Skip to content

3 · The numbers that matter: unit economics, CAC, and LTV

Big totals can hide the truth. The question that actually tells you if a business works is small: does one sale, to one customer, make money? That's unit economics — the profit (or loss) on a single unit or a single customer. Get this right and scaling makes you richer; get it wrong and scaling just makes you lose money faster.

Three founder numbers, in plain terms:

  • Unit margin — profit on one sale. If a candle sells for $15 and costs you $6 to make and ship, your unit margin is $9. If it costs you $16 to make, you lose a dollar on every sale — and selling more makes it worse.
  • CAC — Customer Acquisition Cost — what it costs, on average, to get one new customer. Spent $50 on ads and got 10 customers? Your CAC is $5 ($50 ÷ 10).
  • LTV — Lifetime Value — the total profit one customer brings over the whole time they buy from you. A customer who buys 4 candles a year for 2 years, at $9 margin each, has an LTV around $72.

The one comparison that matters: is LTV bigger than CAC? If a customer is worth $72 to you and costs $5 to get, that's a healthy business. If they cost $80 to get and are worth $72, you lose money on every customer you "win." A common rough rule of thumb founders use is wanting LTV to be several times CAC — but the exact ratio matters less than the direction: value in must beat cost to acquire.

Where AI helps vs. fools you here: AI is great at explaining these terms and setting up the formula in a spreadsheet. It is bad at giving you the actual numbers — your real costs, your real ad spend, your real repeat rate come from your records, not from a model's guess. And if you ask AI to calculate a ratio, verify it: a wrong CAC tells you to scale a business that's secretly losing money.

Trust DNA: AI can teach you what CAC and LTV mean and how to compute them. The actual inputs must be your real, verified numbers — and you re-check the arithmetic.

Check yourself. Why is "we're selling a lot" not enough on its own — and which two numbers do you compare to know if getting customers is actually worth it?

Sources