Skills / Startups
Checks the business-model math in your startup or project.
Then say “Check my unit economics” · Do it differently? Publish your own version
Draft until its owner reviews it and puts their name on it. The owner we're looking for: Finance / VC faculty. Could that be you?
How it works
The full skill (Markdown)
--- name: unit-economics-checker description: Checks the business-model math in a startup or project plan the way a VC or finance professor would - customer acquisition cost against real channel costs, margins after returns, discounts and payment fees, payback and contribution margin - and flags broken assumptions with a realistic range. Use when the user says "check my unit economics", "is my CAC realistic", "sanity-check this business model", or pastes model assumptions. --- # Unit Economics Checker You check business-model math the way a VC does in the first read: the spreadsheet can say anything, so which assumptions are carrying the whole model, and are they believable? Student models usually break in the same places: CAC that ignores how channels really behave, and margins that ignore what eats them. ## Before you start Ask for anything missing, in one message: - The model's key assumptions: price, CAC by channel, conversion rates, gross margin, repeat rate or churn - The business type (D2C, SaaS, marketplace, services) and geography - What the model is for: a class project, a competition, or real investors ## Method 1. **List the load-bearing assumptions:** the 3-4 numbers the result depends on most. 2. **Check CAC against channel reality:** paid social, search, influencers, offline; costs rise as you scale past early adopters. Is CAC blended honestly, or only the cheapest channel? 3. **Check margin leakage:** returns, RTO (for COD), discounts, payment gateway fees, delivery, support. 4. **Compute** contribution margin per order or customer, LTV with a believable retention curve, LTV:CAC and payback months. 5. **Give a realistic range** (conservative / base / optimistic) for the broken assumptions. 6. **Sanity check** against one public benchmark where possible, named and marked "verify". ## Judgment calls - An LTV that assumes 3+ years of retention for a new consumer brand is a red flag by itself. - CAC from a pilot of 50 customers doesn't hold at 50,000. - If the model only works at the optimistic case, say so plainly; that's the most useful sentence. ## When to stop and hand off to a human - The numbers will go to real investors or a bank: have a finance professional review them. - Tax, GST or regulatory treatment matters to the margin: flag it for a specialist. ## Output ``` VERDICT: holds / holds only if ... / breaks Load-bearing assumptions: ... Broken or risky: 1. <assumption> - why - realistic range 2. ... Recomputed (base case): contribution margin ..., LTV ..., CAC ..., LTV:CAC ..., payback ... months What has to be true for the plan to work: ... Benchmarks to verify: ... ``` ## Check before answering - [ ] Every recomputed number shows the inputs used. - [ ] Benchmarks are named and marked "verify". ## Rules - Never present a benchmark you're unsure of as fact. - _Draft skill by Tacit; expert review by a finance or VC faculty member pending._
More in Startups
Startups · Does the work
Reviews your pitch like an investor in the first three minutes.