Contribution Margin

There was a long stretch where nobody could explain a company whose revenue was climbing while its bank balance fell. The quarterly report says customers grew 30 percent, and the line underneath says the loss grew too. You want to argue that growth makes it fine, but there's nothing to argue from; you want to call it dangerous, but you can't point at which part is dangerous.

The cause was treating cost as one lump. If $150,000 went out last month, that figure mixes money that leaves at roughly the same rate whether there are 100 customers or 200 with money that rises as customers are added. Office rent and engineering salaries are the first kind; server usage and payment processing fees are the second. Blend them and there is no way to work out what one additional customer leaves behind.

So the question that kept coming up in investor reviews was whether there's money left over on one more customer. Answering with a percentage conveys that something is left but not how much. Contribution margin answers in currency. Once you know what one customer actually contributes toward the company's fixed costs, the number of customers it takes to end the losses follows directly.

Take the revenue from one unit and subtract the variable costs that unit caused. The unit can be an order in some businesses; in a subscription business it's usually one customer account.

1. Write down what that unit pays per month. For a Pagedeck team-plan account, $300.
2. Subtract only the costs that rise because of that account. Servers and storage at $45, payment processing at $9, and the support time carried for that account at $21 come to $75.
3. The remaining $225 is the contribution margin. Divided by revenue it's 75 percent, which is called the contribution margin ratio.

One test sorts the costs. If it rises when one customer is added, it's variable; if it doesn't, it's fixed. An engineering salary is identical at 101 customers, so it's fixed. Server usage climbs with the data stored, so it's variable. For anything ambiguous, check whether it actually moved with customer count over the past six months.

The first place this gets used is the break-even point — where money coming in equals money going out — which is just fixed costs divided by contribution margin. Pagedeck's fixed costs run $67,500 a month and each account contributes $225, so break-even sits at 300 accounts. At 180 accounts today, 120 more are needed, and that figure becomes the sales target.

Redo the arithmetic whenever fixed costs move. Hire ten more people, push fixed costs to $90,000, and break-even climbs to 400 accounts. That's the real value of having the number: a hiring decision arrives with the customer count required to carry it attached.

Look at the share of revenue before the amount. Knowing that $225 is left tells you nothing on its own. Charging $300 and keeping $225 is 75 percent; charging $1,500 and keeping the same $225 is 15 percent, and that's a different business.

Software subscription businesses commonly land between 70 and 85 percent, because one product built once gets shared across customers and each new one adds little cost. A number down in the 40s usually means one of two things: someone is sitting with every customer doing setup by hand, or pricing has fallen behind usage and heavy accounts are running up server bills the company absorbs.

Below zero the number changes character. Each additional customer deepens the loss, so pausing sales is the profitable move. This isn't rare early on. A generous free tier with real usage, or an enterprise account given a deep discount plus a dedicated person, will push that account's contribution margin negative. The response is fixing the pricing, not selling harder.

Read it alongside customer acquisition cost (CAC) payback and it earns its keep twice. Payback divides the customer acquisition cost spent on one account by the profit that account returns each month, and the number in the divisor is the contribution margin. Spend $2,700 to win an account contributing $225 a month and payback takes twelve months; lift the contribution to $300 and the same spend clears in nine.

"This plan has a negative contribution margin. As it's structured, every sale costs us money."

The variable costs the customer causes exceed what they pay, so pricing has to change.

"Fixed costs are $67,500, so at $225 each we break even at 300 accounts."

Feeding the contribution margin straight into a break-even calculation to get a customer target.

"Let's cap enterprise discounts so the contribution ratio never drops below 60 percent."

Discount, but put a floor under what a single account contributes.

The number becomes the meeting topic whenever Pagedeck has an annual enterprise contract in front of it. The terms Sofia brings back are half list price with a dedicated contact, and on revenue alone it's a welcome deal. What Grace checks is what one account contributes under those terms. The dedicated person's time is a cost caused by that customer, so it belongs in variable costs, and the arithmetic lands the ratio in the 30s.

The conclusion wasn't to walk away from the deal. It was to limit the dedicated support to the first three months and move the account to standard support after that. Revenue alone would never have produced that adjustment.

The most common error is reading a positive contribution margin as profit. The word margin makes it sound like money in hand. But the figure comes before fixed costs, and the word contribution refers to contributing toward paying those costs. A hundred accounts producing $22,500 against $67,500 of fixed costs is a $45,000 loss that month. Stop at this number and you miss when the cash runs out.

The other mistake is counting only server bills as variable. They arrive on their own invoice, which makes them easy to count. But payment processing fees, the support hours attached to each account, and per-contract third-party tool licences all rise with customer count too. Leave them out and the ratio reads 10 to 20 percentage points too high, which makes break-even look much closer than it is — you plan for 300 accounts and don't actually clear the line until 380.