auto-synced 12 Aug 2026 · live spend + queue figures auto-sync from canonical.
Canonical view = the Growth Cockpit → 🧪 Paid Roadmap tab (
https://tma-growth-cockpit-2026.netlify.app). This doc is the narrative source; the cockpit tab is the operative allocation. On 10 Jul a 7-seat paid-ads fleet (Meta · ASA · Google · Attribution · Creative · Analyst · CMO) reviewed the V2 plan below independently and converged unanimously: the plan is directionally right but backs the least-proven channel (Meta cold) first and starts spending before its own measurement gate. The V3 corrections (envelope unchanged — still $3K/mo):
- $0 Meta until the CAPI purchase event is verified end-to-end; then $20–25/day warm (not $65). (Meta · Attribution)
- ASA category $17/day NOW, not September — it's the only confirmed purchase signal (7.4% install→purchase, 2.6× organic). (ASA)
- Google Search forward to Month 1 ($8/day, feeds the web quiz); only UAC stays last. (Google)
- Ramp the dollars, don't splash: ~$40/day Month 1 → $85 → $100 as CPP proves. $3K/mo is a ceiling, not a day-1 run-rate; the balance stays in cash = runway. (Analyst · CMO)
- Tighten the breaker: scale <$50, pause >$87 (plan-mix-aware); define what a "purchase" is before any spend. (Analyst · Meta)
- 3 cold creatives at a time, not 6 (6 at $16/day = noise); add the missing UGC injury anchor + quiz pre-frame. (Creative)
- Judge on Cost-per-Trial-Start reconciled to RC (CPTS), not CPP, until purchase volume is statistically usable. (Attribution · Analyst)
- Honesty flags: the 80/135/216 projections are reverse-engineered from the $4K gap (all imply ~$97.50 upfront); the pessimistic case ($131 CAC) violates the plan's own >$130 PAUSE breaker; cash-upfront ≠ MRR (an annual $157 upfront = $13.08/mo MRR). Paid alone cannot close the gap — R1 + email + the S01 fix carry the rest.
Efficient allocation, one line: free fruit first ($0, out-earns all paid) → confirmed signals first (ASA + Google Search + Meta warm) → Meta cold gated behind verified purchase measurement → ramp into the $3K ceiling only as CPP earns it.
Nothing paid spends until the wiring is done. Every dollar below is gated on ~9 hours of measurement wiring (keystone: RC→AppsFlyer S2S, ~6h). Until it lands we spend blind. This is needed. It unlocks the entire $3K/mo engine.
👉 Nic — open the Analytics Master Brief and execute the wiring, in this order: Task 7 (ASA cost, 30 min → ASA starts today) → Task 5 ★ (RC→AppsFlyer — repairs Meta + Google + ASA at once) → Task 9 (web purchase event) → Tasks 2/3/8 (Meta CAPI + Google Search). Brief: https://tma-analytics-brief-2026.netlify.app · the paid-optimisation event = the trial-start (card-on-file) = the buying decision, so the early KPI is CPTS reconciled to RC, not CPP.
| Channel | Feeds | Monetises via | Fee drag | Role |
|---|---|---|---|---|
| ASA (brand + category) | iOS app 📱 | App Store install → in-app subscribe | ~15–30% Apple | Best-proven paid bet — funded FIRST ($25/day Mo.1) |
| Google UAC | iOS + Android app 📱 | Store install → in-app subscribe | ~15–30% | LAST — blind + junk geos until rebuilt |
| Meta (warm → cold) | Web quiz 🌐 | Web checkout (Stripe) | ~5% | The scaling bet (better margin) |
| Google Search | Web quiz 🌐 | Web checkout (Stripe) | ~5% | Cheap intent harvest into the web funnel |
The iOS app is NOT neglected — it's our first paid money (ASA, $25/day now) because ASA is the only channel converting today. But the web quiz leads the scaling because a web sale keeps ~25% more of every dollar (annual/quarterly cash upfront at ~5% fees vs Apple's 15–30% cut — the Noom/BetterMe model). Killed/parked: Meta app-install campaigns (2,878 installs → 1 purchaser); Google UAC deliberately last (bidding blind, 90% payment-incapable geos).
