00ROAS verdict — 16 Aug 2026 audit

Ade asked: do we concur with 4:1? No. 10-agent research sweep — 7 evidence angles, 97 sourced findings, 3 adversarial audits (source fidelity · metric definition · applicability). Every load-bearing figure opened and checked at source. Full detail: ROAS_AUDIT.md.

Where the 4:1 actually dies — the conversion leg

4:1 at $250 and $1.25 CPC is the 2.0% cold click-to-purchase assumption. Verified reality: stores with AOV >$200 convert at a 0.79% site-wide median including warm traffic (DTC Pages Q2 2026); cold paid-social course pages at $97–297 run 0.5–1.5%; Meta traffic is 86–94% mobile, converting ~40% below desktop; and realistic health-vertical CPC is $1.38–2.07, not $1.25. Holding 4:1 at realistic CPCs needs 2.8–3.3% cold mobile CVR — 3.5–4× the observed median.

Three structural findings

• Meta health restrictions hit this product squarely. Menopause meets the "specific health status" definition; education positioning does not escape it. 84% of 159 women's-health orgs had ads rejected; agencies call 1.8–2.5x "realistic at scale" for remediated accounts. Remediation infra ($2–5k + $500–1k/mo) now budgeted into month 0.
• No precedent in the menopause market. The category leader (Haver, $43M/yr) prices her course at $99 and rides 3M+ warm followers; every $200+ course sells to a warm audience; the one paid-acquisition DTC menopause brand went through Chapter 11. Nobody sells $250 one-off cold at scale.
• Every verified 4:1+ case rests on assets this project lacks at launch — warm audience, email list, retargeting pool, low-ticket entry or high-ticket backend. 4:1+ is real only as a 6–12-month cohort outcome after building those assets.

Consequence for the £17k/month objective (refunds corrected to 8%)

Bottom line: not fucked — but it is a different play than Manus modelled. A flat $250 page on cold traffic cannot carry £17k/month alone. What changes: (1) offer re-architecture becomes the critical path — order bump + upsell (~$287–300 effective AOV), email spine from day one, VSL/webinar test; (2) price architecture gets tested against the market's proven $99 anchor — Stage 1 of the ladder is now a live price test, and a stacked $99–149 front-end may out-earn a flat $250; (3) Gemma-fronted US organic content is the mechanism behind every verified 3x+ case — it moves from garnish to core; (4) the engine's Money Plan role resets to a £9–13k/month line by month 15–18 at audited central, with upside from the levers, not from hoping the old 4:1 shows up. The stage-gate ladder survives unchanged — discovering the true number still costs ~£2–3k, and Stage 1's live data replaces every benchmark on this page.

01Executive answer — engine mechanics by ROAS level

Counting from the first paid month (after a 6–10 week build/entity phase). The bottleneck is never cash and never audience size — it is platform scaling speed and whether ROAS holds. Post-audit (§00): read the 2.0–2.5x rows as the central planning rows; the 4:1 row is the top-decile case, kept for reference.

Calendar terms: ~2 months of product build and entity resolution, then 12–16 months of compounding ⇒ 12/day around late 2027 if ROAS proves ≥3:1, full £17,000/month extraction from roughly Q1 2028. Faster only if 4:1 proves early and Meta permits aggressive budget growth.

02Reality gates the Manus pack could not know

Local knowledge that reshapes month 0. These are sequenced before the first paid dollar.

03Offer & validation ladder

Free users do not behave like $250 buyers. Engagement ≠ testimonial ≠ paid conversion ≠ retained purchase. Each stage produces the evidence class the next stage needs — and three cheap kill-switches precede the expensive 300-purchase full proof (~US$18,750 at target CPA).

Payment-plan sales count at 90% value in ROAS maths (instalment default risk). Test the plan only after Stage 3 stabilises.

04The compounding engine — central 4:1 path

$25/day founding months, $50/day from month 3, budget grows ≤40%/month and never exceeds 60% of opening cash. ROAS decays with scale (4.0 → ~3.3 at ~$1,000/day). Overhead £1k/mo → £3k/mo once spend ≥$250/day. 25% Corporation Tax reserved monthly.

Gate months across scenarios (first paid month = 1)

Scaling-aggression sensitivity at 4:1: cautious +25%/mo → 12/day month 17 · central +40% → month 12 · aggressive +60% → month 10. Full monthly CSVs in menopause-us-engine/model/.

05Minimum starting cash

Binary-searched minimum float keeping cash ≥ £0 throughout (central decay + scaling). The engine self-funds at ≥2.5:1 because every $1 of ads returns $2.24+ of net receipts in-month — the float covers timing risk, not losses.

06Reinvestment waterfall & hard rules

Order of application of every month's net receipts. No discretion in the moment — the rules are set now, followed later.

07Review generation — genuine, disclosed, compliant

Governing rules to re-verify before launch: FTC Rule on Consumer Reviews & Testimonials (16 CFR Part 465, in force Oct 2024 — civil penalties for fake/purchased reviews, sentiment-conditioned incentives, review suppression) and the FTC Endorsement Guides (material connections disclosed; typical results; health claims need competent, reliable scientific evidence).

1. Founding cohort gets free access never conditioned on any review — and never on a positive one. Ask: "share your honest experience, positive or negative".
2. Every founding testimonial displayed carries: "Received free early access."
3. Collect via an independent platform (Senja/Trustpilot) + in-product survey; keep consent records.
4. No outcome claims ("cured my hot flushes" cannot run). Dr Gemma Lewis reviews every testimonial pre-publication.
5. Negative feedback is product R&D. It is never suppressed — Part 465 explicitly bans suppression.
6. Unincentivised paid-stage reviews become the primary social proof; founding reviews rotate out.

08Meta architecture & creative cadence

09Landing page, checkout & email architecture

10Tax & entity

• US: no federal VAT; state-by-state digital sales tax — the MoR route removes registration/remittance risk entirely at launch. Direct-Stripe economic nexus (~$100k or 200 transactions/state/yr) arrives fast at 12/day ($1.1M/yr) — MoR stays sensible longer than intuition suggests.
• UK VAT: US B2C sales are outside UK VAT scope. UK buyers (minority) attract 20% — MoR handles it.
• Corporation Tax: 25% main rate at modelled profits — reserved monthly in the waterfall.
• Extraction split: per the handover's corrected bridge — £32,540 employer pension + dividends, until AWR confirms actual taper headroom (threshold/adjusted income, carry-forward).
• Entity: model is entity-agnostic; confirm the vehicle before the first sale (§02).

11Key risks & mitigations

12Five decisions for Ade — everything else runs on labelled assumptions