The companion to the growth plan. The growth plan gets a stranger to their first payment; this plan keeps them, and revives the ones who've lapsed. Built entirely from what members already told us — 4,434 real exit surveys + our testimonials — no new research needed.
Our churn is an execution problem, not a product problem. Members still want what TMA does — they leave because the app breaks, the value isn't visible, or life gets busy and there's no way to pause. The single biggest lever is fixing the app: it's 16% of why people leave and the #1 thing (35%) they say would make them stay. That's the case for the RN team — retention is partly a fix problem, not an email one.
The number we drive: monthly churn % · D30 / D90 retention · reactivation rate. (Separate from the growth plan's lead→trial.)
| Reason | Share | What it really is | |
|---|---|---|---|
| Cost / "not worth it" | 20.5% | Value not seen — 64% said they'd pay nothing to stay (they don't see results worth it) | |
| No time / life got busy | 17.7% | Temporary — they'd come back if they could pause instead of cancel | |
| App problems 🔧 | 16.0% | Crashes, doesn't save progress, confusing — immediate churn of engaged members | |
| Never really started | 6.9% | Paid but never activated — an onboarding failure | |
| Found an alternative | 4.3% | Gym, another app, "can do it myself" | |
| Injury · fit · billing | ~4.6% | Smaller, mixed |
When members were asked "what would make you stay?", the #1 answer wasn't a lower price — it was a better app (35%). Price was second (25%). So the biggest retention win is app quality, not discounts. And the "wanted fast results, got slow progression" expectation gap drives a lot of early churn — we can close that with better first-week framing.
§1 is what members say when they leave. This is what they do — the first-ever native retention cohorts (GA4 prop 177471782, iOS+Android). Caveat: this is engagement retention (did they reopen the app), not subscription retention (RevenueCat), and not a money number. Baseline = pre-promo (the truth); promos: May 50%-off, 4–12 Jul lifetime.
| Cohort era | D7 | D28 | n | Read |
|---|---|---|---|---|
| BASELINE (Feb–Apr 2026, pre-promo) | 20% | 6% | 1,378 | the truth — stable a full year (early-2025 also 19%/6%) |
| May 50%-off | 26% | 7% | 807 | HEALTHIER than baseline — a discount attracts intent |
| July lifetime (4–12) | 11% | ~0% | 4,282 | TOXIC — 15× volume of non-retaining cheap-geo installs (polluted every recent blended number) |
Baseline splits: iOS 24%/7% vs Android 16%/5%; paying geos (US/UK/CA/AU/DE/FR) 24%/8% vs cheap geos (IN/PK/ZA) 14%/2%. The daily curve is a cliff, not a slope — D0 100% → D1 ~10% → D7 2%. The first-session-to-second-session gap is where users are lost — which is exactly why activation (§2 #3) and the first-workout moment are the durable levers.
50%-off on a recurring plan is a healthy mechanic (it retained better than baseline and brings MRR). Ownership / lifetime is banned — it kills both the reason to return and the ability to re-monetize, and it retained at ~0% D28. Audience quality outranks offer mechanic (warm + paying geos only; no cold cross-network promo). Judge every promo on D28 cohort retention + net MRR — never install volume.
We just ran the deepest offer possible to a big list; those people are 89%+ gone by D7 and any follow-up offer is mathematically worse than what they already declined. Don't re-promo the wave. But it's three different populations, three different plays:
The win-back sequence is 976 Cancelled Win-Back (live in ActiveCampaign; triggers fixed 21 Jul so only genuinely-expired members enter). The ruling on what to offer them:
64% of churners would pay nothing and ~89% of discount-saved re-churn within a year, so blanket discounting is out. Only the price branch ever sees money — and it's a come-back ANNUAL, not a monthly cut.
| Cancel reason | What we offer |
|---|---|
| Price / value | Value receipt, then a come-back annual: SAVE30 → $109.90/yr (30% off $157, effective $9.16/mo, 12-month lock, cash upfront). Fallback for decliners: capped $12.97 × 3 months then standard. |
| No time / life busy | A pause / hold — not a discount. Their access is held; nothing resets. |
| App bugs | "Here's what we fixed" + reply-to-Jesse. Zero discount (a discount on a broken experience insults them). |
| Never started | Activation restart — reuse the $9.99 / 30-day entry offer, not a retention discount. |
| Injury | Recovery-first, place held, no price talk. |
1. $9.99/month forever is retired — it permanently halves ARPU and you never collect the "forever" (~$50–60 expected vs the annual's ~$110 day-one). 2. Reply-based delivery — the offer email carries no link; it says "reply and I'll send it over," then we send the annual checkout + SAVE30 (better deliverability, and the post-cancel offer must never be deeper than the pre-cancel R1 save — both are SAVE30, at parity).
| Stage | Why they go | What stops it |
|---|---|---|
| First 30 days | App breaks · expected fast results · price shock · overwhelm | #1 bug fixes · #3 activation + quick wins · #7 expectation reset |
| 30–90 days | Slow progress · got busy · boredom · "is it worth it?" | #2 pause · #4 check-in · #5 visible value · content variety |
| 90+ days | Goal achieved · life change · found alternative | #5 value view · new goals/skills to chase · #6 win-back |
These run in parallel, not in sequence:
They share two things: the RN team (app fixes serve both), and the measurement wiring (the same RevenueCat events power both scoreboards). Both report into the one daily dashboard so nothing hides.
No new member interviews — we already have 4,434 exit surveys and 8 years of testimonials; the "why" is known. No big rebuild — the top four moves are bug fixes, a pause feature, a better first week, and one email. Start there.