How to run a unit economics review before you raise money
Work out contribution margin, acquisition payback and churn-adjusted lifetime value from your own records, and write down each assumption before an investor tests it.

Most founders who walk into a funding conversation can recite their revenue growth. Far fewer can say what one new customer costs them, how many months that customer takes to repay the cost, or how many customers leave before they do. A unit economics review answers those three questions from your own ledger and your own customer records. It should be finished before an investor asks the questions for you, because the answers you give under pressure are rarely the ones you would have chosen.
Before you start
Three things must be true before the arithmetic means anything. First, you need one agreed definition of a customer: a paying account, a signed contract, or a person who has used the product. Second, your revenue and cost figures must come from the accounts, not from a spreadsheet built for the last pitch deck. Third, you need enough history to see a pattern. One month of data shows a result; several months of cohorts show a trend. The review is only as good as its inputs, and the usual failure is to skip them and go straight to the formula.
- A revenue schedule by customer and by month, reconciled to the general ledger.
- A list of every cost that rises when you serve one more customer.
- Acquisition spend by month, including the share of sales salaries and commissions that belongs to new business.
- Start and end dates for every contract or subscription in the period.
- The discount and refund history for the same period, so that net revenue can be checked against gross.
- Step 1
Fix the definitions
Write down what counts as a customer, a new customer, a churned customer and a month. Decide whether a discounted pilot is a customer, and whether a refund reverses revenue in the month it happens or the month of the sale. The usual mistake is not the choice itself but switching between choices from one slide to the next. Keep the definitions on one page and attach it to the review.
- Step 2
Work out contribution margin
Contribution margin is what remains from each unit of revenue after the costs that rise with serving a customer. In our working form, take revenue per customer, subtract the variable costs of serving that customer, and divide by revenue. Payment fees, hosting, onboarding labour and support often belong in the variable column; rent and founders' salaries usually do not. Calculate the figure for each plan or segment, because a blended margin can hide a product that loses money on every sale. Do no
- Step 3
Measure acquisition cost and payback
Add up the spend that brought in new customers in a period: marketing, sales salaries, commissions and the tools used to sell. Divide that total by the number of new customers won, which gives acquisition cost per customer. Then divide acquisition cost by the monthly contribution one customer produces to get the number of months to payback. Match spend and wins to the same period, or lag the spend by your typical sales cycle and state that you have done so. Founders often leave out the sales tea
- Step 4
Measure churn from cohorts
Group customers by the month they started and track how many remain in each following month. An average churn rate can look healthy while the newest group leaves faster than the older ones, and an average hides that difference. Show the curves for at least the last several cohorts, side by side. If your product is billed annually, measure renewals rather than monthly exits, and say which measure you used.
- Step 5
Build lifetime value with churn in it
Our working form is monthly contribution per customer divided by monthly churn rate. It assumes the rate stays flat and it ignores the timing of future cash, so it flatters a business whose churn is rising. Show it next to a second version that caps a customer's life at a fixed number of years. If the answer depends heavily on a churn rate measured over only a few months, say so in the review rather than letting the figure carry the argument.
- Step 6
Write down every assumption
Make a table with one row per assumption, listing its value, its source, the date it was measured and the person who owns it. Price, churn, margin, sales cycle and the share of spend attributed to new customers are the usual rows. Investors will test these rows one at a time, so the document must show which figures are measured and which are estimated. A labelled estimate reads as honest. An estimate presented as a measurement is the kind of problem that, in our experience, surfaces late in dili
- Step 7
Test what breaks the answer
Move each key assumption up and down by a plausible amount and see which ones change payback or lifetime value the most. Those are the figures to defend first. If one unmeasured number decides whether the business looks healthy, it belongs at the top of your measurement list, not in an appendix. This step also tells you which numbers need a second source before a meeting.
- Step 8
Get an outside read
Ask your accountant to check revenue and cost inputs against the ledger. Ask someone who has sat on the investing side of a funding table to read the assumptions document and mark every line they would challenge. Questions about tax, accounting treatment or the terms of any instrument belong with a qualified accountant or lawyer, not with the review itself. The cost of an outside read is usually small next to the cost of a correction that arrives in the middle of a process.
The mistakes that cost the most
The mistakes cluster in three places. Founders mix definitions across slides, so that a customer in one chart is a pilot in another. They report averages that hide the newest cohorts, which are usually the ones an investor most wants to see. And they present estimates as measurements, often without noticing the difference. A quieter mistake concerns the source of the money. The federal investor education site investor.gov advises caution with investment offers you did not seek out, and it lists red flags to watch for. A review that prepares you to pitch also prepares you to read an unsolicited offer more carefully. Put any offer you did not ask for through the same scrutiny you would give your own numbers: who is offering, what they want from the business, and what they have actually checked.
A review can also end the case for a raise. Research published by the National Bureau of Economic Research, in a randomised, multi-year field experiment on innovation-driven entrepreneurs, found that effective training may cause some startups to shut down sooner while improving outcomes for others. The same research describes training as a way for founders to recognise when a venture no longer warrants further investment of time and resources. Treat a weak result in this review as information for the next decision, not as a verdict on the founder. Some businesses should fund a fix before they fund growth, and a few should stop.
Be wary of investment opportunities that you didn't seek out.
- Which figure should I show first?
Lead with acquisition payback in months, because it tells an investor how long the cash spent to win a customer stays out of the business. Contribution margin and lifetime value should follow, each with its assumptions beside it. Showing lifetime value first invites the question of how it was built, and that is harder to answer if you have not yet checked the inputs.
- Do I need a finance hire to do this?
Not necessarily for a first pass, but someone who can read the ledger should reconcile the revenue and cost inputs. An outside accountant can do that check for a fixed fee. The founder should still be able to explain every line of the calculation, because an investor will ask about the lines the founder cannot explain.
- What if my churn data covers only six months?
Say so in the document and show the cohorts you have. Present lifetime value as a range, using the lowest and highest churn figures that the data supports, rather than as a single number. A short history is a limitation to disclose, not a reason to leave the review out.
- Should I include forecast customers in the numbers?
Keep actual results and forecasts in separate columns, with the forecast clearly labelled. Mixing them makes every figure harder to check. If a forecast drives the payback calculation, list it as an assumption in the table and test how far the answer moves when it is wrong.
- What if the review says the business is weaker than I thought?
That is the purpose of the review. The National Bureau of Economic Research describes effective founder training as something that may cause some startups to shut down sooner while improving outcomes for others. Use the result to decide what to fix, what to measure and whether to raise at all. Take questions about legal or tax consequences of any wind-down to a qualified adviser.
What to do this week
Pull the last three months of revenue and direct cost by customer, and write the definitions page before you open the model. Calculate contribution margin for your largest plan only, and send the result to your finance lead or accountant before you build anything else. Hold the payback and lifetime figures until the inputs have survived one honest check. Once they have, build the assumptions table, mark every estimate, and find the two numbers that decide the answer. Those two are the ones to measure properly before the first meeting, and the ones an investor is most likely to ask about.