Unit economics in plain words: why the numbers lie until you measure them

Unit economics answers one question: do you earn more from one customer than you spend to get them. If yes, growth brings money. If not, every new customer makes the hole in your cash a little deeper, and the faster you grow, the faster the cash runs out.

Three numbers are enough to answer.

What it costs to bring in a customer — CAC

CAC is everything you spent on acquisition over a period, divided by the number of new paying customers over the same period. “Everything” means everything: ads, salespeople's salaries, swag at a trade show.

An example: over two months, $50,000 went on ads and $25,000 on a salesperson's salary, and 30 paying customers arrived. CAC = 75,000 / 30 = $2,500.

A common mistake is to count only the ads. Then CAC looks twice as good as it really is, and the decision “let's pour in more” seems obvious.

What a customer brings over their lifetime — LTV

LTV is how much money a customer leaves before they go. The simplest formula:

LTV = monthly payment × margin ÷ monthly churn

If a customer pays $600 a month, 70% is left after servers and support, and 5% of customers leave each month, then LTV = 600 × 0.7 ÷ 0.05 = $8,400.

Look at the denominator. Churn of 5% and churn of 10% isn't “a bit worse”, it's half the LTV. Where churn comes from and how to count it honestly — in a separate post: Customer churn: when they leave as fast as they arrive.

How to read LTV and CAC together

LTV three times CAC or more — the usual benchmark for a healthy model: acquisition money comes back with a margin for salaries and mistakes.
Roughly equal — you're working for the ad platform.
LTV below CAC — every new customer is a loss. You can't scale that; fix it first.

The second question is how many months a customer takes to pay back their CAC. In the example above a customer brings $420 of margin a month and CAC is $2,500 — it pays back in about six months. All that time the money for them has already been spent and only comes back later. That's why a fast-growing company with good unit economics can still run out: cash leaves earlier than it returns.

Where's the catch

All three numbers — CAC, margin and churn — you don't know early on. You estimate them.

There are few paying customers, and one who leaves changes churn by several percent.
It's unclear which channel brought a customer: they saw an ad, then came on a friend's advice, and got recorded as “word of mouth”.
The trickiest one: if the way you measure is skewed, repeating the measurement doesn't help. Look a hundred times the same skewed way and you'll get the same error a hundred times.

Hence the practical conclusion: analytics is an investment, not a trifle “for later”. It costs engineering time, but without it decisions on price, hiring and advertising are made by guesswork. And the later you start measuring, the more old decisions have already been made blind.

Work out LTV, CAC and payback on your own numbers — unit economics calculator.

How it works in Runway 200

In the game, unit economics isn't set with a slider — it comes out of your decisions, and you see it exactly as far as you've invested in instruments:

Without analytics you see only cash and the number of paying customers.
Cohorts (level two) show churn and how many customers made it to week eight.
Channel attribution (level three) reveals CAC and LTV.

Each level is a separate task for the engineers lasting several weeks, and it adds tech debt. So you pay for precision with the same thing you pay for the product — team time.

While the instruments are weak, the numbers on the screen are noisy, and every metric has its own skew that doesn't average out with repetition. Analytics bought later doesn't fix old measurements. If you played a run almost blind, the review at the end will say so: “Decided on noisy numbers — churn and CAC were estimates, not measurements, the whole run.”

Try it yourself: the first ten weeks are free, no sign-up. See in which week you decide it's time to pay for analytics — and what you manage to decide without it in the meantime.