How much of each sale is left to cover everything else. One number, and it answers most of the questions founders take to spreadsheets.
13 min read · Level 2
Contribution margin is the question "is this sale worth making?" reduced to a single number, and it is the one piece of finance that changes daily decisions rather than quarterly ones. Should we take this discounted deal? Is this customer segment worth serving? Can we afford this channel? All three are contribution-margin questions, and founders answer them by instinct because nobody told them there was a number. Revenue tells you how busy you are. Contribution margin tells you whether being busy is helping.
Contribution margin = Revenue − Variable costs. What each sale contributes towards fixed costs, and then to profit.
CM ratio = Contribution margin ÷ Revenue, as a percentage.
It can be computed per unit, per customer, per segment, or across the whole company.
Because it is the number that answers "what if we sell more?" — and, crucially, "what if we sell more of this rather than that?"
Gross margin tells you about the product. Contribution margin tells you about the decision: whether a discount is survivable, whether a channel can pay for itself, whether a segment is worth the effort. It is also the direct input to break-even, which is the single most useful thing a small company can know about itself.
A product at £200/month. Variable costs: £24 hosting and API, £6 payment fees, £30 of support time.
Contribution margin = £200 − £60 = £140 per customer per month CM ratio = £140 ÷ £200 = 70%
So every £1 of new revenue puts 70p towards fixed costs. With a fixed base of £45,000/month:
Break-even = £45,000 ÷ 0.70 = £64,286 of monthly revenue, or about 322 customers.
Now the decisions it settles instantly:
"A prospect wants 30% off." At £140/mo the contribution falls to £80 — the CM ratio drops from 70% to 57%, and break-even rises from 322 to 394 customers. The discount is not "£60 less"; it is 72 more customers before you make a penny.
"Should we serve smaller customers at £80/month?" Variable costs do not fall proportionally — support is still £30. Contribution is £80 − £36 = £44, a 55% ratio. Viable, but each one contributes a third of what a full-price customer does, and it consumes similar support attention.
"Can we afford a channel at £400 CAC?" At £140/month contribution, payback is under three months. Yes — if you have the cash to wait.
One number, three decisions, no spreadsheet.
⚠ CM ratio is not gross margin. Gross margin uses COGS as your accountant defines it. CM uses everything variable, which usually includes sales commission and per-customer support that may sit in OpEx. CM is almost always the lower and more decision-useful of the two.
First-time founders use gross margin where they should use contribution margin, and conclude a discount is affordable because the product 'costs almost nothing to deliver'. It is the commission, the support hours and the onboarding that make a discounted deal marginal.
The second mistake: computing one blended CM ratio and treating all revenue as equivalent. Segments differ enormously — the enterprise deal with three months of implementation may have a worse contribution margin than the self-serve customer at a fifth of the price.
The third: discounting without recomputing break-even. A 30% discount does not cost 30%; it costs a disproportionate share of contribution, because variable costs do not fall with the price.
Everything that only happens because the sale happened — including commission and per-customer human time.
Per customer, per month for subscriptions; per unit for one-off sales.
This percentage is the one to remember. It is the share of each new pound that helps.
Enterprise, mid-market, self-serve. A blended figure hides the segment that is quietly unprofitable.
Fixed base ÷ CM ratio = the revenue you need. This is the whole reason to compute the ratio.
Recompute break-even at the new price. That is the real cost of the discount, and it is always larger than it looks.
Before pricing decisions, before discounting, before entering a new segment, and any time someone asks whether more volume would help.
With one or two customers there is no reliable variable cost to average. Get to a handful of customers before trusting the number, and mark it as an estimate until then.
Once a flight is scheduled, nearly all its cost is fixed — crew, fuel, gate, aircraft. The variable cost of one more passenger is close to a meal and a little fuel. So a seat sold for a fraction of the standard fare still contributes almost all of that fare to fixed costs. The same logic drives last-minute hotel pricing and off-peak discounting across capacity businesses.
Where variable cost is near zero, almost any price above it helps — as long as it does not cannibalise full-price demand. The discipline is in that second clause, which is where most discounting goes wrong.
A recognisable pattern in B2B: a company celebrates a large contract at several times its usual price, then discovers the implementation consumed four months of engineering, a dedicated support contact and a custom feature. Computed properly — with that engineering time as a variable cost of the deal — the contribution margin was worse than the self-serve customers, and it displaced the roadmap as well.
Compute contribution margin per segment before you chase a segment. A bigger number on the contract is not a bigger number on the bank statement.
Create a free account to read the rest and apply it to your own startup.