The split that decides what happens when you grow, shrink, or have a bad quarter. Most founders classify their own time as neither, and that is where the trouble starts.
11 min read · Level 2
Fixed costs are promises; variable costs are consequences. A variable cost only happens because you sold something, so it is never a decision you regret. A fixed cost happens whether or not anyone buys, which means every fixed cost you add is a bet that revenue will show up. That is why hiring feels different from raising ad spend even when the monthly number is identical — one you can stop on Friday, the other you have made a commitment about someone's life.
Variable costs move with each additional sale: payment fees, per-seat licences, hosting that scales with usage, shipping, delivery labour.
Fixed costs do not, at least in the short term: salaries, rent, insurance, most software.
The boundary is a judgement, not a fact, and it holds only within a range and a timeframe.
Because it determines what happens to profit when revenue moves, which is the most important thing to know about your own cost structure. High fixed costs mean profit swings violently in both directions. High variable costs mean a flatter, safer, less spectacular business.
And because it is the input to everything downstream — contribution margin, break-even, operating leverage. Get the split wrong and all three are wrong, confidently.
A software company, £48,000 of monthly revenue:
Variable — payment fees £1,300, hosting that scales £3,400, per-seat API £900. £5,600, about 12% of revenue.
Fixed — salaries £39,000, rent £2,400, base infrastructure £2,000, software £1,900, insurance £400. £45,700.
So roughly 88% of every extra pound of revenue drops through to cover fixed costs. That is a high-leverage structure: wonderful above break-even, brutal below it.
Now the same numbers for an agency at £48,000 of revenue: delivery salaries of £30,000 are variable in substance — take on twice the work and you need twice the people. Only £9,000 is genuinely fixed.
Same revenue, completely different companies. The software company should protect its fixed base and push volume. The agency should protect its margin per project, because volume alone will not save it.
⚠ The trap in both cases is founder time. Six hours of your own onboarding work per customer is a variable cost whether or not it appears on the P&L, and treating it as free makes the software company look like it has agency economics discovered only after the first hire.
First-time founders classify by account name rather than behaviour. "Salaries" goes to fixed, always — but delivery salaries at a services business are variable in every way that matters, and calling them fixed makes the break-even calculation nonsense.
The second mistake: treating founder time as neither. It is the largest variable cost in most early companies and the only one nobody records.
The third: forgetting that fixed costs are fixed in steps. Hiring does not raise your fixed base smoothly; it raises it in a jump you cannot easily undo, which is why a hire deserves more analysis than a marketing budget of the same size.
Which costs roughly double? Those are variable. Which stay flat? Fixed. Ignore what the account is called.
If a human must do work per customer, that is variable, whoever absorbs it. Price it at what you would pay someone.
Fixed costs jump — at the next hire, the next office, the next licence tier. Write down where the next step is.
This is what break-even is calculated against, and it is the number that tells you how bad a bad month can be.
Moving from services to product, or from self-serve to enterprise, moves the boundary substantially.
Before computing contribution margin or break-even, and before any hire.
Do not spend hours on borderline classifications pre-revenue. Roughly right is enough; the ratio matters far more than the last few hundred pounds.
The structural difference is entirely here. A software company's costs are mostly fixed, so each additional customer is nearly pure contribution — the margin improves as it grows. A services company's costs are mostly variable, so the hundredth customer is about as profitable as the tenth. Both can be excellent businesses; only one improves with scale, and that is what venture capital is buying.
Before deciding whether to raise, work out which shape you actually have. The answer is in this split, not in the pitch.
The recurring pattern in early-stage companies: gross margin looks like 85% because the founder personally does onboarding, support and implementation for free. The first hire into that role reveals the real number, often ten to twenty points lower, and it arrives as a nasty surprise in the same month payroll went up.
Cost your own per-customer time at replacement rate from the start. You will hire eventually, and you want the margin shock to happen on a spreadsheet rather than in the bank account.
Create a free account to read the rest and apply it to your own startup.