| Channel | Mo.1 $/day | Funnel / fees | Proof today | Decides | Start |
|---|---|---|---|---|---|
| ASA brand defence | $8 | iOS app | We bid $0 — conquestable NOW | Stops rivals buying our name | 🟢 today, $0 wiring |
| ASA category | $17 | iOS app (Apple cut) | Best signal: 7.4% install→purchase | Scalable iOS channel? ($3–8K/mo path) | 🟡 30-min ASA cost pull |
| Google Search | $8 | web quiz (~5% fees) | Harvests existing demand | Buy intent cheaply into the good funnel? | 🟡 GA4→Ads import (4h) |
| Meta warm | $0 → $20–25 (~day 10) | web quiz | Only-ever signal (15 July-4, MODELLED) | Warm pool buys <$50 CPP? | 🟡 RC→AF S2S + CAPI dedup (~9h) |
| Meta cold | $0 (→ $50/day test in Aug) | web quiz | UNPROVEN — pixel Lead-trained; installs 2,878→1 | Can Meta buy purchasers <$40? | 🔴 gate: warm proven + CAPI verified |
| Google UAC | $0 | app | Blind 38d; junk geos | Junk or blind? | 🔴 LAST |
Month-1 paid deployed ≈ $40/day, not $100. The confirmed signals can't absorb $100/day yet and Meta cold can't start until wired; the unspent ~$55/day of the ceiling stays in cash = runway. That is the efficient allocation.
| Month | Total $/day | What's live | Gate to release more |
|---|---|---|---|
| Mo.1 · Jul | ~$40 | ASA brand $8 + ASA category $17 + Google Search $8 + Meta warm $20–25 (from ~day 10) | ASA CPT read · CAPI purchase event VERIFIED |
| Mo.2 · Aug | ~$85 | above + Meta cold $50/day single-question test | Meta cold CPP < $50 on ≥25 CAPI-confirmed purchases |
| Mo.3 · Sep | ~$100 | scale the winner to ceiling; ASA category scales if CPT<$4 & install→trial≥15% | blended CPP inside breaker |
| Mo.4 · Oct | $100 | full ceiling on proven channels; UAC only if rebuilt + baseline passes | paid-cohort MRR ≥ $800/mo |
Runway: full $3K/mo raises burn −£5.7K → −£8.1K/mo; two months of near-zero return = £14.8K left = ~1.8 months → mid-October, no buffer before 1 Nov break-even. Stage-gating holds back ~£3.2K/mo and still answers the biggest question (does Meta cold CPP clear ~$50?) for ~$1,000–1,250, not $3,000.
Corrected circuit-breaker (plan-mix-aware): scale <$50 · hold $50–70 · no-increase $70–87 · pause >$87. (v2's $62–98 hold loses money on non-annual plans; allowable CAC at 3-mo payback ≈$35.)
Aga set the envelope: $3,000/month (~$100/day), it must bring results. Strategy SoT amended same day. The complete strategy lives on the cockpit's 🧪 Paid Roadmap tab — this is the document of record.
Paid → web quiz → annual $157 / quarterly $50 cash upfront (Noom/BetterMe/Simple web-to-app model; ~5% fees vs 15–30% store cut, clean tracking, fast payback). App-install campaigns are NOT the scaling path. Meta fitness CPA benchmarks $22–45 → our $62–98 breakeven band is realistic.
| Period | Allocation | Gate to unlock next |
|---|---|---|
| Jul 9–31 | Meta warm retargeting ONLY ~$60/d + ASA brand defence ~$5/d (~$880 of July stays in cash) | CAPI purchase event verified end-to-end; off Lead objective; RC→Amplitude live |
| Aug | Meta retarget $40 + cold $57 (6-ad creative matrix) | CPP < $120 after first $1,500 (CAPI-confirmed) |
| Sep | Meta cold+LAL $65 + ASA category pilot $32 (RC postback live first) | Meta CPP < $100 on ≥25 RC-reconciled purchases |
| Oct | Meta scale $70 + ASA $27 | CPP < $85 · paid MRR ≥ $800/mo run-rate |
| Google UAC | LAST: wiring + trial-event rebuild + UK/US + 200-install ≥5% baseline | — |
Pessimistic 80 subs → +$1,106 MRR by 1 Nov (28% of gap) · Base 135 → +$1,914 (48%) · Optimistic 216 → +$3,168 (79%). Cash collected upfront $7.8K/$13.2K/$21.2K. Paid alone cannot close the $4K gap — R1 + email + S01 fix carry the rest.
