Unit economics: the three numbers that tell you if you're actually making money
What each customer costs to win and serve, what they're worth, and how long before they pay you back. The three numbers behind every pricing decision.
Unit economics has a reputation as a startup metric — something venture investors ask about. That framing has done real damage, because the underlying question is one every business needs answered: does one more customer make me better off or worse off?
You can be profitable in aggregate and still be losing money on some of your customers. Growth then makes things worse, quietly, while revenue goes up and everyone congratulates each other.
The three numbers
1. What it costs to win a customer
Total sales and marketing spend in a period, divided by customers acquired in that period. Include everything: advertising, the commission, the proportion of your own time spent selling, the tools.
The most common miss is the owner's own selling time. If you personally spend two days a week selling, that's a real cost even though it doesn't appear anywhere on the P&L.
2. What a customer is worth
Gross margin per customer per period, multiplied by how long they stay. Note gross margin, not revenue — a customer paying $2,000 a month who costs $1,400 to serve is worth $600 a month to you, not $2,000.
For lifespan, use what actually happened. If your average client relationship has run three years, use three years. Don't use the longest one.
3. How long until you're square
Acquisition cost divided by monthly gross margin. If it costs $6,000 to win a client who contributes $1,000 a month, payback is six months.
This is the number that matters most for cash, and the one most often ignored. A customer who is highly profitable over three years but takes fourteen months to repay acquisition cost is a cash problem right now, however good they look in the lifetime figure.
How much to trust the lifetime figure
Lifetime value is the softest of the three numbers, and it is worth knowing why before you make a decision on it.
If you derive lifespan from churn — the usual shortcut, where a 2% monthly churn implies a fifty-month life — you have assumed churn stays flat forever. It rarely does. Most businesses lose customers fastest in the first few months and then the survivors stick, so a single blended rate overstates the life of new customers and understates the life of old ones.
If instead you use observed average tenure, you have a different problem: you can only observe customers who have already left. A four-year-old business cannot produce a five-year average, so young businesses systematically understate lifespan, and businesses that have just been through a bad patch overstate their churn.
The honest version is to look at cohorts — group customers by the quarter they joined and watch each group's retention age — rather than one number for everyone. Short of that, treat lifetime value as a range, and lean on payback period for anything with a cash consequence. Payback is measured from real money that has already moved, so it is the number you can act on with the most confidence.
What good looks like
Rules of thumb, not laws:
- Value to acquisition cost of 3:1 or better. Below that, too much of the customer's worth is consumed winning them.
- Above roughly 5:1, read it as a spending signal rather than a gold star. A very high ratio usually means you are underfunding acquisition, not that you have found something free. Either retention really is that strong, in which case the question is how much more you could profitably spend to win customers, or the lifetime figure is optimistic for one of the reasons above. The calculator flags this the same way, which is deliberate — a 16x ratio is an instruction to investigate, not to celebrate.
- Payback under twelve months. Under six is strong. Beyond eighteen, you are financing your own growth for a long time and need the balance sheet to support it.
- Gross margin stable or improving as you scale. If it erodes as you grow, you're buying revenue by discounting or by serving customers you're not set up to serve.
Cut it by segment or don't bother
This is where the analysis earns its keep. A blended average across the whole business almost always hides the answer.
Cut it by service line, by client size, by acquisition channel, by industry. An illustrative example of what the cut can show: referred clients pay back in three months and stay four years; clients from the paid channel pay back in eleven months and leave in eighteen. Both are inside the same average, and only one is worth more of.
Blended unit economics tell you the business is fine. Segmented unit economics tell you what to do on Monday.
The decisions it changes
- Pricing. If payback is too long, a modest price increase usually fixes it faster than cutting acquisition spend.
- Where to spend. Put money into the channel with the shortest payback, not the largest volume.
- Who to fire. Some segments are worth declining. That's easier to act on when you can see the number.
- How fast to grow. Payback period times monthly acquisition spend tells you roughly how much working capital growth will consume.
Start rough
Perfect cost allocation isn't the point and will stop you finishing. Take a quarter, split customers into three or four sensible groups, and work the three numbers for each. The differences between groups will be large enough that precision doesn't change the conclusion.
The unit economics calculator takes monthly revenue per customer, gross margin, monthly churn and acquisition cost, and returns lifetime value, the ratio and the payback period. It will run one segment in a couple of minutes. Note that it derives lifespan from the churn rate you give it, so the caveat above applies to its lifetime figure as much as to your own.
