Unit Economics Review — by an AI council
Find out if your business actually makes money per customer — and when. Four rival AIs debate it and hand back: A clear read on CAC, LTV, margin and payback, with the lever that fixes it.
A clear read on CAC, LTV, margin and payback, with the lever that fixes it.
Signed off by an independent Final QA audit — with a “verify before you rely on this” list, never a black box.
Why this team
Unit economics flatter their author: definitions drift, averages hide cohorts, and the deck version of CAC rarely survives an audit. So the Forensic Accountant re-derives the numbers from their definitions, the Data Skeptic hunts the cohort behaviour underneath the averages, and the CFO judges whether the machine makes money at scale. The Pricing and Retention Strategists own the two levers that usually fix broken economics, and the Devil’s Advocate argues the bear interpretation of every ratio.
Each seat runs on a different frontier model — GPT, Claude, Gemini, Grok — so it’s a genuine cross-model review, not one AI in costumes. How a run works →
What the debate sounds like
An illustrative excerpt — A subscription meal-kit company reports LTV:CAC of 4:1 in its board deck, but blended CAC mixes paid and organic, and LTV is computed from the twelve-month-old cohorts only.
Two definitional problems make the 4:1 unusable. Blended CAC averages cheap organic customers into the paid number — on marginal paid spend, the true acquisition cost is what the next dollar buys, and that is what decides whether to scale spend. And LTV from the oldest cohorts assumes today’s customers behave like the early adopters. Recompute both before this ratio appears anywhere.
The cohort curves in the appendix contradict the headline: recent cohorts churn faster in their first eight weeks than the old cohorts did. If that trend holds, the real forward-looking LTV:CAC is materially below the deck number. The average is not lying, exactly — it is answering a question nobody should be asking.
The interesting fact is where the churn concentrates: the first month, around the moment the novelty fades and the price renews. That is not an economics problem, it is an onboarding problem wearing an economics costume — the highest-leverage fix is the week-three experience, not the acquisition budget.
Questions
Which numbers do I need before running this review?
Revenue per customer, gross margin, CAC (split paid versus blended if you can), churn or retention by cohort, and the payback period as you currently calculate it. Imperfect data is expected — identifying which of your numbers cannot be trusted is explicitly part of the deliverable.
My LTV:CAC looks great — why would I stress-test it?
Because the flattering version of the ratio is the default output of every dashboard: blended CAC, average LTV, oldest cohorts. If the number survives re-derivation from definitions, you get to trust it — and a stress-tested 3:1 is worth more in a board meeting or a fundraise than an unexamined 4:1.
Does the review tell me what to fix, or just re-audit the maths?
It commits to the single biggest lever — pricing, retention, margin or acquisition cost — with the expected effect and the reasoning, because “improve everything” is not a plan. In the example above, the verdict names week-three onboarding, not ad spend, as the fix.
Your material is used only to run your review — never to train public models. Encrypted in transit and at rest. Security & privacy →
Want full control — pick your own minds, set the depth? Open the full council →