasks, self-monitoring and reminders into one package tested only as a whole against a waiting list, so which part did the work is unknown, and the trial never measured what people knew or how much they logged.
Why 79% will not happen to you
Now the part that matters more, because it is the part most fitness marketing omits.
That 79% is enrolled-trial retention. Participants were recruited and screened, they consented, they had scheduled follow-up visits, and researchers were measuring them throughout.
Set against it: a peer-reviewed analysis of 2,771 real users of a commercial app found sessions down 69.3% by month one and 80.6% by month three, a fall in session volume across the cohort rather than a count of people who stopped (JMIR mHealth and uHealth 2026;14:e72201, PMID 41505740).
The same gap appears everywhere in this literature. A supervised twelve-week trial in inactive night-shift nurses: 96.2% of prescribed sessions completed, under scheduled supervision at a research facility (The FASEB Journal, PMID 42444395, July 2026). A twelve-week wearable programme in 304 adults: 99.3% completion, in a single-arm study with no control group at all (Journal of Yeungnam Medical Science 43:53, PMID 42571925, August 2026).
What the high numbers share is not the technology. It is a study setting where a human was somewhere in the loop and somebody would notice if you stopped. None of these trials ran the same programme without that, so take it as a pattern across the designs.
So when a product cites adherence research in the nineties, the honest reading is that you are looking at a study’s conditions rather than a product’s performance.
What transfers anyway
Three things do, and they are not the ones usually marketed.
Design for the frequency people manage. Real users averaged 1.9 sessions a week. A plan built for five days a week is a plan you fail in the first week.
Whatever helps has to help in the first fortnight. A corporate wellness study lost users from 84% at onboarding to 14% by week two (Journal of Occupational and Environmental Medicine, PMID 42579843, August 2026). Any mechanic that pays off at week eight is arriving to an empty room.
Onboarding completion is close to meaningless as a signal. In the same study, by week six only 6% of everyone who registered was still interacting with it. What matters is how many are still there a month in, and that is the number almost nobody publishes.
The gap ZeroFit has not closed
Honesty is cheaper than positioning here.
Supervision, or at least a human touchpoint, is present in most of the studies with the good adherence numbers. A scoping review of just-in-time nudges lists support from health-care professionals and the care environment among its facilitators (JMIR mHealth and uHealth, PMID 42566750, August 2026), and the 99.3% pilot had study-team contact built into the design. None of those studies removed the human to see what happened without one, so this is co-occurrence rather than a measured ingredient.
ZeroFit has no human in the loop, and no notifications either. Neither is built.
What it has is the other lever: making the session small enough and the decision light enough that week two does not require resolve. Two questions, one session, a floor of five minutes, and nothing that resets or scolds when a week goes missing. Whether that substitutes for a person who notices is genuinely unproven, and pretending otherwise would be the same overclaiming this article is about.
Do fitness apps work has the retention evidence in full, and why fitness apps make you feel worse covers the design side of why people leave. The workout you enjoy beats the one you’re supposed to do covers the input that predicts coming back, and does goal setting work for exercise covers the mechanic most products reach for first.
Two questions, one session. Nothing to rebuild when you come back. Get beta access.
ZeroFit content is educational and not medical advice. If you have an injury, a health condition, or you’re pregnant, check with a qualified professional before starting any new movement routine.