Your gross margin is probably wrong

Gross margin is the number leadership teams quote most often and interrogate least. It appears on the first page of the pack, it is compared to a benchmark somebody read in an industry report, and it is used to justify decisions about pricing, hiring, and which parts of the business to grow. Very few of the businesses we come into have ever gone back to check what is actually in it.

This matters because gross margin is not a rule handed down by accounting standards. Above a small core of obvious direct costs, it is a set of choices your predecessors made, often quickly, often years ago, and usually without writing down why.

The line is a judgement, not a standard

In a service business, is the delivery manager a cost of sale or an overhead? What about the project management tool the delivery team lives in, the recruiter fee for a billable hire, or the time consultants spend on internal work? In an online business, do you put customer support above the line? Payment processing fees? Hosting? The engineering team that keeps the platform running?

Reasonable people answer these differently, and every answer produces a different gross margin from identical underlying economics. That is fine. What is not fine is answering them inconsistently, or not knowing which answer your own accounts have chosen.

A test that settles most arguments

The question we use is simple: if you won one more unit of this work tomorrow, would this cost go up? If the answer is yes, it belongs in cost of delivery. If you lost a third of your customers, would this cost fall within a quarter? If yes, again, it belongs above the line. Costs that sit flat whatever happens to volume are overheads, however close to the customer they feel.

• Billable salaries, contractors, and the employment costs that go with them sit above the line, including the portion of a manager who is genuinely billable.
• Software licensed per seat or per customer sits above the line. Software licensed for the firm sits below it.
• Hosting, payment processing, and third-party data that scale with usage sit above the line in an online business.
• Recruitment, training, and idle time are overheads. Putting them in cost of delivery hides a utilisation problem inside a margin number.
• Whatever you decide, write it down in a one-page policy and apply it to the prior year as well, or you have created a trend that is not real.

One number is not enough

Even a correct blended gross margin is close to useless on its own. It is an average of things that behave differently: a service line running at fifty per cent and one running at twelve, a large account you discounted to win and a small one paying list price, a new product still carrying its launch costs. The average moves and nobody can say why, because the average was never the thing that moved.

Margin has to be cut by the dimension the business is actually managed on. For most of our clients that is service line or product, then customer, then delivery team. Getting there is a data exercise before it is an accounting one: revenue and cost both need to carry the same tags, which usually means tracking categories in the ledger, a consistent job or project code, and a payroll allocation that is reviewed rather than inherited.

What changes once you can see it

The pattern we see most often, once margin is split properly, is that a meaningful slice of revenue is being delivered at or below cost, and it is not the slice anyone expected. Usually it is a long-standing account that was priced years ago and has quietly absorbed scope, or a service line that was added to win a deal and never re-examined. Sometimes it is the product the founder is proudest of.

None of that is an argument for firing customers on the spot. It is an argument for knowing, so that the next renewal, the next price review, and the next decision about where to put your best people are made with the arithmetic in front of you rather than a feeling about who shouts loudest.