Gross Margin Is the Number That Tells the Truth About an LSP

When a buyer receives the financial statements of a language company, the first number they look at is EBITDA. Everybody knows this.

The second number is less famous. In my experience it decides more.

It is gross margin. What remains of each euro of revenue after paying the people who actually did the work: freelance translators, partner agencies, interpreters, sometimes in-house linguists.

EBITDA tells a buyer how much the company earns. Gross margin tells them how it earns it, and how long it can continue.

Most sellers have never calculated it properly. Most buyers will calculate it for them, in the first two weeks of due diligence.

Schedule a Discovery Call

1. The Number Buyers Calculate Themselves

Your profit and loss account rarely shows gross margin directly. Vendor costs are often mixed with software subscriptions and travel. In-house translators sit in personnel costs, together with sales and administration.

A buyer will separate all of this. They will ask for vendor invoices by month, a list of employees by function, and a project export from your management system.

Then they rebuild your gross margin with their own definitions.

If their number is lower than the one you had in mind, you start the negotiation in defence. Better to arrive with your own calculation, with the method written down, so the discussion is about the method. Your credibility stays out of it.

2. What Goes Above the Line

The definition seems simple. In an LSP it creates discussions.

Freelancers and partner agencies are clearly direct costs. In-house translators and revisers are direct costs too, even if they are on the payroll. Their time goes into the product, exactly like a freelancer’s.

Project managers are the grey zone. Some buyers put them above the line, because without PMs no project is delivered. Others keep them in overheads. Both approaches are defensible. What matters is that you choose one, apply it the same way for every year, and say clearly which one you chose.

Machine translation and AI costs are a newer question. Engine licences, API usage, post-editing tools. If they scale with volume, they belong with direct costs. A buyer comparing your 2023 margin with your 2026 margin will want to see them in the same place.

Consistency is worth more than the perfect definition. A gross margin that moves five points because you changed classification looks like a company with something to hide.

3. Same EBITDA, Different Companies

Take two LSPs. Both have €4 million in revenue. Both have an EBITDA of €600,000, so 15%.

Company A has a gross margin of 55%. After vendors and in-house linguists, it keeps €2.2 million. Its overheads are heavy, €1.6 million: a large sales team and a management layer the founder built over twenty years.

Company B has a gross margin of 35%. It keeps €1.4 million. Its overheads are lean, €800,000. Very efficient, almost nothing to cut.

On the EBITDA line they are identical. A buyer sees two different businesses.

Company A has pricing power. Its clients pay for something they value. The heavy overheads are a management choice, and a strategic buyer with its own back office can remove part of them after closing. The buyer sees synergies, and may pay for some of them.

Company B has already squeezed everything from its structure. Its profit depends on buying translation cheaply and selling it at a small premium. If vendor rates go up 5%, Company B loses €130,000, more than a fifth of its EBITDA. Company A loses €90,000 on the same increase.

The two companies have the same EBITDA. For most buyers, Company A is worth more.

4. Margin by Client: Where the Surprises Are

The average gross margin hides more than it shows. The real information is in the margin by client.

When we prepare a company for sale, we calculate contribution margin for each of the top twenty or thirty clients. Revenue from the client, minus the direct costs of the projects for that client. Almost always there are surprises.

A large, prestigious account that brings 18% of revenue at a margin of 22%. Three smaller clients with margins above 60% that nobody in the company considered important. One client that, after internal PM time, probably loses money. A client billed in dollars, whose margin depends mostly on the exchange rate.

The owner usually knows the large client is “a bit tight on price”. Few owners know the exact number.

A buyer will find it. And a large client with low margin is a double problem for them. Concentration risk, which I discussed in The One-Client Risk Buyers Worry About, and a margin that drags down the whole company.

Margin by project is useful too, especially by service line and by language pair. It shows where the company is really good, and where it simply accepts work to keep volume.

5. Vendor Cost as a Share of Revenue

There is a simple ratio that buyers follow closely: total vendor cost divided by revenue, year by year.

In itself, a high or low ratio says little. It depends on the service mix. An LSP doing mostly interpreting, or one that outsources whole languages to partner agencies, will have higher vendor costs than one with a strong in-house team.

The trend is what matters.

If vendor cost moves from 45% to 52% of revenue in three years, the buyer will want to know why. Rates paid to linguists going up while client prices stay flat. Or more work subcontracted to other agencies, who keep their own margin.

Each explanation has a different consequence for value. The first may continue. The second may be corrected by a new owner. A buyer needs to know which one it is.

The quality of your vendor management shows up here, in the numbers. I wrote about this side of the business in The Hidden Engine of Growth.

6. What AI Is Doing to Gross Margin

In the last two years, many LSPs have seen something strange. Gross margin percentage goes up, revenue goes down.

Machine translation and post-editing reduce the cost per word. Clients know this, and ask for lower prices. Often the company keeps a better percentage on a smaller amount of money.

A buyer will look at both numbers together. A margin that improves only because revenue shrinks faster than costs is a warning sign. A margin that improves while the company keeps its clients and its prices is the best signal you can show.

The difference is visible in the data, by client and by year. This is why the margin by client is so important to prepare. It lets you show which clients have accepted the new model and which ones are only negotiating down.

How buyers read AI adoption more generally is in How AI Adoption Is Changing LSP Valuation.

7. Preparing Your Numbers Before a Buyer Does

You do not need a consultant to start. You need your project data and some patience.

Export three years of projects from your management system, with revenue and direct costs per project. Most LSP platforms can do this. Group the projects by client, and then by service line.

Decide your definition of direct costs, write it down in one paragraph, and apply it to all three years.

Then look at the result honestly. Clients with low margin, a vendor cost trend going in the wrong direction. Some of these things can be corrected in the year or two before a sale. A price increase, or a difficult conversation with a large client.

What cannot be corrected can at least be explained, with numbers, before the buyer asks.

For buyers the advice is the same, in the other direction. Before you discuss a multiple of EBITDA, ask for the margin by client. If the seller cannot produce it, you have learned something important about how the company is managed.