Every leadership dashboard I’ve seen in a services business carries revenue per client. Pipeline, win rate, utilisation, sometimes churn. All useful. All incomplete in the same way.
None of them answer the question that determines whether the business works: what does it cost us to serve this client, and is the fee above that number?
Revenue per client tells you which accounts are large. It says nothing about which ones are worth having, and those are not the same list. In most firms I’ve looked at, they overlap less than the leadership team expects.
Why revenue per client is a misleading metric
The logic that makes the top line attractive is also what makes it deceptive.
Two clients pay the same annual fee. One is straightforward: clear briefs, decisions made quickly, scope respected. The other generates a constant stream of small requests, changes direction twice a quarter, and requires a senior person in every meeting. On a revenue report they’re identical. On a margin report they’re not in the same business.
Leadership teams usually sense this. What they lack is the evidence to act on it, because the cost side of the equation, which in a services firm is almost entirely labour, isn’t measured at client level. Payroll is known in total. Hours are known in total. The allocation between accounts is an estimate, and estimates in this area are reliably wrong in a predictable direction: the difficult client’s true cost is understated, because nobody counts the ten minute interruptions.
The result is a firm that grows revenue while margin quietly erodes, and a leadership team debating headcount when the actual problem is pricing.
Measuring client profitability
Getting this number is less work than most executives assume. It requires two inputs: hours attached to a specific client and engagement, and a cost rate per person.
The operational change is at entry. Hours need a client, project and task as they’re recorded, not sorted afterwards, and billable work needs separating from internal work at the same moment. Software built for professional services does this by design, with billing rates applied by type of work, an approval step before anything reaches billing, and reporting that sets billable revenue against labour cost by client and engagement.
What you get out is a ranked list: margin by client, worst to best. It takes a quarter of data to be credible and it changes the conversation permanently. Board discussions about “growing the account” become discussions about whether that account should be repriced, rescoped or released.
There’s a second output that’s arguably worth more. Comparing estimated hours against actual, by engagement type, turns pricing from an act of memory into an act of analysis. Firms that do this stop repeating the proposal that lost money last time.
The uncomfortable part
Two findings tend to come out of this exercise, and executives should expect both.
The first is that a named, prestigious client is losing money. This is the harder one politically, because that logo is on the website and someone senior owns the relationship. The data doesn’t decide what to do about it. It does remove the option of not knowing.
The second is that a category of work the firm has been chasing is structurally unprofitable at market rates. That’s a strategy finding dressed as an accounting one, and it’s usually the more valuable of the two.
Capacity planning around staff availability
A related blind spot sits next to this one, and it’s worth fixing at the same time.
Cost to serve assumes the people are there. Delivery commitments made without knowing who’s on leave produce exactly the kind of last minute reshuffling that inflates the cost of an engagement without anyone recording why. A maintained shared calendar handles it; so does a leave planner such as actiPLANS, where requests are approved and availability visible before dates are promised to a client. The mechanism matters less than the fact that someone can see it.
Where to start
You don’t need a systems project. Pick your five largest accounts by revenue, track hours against them properly for one quarter, apply your cost rates, and rank the results.
If the ranking matches your revenue ranking, your pricing is sound and you’ve lost nothing. In my experience it rarely does, and the gap between the two lists is the most actionable thing a services CEO can put in front of a board.
