Everything decided on 8 Jul 2026, in one place. This is the strategy we follow and the index of every doc behind it. When something contradicts this, this wins — until we amend it here.
The core problem isn't traffic — it's visibility. RevenueCat shows ~763 paying subs, but the analytics layer recorded the trial event 3 times in 90 days. We can see that sales happen, not which step or which ad made them. Fix that first; everything else is a guess until we do.
✅ 21 Jul 2026 — visibility is improving on the web side. Nic shipped Google Ads Lead + value-bearing Purchase conversions (GTM) and a browser-independent server-side Stripe → GA4 + Meta CAPI webhook (transaction_id deduped, 200 OK on a real payment). The web measurement gate is nearly closed pending an Ads-UI confirmation; the app-side (RC→AppsFlyer) and gclid→revenue join remain. Full architecture + proof: docs/company/ANALYTICS/TRACKING-SYSTEM/.
We do not have a normal trial→paid funnel. A lead becomes a paying customer at the paywall, on one of two paths — and the number we steer by must count both:
So: the funnel below reads "…→ trial → paid," but the true gate metric is Lead → Paid, by source (currently unmeasured by channel — the top instrumentation job). It flows to the meter (count both paths) and Task 3 (import both, deduped by transaction_id). Detail: MESSAGE-TO-NIC-FUNNEL-CORRECTION-2026-07-16.md · event catalog §1 · $15K doc §⑮/§⑰.
Locked fact (auto-synced from canonical_strategy): Lead → Paid, by source — two paths: carded free trial (trial-start IS the purchase) + direct ~50%-off purchase (no trial). Not "trial → paid".
Two funnels bring people in, an email backend converts them, an offence engine fills the top, and a retention plan keeps them. Two things run at once:
| Funnel | Path | North-star metric |
|---|---|---|
| 1 · Web | quiz → paywall → email → nurture → trial → paid | lead → trial per funnel · MoM new-MRR growth |
| 2 · App (iOS-first) | install → onboarding → paywall → 7-day trial → subscribe | |
| Backend | email (AC → Customer.io; Intercom = in-app, already ours) | trial→paid conversion |
| Retention | keep a payer · revive a lapsed one | churn % · D30/D90 · reactivation |
| Question | Decision |
|---|---|
| Prioritisation | THE MRR-30 FILTER "Does it touch MRR in 30 days?" No → WAITS. The three payers: trial launch (19→GO) · purchase signal (9/5/2/3) · non-converter emails. See the box above. |
| Run paid now? | YES — to the web quiz Meta + Google → free assessment, measured on cost-per-lead. Nic runs it, small & capped. iOS when the app funnel's instrumented. |
| Rebuild the checkout? | NO payment step works (72%); ship safe copy wins, gate a rebuild on data. (Ghost-PIs were the lifetime cart, not the evergreen checkout.) |
| Analytics: Nic or RN team? | HYBRID Nic owns Phase 0 + the two attribution-critical merges; RN team does the native build + app bug fixes. Nic is full-time. |
| Retention | Churn is an execution problem — #1 lever is fixing the app (16% of churn + the top "what would make me stay"). Own plan. |
| Tooling | RevenueCat Experiments (not Adapty) · GrowthBook (web A/B) · AC → Customer.io · Intercom kept (in-app) · AppsFlyer kept · n8n fan-out. |
| Focus (founder lens) | The real constraint is diffusion. Say no to net-new builds/fleets; yes to this one chain. Nic on plumbing; growth runs parallel, never queued behind him. |
| Organic vs paid | PAID is the reach engine Organic social is ~dead (~0 reach) — a lottery ticket, not a pillar. Test creative in the paid account, not organically. |
| Brand / founder story | The injury→rebuild origin story is Aga's — email-only, never public. Public coach persona stays Jesse. |
| Owner | Owns |
|---|---|
| Nic (full-time) | The plumbing (measurement wire, user_id), the paid ads, checkout/paywall fixes. Reviews the RN team's attribution merges. |
| Lee + Ritesh | Cheap organic + ManyChat "comment ASSESS" capture → free-assessment email (upside-only — social ~dead, not a pillar). One creator/affiliate shot. |
| RN team | Native instrumentation (T4–T7) + the app bug fixes behind the churn. |
| Email fleet | Lead→trial nurture, trial→paid sequence, 14-day churn trigger, win-back (live). |
| Aga | The cut & the focus, pricing decisions, creative approval, the weekly review. |
Hard in-platform caps · weekend lock · a daily spend alert to Aga + Nic · lead quality reviewed weekly by source · one shared daily scoreboard where "done = a moved number." No paid scale-up before Lead→Paid is visible and the quiz screen-1 leak is fixed. (The original launch guardrail was ~$30–40/day — RETIRED 9 Jul, see below.)
