A buyer does not just count revenue, but asks about its source. A company where the largest part of revenue comes from one customer or a small number of customers is priced differently than a company with broad spread. Not because the revenue is worth less, but because the chance that revenue disappears is higher and the consequences of that are greater. Customer concentration is therefore a separate value driver, separate from profit and growth: it determines how certain that profit and growth are.
What a buyer looks at is not just the percentage of revenue with the largest customer. It also concerns the nature of the relationship: is there a contract with a term, or does it run on goodwill and habit. Is the customer replaceable by a comparable customer, or is the service so customized that the departure of that customer also renders the process behind it worthless. And: does the contact run through the owner personally, or through a team and a system that continues to exist if that person leaves. The latter directly touches on the degree to which the company revolves around the owner, because a large customer who only knows the owner is a double risk.
Customer concentration is measurable without special software. Put revenue per customer over recent years side by side and look at the share of the largest customer and of the largest five. Also look at the trend: is that concentration growing, because the large customer is growing faster than the rest, or is it decreasing because new customers are being added. A second signal is the contract form: framework agreements with notice periods weigh differently than one-off orders. A third signal is the margin per customer — a large customer at a low margin pushes value down harder than a large customer at a healthy margin, something that becomes visible as soon as recurring revenue is distinguished from one-off revenue.
Whether these figures are also reliable and up to date depends on how the administration is set up. Companies where customer revenue, margin, and contract status sit separately in spreadsheets often have no up-to-date answer to the question of how concentration stands this quarter. That is a matter of how the management information is built up and whether that information is available without manual work.
The shift is not in the risk itself — a buyer already counted customer concentration before AI existed — but in what is needed to reduce the risk and to demonstrate it. Three things run through each other here.
Compiling the customer concentration picture is work that AI can take over for the greater part, if the underlying data — invoices, contracts, CRM records — is structured enough to link. Where this still happens manually today in a spreadsheet that is updated once a quarter, the overview can then be available continuously and without extra hours.
Tracking signals that point to greater risk — a customer who orders less often, a contract that expires within a year, a contact person at the customer who changes — is partly transferable. AI can flag patterns, but assessing what a signal means and deciding on a response remains human work with oversight.
The third point is where the real leverage lies: if part of the service delivery to the large customer consists of work AI can take over — reporting, processing, first-line contact — that changes the nature of the dependence. The relationship then leans less on the specific people who do the work now, and more on a process that is also deployable with another customer. That does not make that customer's revenue less large, but it does make it less fragile, and that is exactly the distinction a buyer makes between revenue that weighs down and revenue that counts.
This shift does not run at the same pace across companies. A service provider with standardized reporting to a large customer is positioned differently than an agency where that same customer is used to a customized, informal way of working. The difference is not in the sector, but in how the work for that customer is set up — hence the question of which tasks are transferable in this specific company has no general answer.
Improvement begins with making the current concentration visible, per customer and per contract form, so it is clear which revenue carries the most risk. Next comes the question of which part of the work for that large customer is replaceable by a process that also works for other customers — that is a setup question, not a personnel question, and touches on how the team and succession within the company are organized. Where this touches on decisions about who does which work, separate statutory requirements apply that are not addressed here.
The underlying question — which work in this company can truly be taken over by AI, including the work tied to the largest customer — is answered per task in the work scan of FTE TO AI.
Start by adding up the revenue share of the largest customer and the largest five customers over the last two years, and alongside that figure record which part of the work for those customers is customized and which part follows a fixed process. Do you want to know whether customer concentration is currently the driver pressing hardest on your price, or whether another driver weighs more heavily — see also the overview of what depresses the value of a company — then the free value check, eight short questions, one per driver, gives a first picture of that. The full value scan, with evidence per driver and a two-year calendar toward the exit moment, is under construction.