The app spending the most on App Store ads in this category is not the one with the longest-running ad. BetterMe ran 8,760 ad records across the EU in the last year, more than six times any rival here, and its longest creative still live has run 137 days. Muscle Booster ran 1,431 records, and its English line "Gym, Home Fitness, Calisthenics" has run over a year across 14 countries and is still live. Fitbod ran only 140 records and has an English line past a year too. Spend bought BetterMe the most testing. It did not buy the longest survivor.
We pulled every App Store ad these apps ran across EU storefronts from Apple's public Ad Repository, the transparency dataset the EU Digital Services Act requires. Here is the survivor board, sorted by how long each app's longest still-live creative has stayed up. Data window ends 18 July 2026.
The survivor board
| App | Ad records tracked | Distinct creatives | Longest still-live creative |
|---|---|---|---|
| Muscle Booster | 1,431 | 10 | over 1 year ("Gym, Home Fitness, Calisthenics", en-GB, 14 countries, recorded 17 July 2026) |
| Fitbod | 140 | 2 | over 1 year ("Strength Training & Fitness", en-GB, Ireland and Sweden, recorded 17 July 2026) |
| Freeletics | 1,159 | 8 | 341 days ("Bodyweight, HIIT Fitness Coach", en-GB, 5 countries, recorded 17 July 2026) |
| BetterMe | 8,760 | 8 | 137 days ("Pilates, Calisthenics, Yoga", 4 languages, 18 countries, recorded 17 July 2026) |
| Centr | 32 | 5 | 209 days ("Build Muscle. Get Fit.", en-GB), stopped 17 May 2026 |
A few readings of this table.
The biggest budget has the shortest survivor. BetterMe ran 8,760 records, the most in the category by a wide margin, across 18 countries. Every creative it runs today started on 2 March 2026 or later, so nothing it runs has had the chance to prove a year, and its longest live line is 137 days. That is not a criticism, it is a strategy: BetterMe treats the App Store as a high-volume test bench and refreshes constantly. But if you are looking for the ad that pays for itself, the heaviest spender is not where you find it.
The disciplined mid-spenders hold the winners. Muscle Booster and Freeletics each run around a tenth to a sixth of BetterMe's volume, and each holds a near-permanent English line: Muscle Booster's "Gym, Home Fitness, Calisthenics" past a year across 14 countries, Freeletics' "Bodyweight, HIIT Fitness Coach" at 341 days across 5. An ad still live after a year is the strongest free signal a category gives you that a creative converts.
Small spend, evergreen ad. Fitbod ran only 140 records, in just two countries, but its English "Strength Training & Fitness" line is marked over a year old and still recorded. You do not need a big budget to hold a winner. You need one line that works and the discipline to leave it on.
One app already gave up on its winner. Centr's "Build Muscle. Get Fit." ran 209 days, the longest single span of any creative outside the year-plus lines, and it stopped on 17 May 2026. It is the fitness version of a proven ad switched off, and it is exactly the mistake the survivors avoided.
Every survivor names the workout, not the body
Read the surviving lines together. Muscle Booster: "Gym, Home Fitness, Calisthenics". Freeletics: "Bodyweight, HIIT Fitness Coach". Fitbod: "Strength Training & Fitness". BetterMe's widest line: "Pilates, Calisthenics, Yoga". Every long-runner names the kind of training you will do, and not one of them promises the body you will get. No "lose 10kg", no "get shredded", no before-and-after. The ads that survive sell the activity, calisthenics, HIIT, strength training, gym or home, because that is what a person types into App Store search when they are ready to start.
The transformation is the fantasy on the landing page. The workout is what the buyer searches for, and the ad that names the workout is the one that keeps converting. The one place this drifts is in translation, where Muscle Booster's German line becomes "Abnehmen mit calisthenics" (lose weight with calisthenics), leaning on the outcome, but the English survivors that carry the spend stay on the exercise itself.
Where these apps buy
Almost every record on this board ran in App Store search results, the slot you see after typing a query, which fits a category built on intent: people search "workout", "calisthenics", "strength training", or a rival's name when they decide to start. Muscle Booster and Fitbod also run the search tab, the browse surface before you type, so they catch the undecided as well as the ready.
How to read this for your own category
Longevity is the cheapest performance signal you can get for free. Nobody keeps funding an ad that loses money, so a creative still live after months, and past a year in three cases here, is one that converts. Pull your own category the same way: list your three closest rivals, sort their ads by how long each has run, and read what the survivors say. Do not be fooled by the heaviest spender: as BetterMe shows, volume is testing, not proof. The ad that has run longest is the one telling you what your buyer types into search. Then check monthly, because a mid-spender holding a year-old line, the way Muscle Booster does, is showing you the hook the category has already validated.
See any advertiser's ads yourself
See which of your competitors' App Store ads have survived longest, free, no account, in about a minute. Type an app, get its live ads ranked by how long each has stayed up.
Method note: figures are computed from Apple's public Ad Repository (DSA Article 39), across 25 EU storefronts, data window ending 18 July 2026. "Days running" is the span from a creative's first to last recorded impression across those markets, and "over 1 year" is Apple's own 365+ marker. Record counts are what the repository reports for each advertiser; per-country slices return the most recent records, so the analyzed sample can be smaller than the repository total. "Still live" means the creative was recorded on 17 July 2026, the last impression date in this window. Translations are literal glosses of the advertisers' own lines. Longevity is a proxy for performance, not a measured conversion rate. We invented no numbers beyond what the repository reports. Refreshed monthly.