Unit economics: does one sale make money before you count the rent

What a customer costs to win, what they pay, how long they stay, and how long until they have paid you back. Simple framework, and most of the mistakes come from what people leave out of each number.

Before asking whether a business can get big, ask whether one customer makes money. If a single customer costs more to win and serve than they will ever pay, more customers only dig the hole deeper.

The four numbers

Customer acquisition cost is everything you spent on sales and marketing in a period, divided by the new customers you got. Contribution per customer is what they pay you each month minus what it costs to serve them. Lifetime value is that monthly contribution times how many months they stay. Payback is acquisition cost divided by monthly contribution, the months until a customer has repaid what it cost to win them.

A made up example

A meal kit service spends $30,000 a month on marketing and gains 500 customers. Acquisition cost is $60. Each customer pays $120 a month and costs $90 in food, boxes, and delivery, so contribution is $30 a month. Payback is two months. If the average customer stays eight months, lifetime value is $240, four times the cost to win them. If customers quit after three months, lifetime value is $90, and the business is spending $60 to earn $90 before a dollar of rent, salaries, or software. The number that is hardest to know early, how long they stay, is doing all the work.

The mistakes are predictable. Acquisition cost gets understated by leaving out the salaries of the people running marketing, or by dividing spend by all customers rather than new ones. Contribution gets overstated by calling customer support, returns, and card fees fixed. Lifetime gets overstated by extrapolating from the first customers, who are usually the most enthusiastic. Each error flatters the ratio, and they compound. A lifetime value three times acquisition cost that survives scrutiny is a different thing from a three built on hopeful inputs.

Where the tidy version goes wrong

"We will make it up in volume" is only true if each sale contributes something positive. Volume multiplies whatever sign the unit carries. And "our lifetime value is high" means nothing without the quitting data behind it and the time it took to collect that revenue. A customer who pays $1,000 over ten years is worth less than one who pays $600 over one, once you count the cost of waiting for the money.

Questions worth asking yourself

What is your fully loaded cost to win a customer, people included? What does it truly cost to serve one, support and returns included? How long do customers actually stay, measured on a group who started together, not guessed? And how many months until a customer pays back what it cost to win them, and can you fund that gap?

Sources

Written by one person and checked against the sources above; no outside expert has reviewed it yet. Rules and dollar limits change every year, so this guide explains how things work and sends you to the official source for this year's numbers.