<$62 scale · $62–98 hold · $98–130 no increase · >$130 (or >$150 with <5 purchases) PAUSE. Worst case ≈ 2 weeks of runway per month — that's what the breaker prevents.
A1 Luke injury-testimonial static · A2 Jesse founder video ("I could bench 300 pounds…") · B1 "your push-up hasn't gotten harder" static · B2 skill-tree demo video · C1 "most training fails in week 2" quiz-native · C2 five-failure-modes carousel. Kill on hook rate; refresh every 3–4 wks; naming META_INJ_STATIC_RW_2026W30. ASA listing leads with the assessment screen, never transformation photos. Anti-list: no aesthetics/aging/cold-lifetime/studio gloss/gym equipment.
CPP vs breaker → hook rates → paid S01 start-rate ≥47% → retargeting frequency <3.0 → one decision logged in the registry.
Wiring unfixed (no purchase events in Events Manager by day 7) · S01 leak uncured (paid landing→start <47% after 14d — fix the page, not the ads) · retargeting pool burnout (frequency >3.0, CPP +15% WoW).
Not 100% — base case covers ~48% of the gap. What is guaranteed if the breakers are respected: never more than $1,500 spent without knowing if it works; never scaling a number that exists in only one dashboard.
9 Jul 2026 · full /growth-fleet run (2 waves, 8 specialist briefs + skeptic + founder coach) PLANNING MODE — nothing executed. Companion: MESSAGE-TO-NIC-GOOGLE-RETARGETING-2026-07-09.md Currency USD · paid budget $3K/mo ceiling — ramp, Mo.1 ≈$40/day (Strategy SoT amended 9 Jul) · burn context: −£5.7K/mo, cash ≈£31K
RC→AppsFlyer S2S purchase postback · GA4→Google Ads conversion import ·
RC→Amplitude project confirm (T2) · label Meta's mystery pixel event + verify
CAPI purchase match-keys/dedup (event_id) · enable ASA cost in AppsFlyer (or pull
CPT from the ASA console — the single most important missing number) · GA4
source-segmented funnel dashboard. Only possible release-dependency: RC
$appsflyerId attribute (2-min check).
| # | Move | Est. MRR impact | Read window |
|---|---|---|---|
| 1 | R1 cancel-queue save (106 subs, $1,487/mo at risk — Wave A SENT 20 Jul, 24/24; B/C after the 48h read) | +$330–825/mo | 2–4 wks |
| 2 | Email retarget: saw-paywall-no-trial (~180/wk pool) | +$600–1,400/mo | 2–3 wks |
| 3 | Email retarget: signed-up-no-workout (~500/wk pool) — "your 3 free workouts are waiting" | +$200–600/mo | 2–3 wks |
| 4 | Quiz S01 hook A/B (#17 — the −51% first-screen leak) | 5–20% lift on ALL traffic | 4–6 wks |
Together: ~$1,100–2,800/mo recoverable at $0 — more than any paid experiment delivers in its window. (All MODELLED save rates; pools CONFIRMED from GA4/Stripe.)
Rule: max 2 concurrent paid experiments inside the Mo.1 ramp (≈$40/day); a third makes all undecidable.
| Experiment | The $ decision it resolves | Cost to decide |
|---|---|---|
| ASA MVE | "Is there a scalable iOS channel at allowable CAC?" (unlocks/kills a $3–8K/mo path) | ~$210–500 |
| Meta ASC MVE | "Can Meta buy PURCHASERS (not leads) at <$40?" | ~$280 |
| Google UAC rebuild | "Was UAC junk or just blind?" | $0 until wiring, then ~$150/2wk baseline |
| Free retargets | "Is the funnel leak friction or intent?" | $0 |