AMENDED 9 Jul 2026 (Aga) — the current rule: the paid envelope is $3,000/mo USD ceiling — ALL ad platforms COMBINED (Meta + Google + Apple Search Ads) — ramp, not splash (the $30–40/day cap is RETIRED, SoT amended 9 Jul 2026; scope clarified to all-platforms-combined by Aga 24 Jul 2026 — the July 2026 overage was the known, accepted lifetime-ads push, not a breach going forward), governed by the Paid Experiment Roadmap (cockpit 🧪 tab): monthly staging $60/day retargeting → $97/day (Aug) → $100/day (Sep–Oct), each month gated by pre-registered CPP circuit-breakers ($750-spend checkpoints; pause at CPP >$150 with <5 confirmed purchases) and by the measurement hard gate (dedupe + the Lead→Paid meter before any scale-up). Weekend lock + daily spend alert unchanged. Spend is API-live daily (Meta + Google) on the cockpit.
| Doc | What it is | Link |
|---|---|---|
| Growth Plan ⭐ | The main plan → send to Nic. Sprint to MRR, the full backlog with hours. | tma-growth-plan-2026 |
| Retention Plan | Keep/revive members — from 4,434 exit surveys. | tma-retention-plan-2026 |
| Analytics Brief | The instrumentation spec Nic builds from. | tma-analytics-brief-2026 |
| Growth Analysis | The data dig (June/July, checkout autopsy, paid ads) — MRR now corrected. | tma-growth-analysis-jul2026 |
| 10x Plan stretch experiment | Founder-focus stretch: what would 10x us in 6 months (6 levers, honest odds). The swing above the floor — not a committed directive. | tma-10x-stretch-jul2026 |
| Change Brief | The restructure record (retention split, offence engine, clock). | TMA-GROWTH-PLAN-CHANGE-BRIEF-JUL2026.md |
| Tracking Reality Check | The proof the funnel's dark (af_start_trial = 3/90d). | TMA-TRACKING-REALITY-CHECK-JUL2026.html |
| Message to Nic | The short send-ready message. | MESSAGE-TO-NIC-GROWTH-PLAN-2026-07-08.md |
All local files live in docs/company/ANALYTICS/. MRR puller: tools/finance/stripe_mrr_pull.py.
Separately, a founder-focus stretch thought experiment: what would 10x us in 6 months. Keystone = a launch to the ~15K active list (£9–54K cash, self-funding, runway-extending) → then offer+upsells → paid on the fixed offer. The swing above the floor, not a directive. Aga owns the email side; needs 5 decisions to fire. tma-10x-stretch-jul2026
Sprint 1 ends 22 Jul: RC→Amplitude reconciled ±5% · 110-queue save sent · cancel-flow live · retarget live (bot-purged list only) · D6 nudge live · UTMs capturing. → By 31 Aug: churn <15/wk · leads ≥250/wk · net-MRR Δ positive 2 consecutive weeks. → By 1 Nov: MRR ≥ $15K (break-even) — recurring only, one-time promo cash never counted. Then the growth-rate phase. The maths behind these: $3.9K gap ≈ 270 net subs; the −29 subs/wk bleed means halving churn ≈ worth 3× lead flow. Full working in the Growth Plan §0.5 Sprint Contract.