Your Largest Client and Your Best Client Are Rarely the Same Client
· 7 min read · Written by Faceela Research & Editorial Team
Reviewed by Ahmed Hassan Algammal — Founder and Enterprise Systems Consultant
Every firm can rank its clients by revenue in about four seconds. It is one query against the sales ledger, it has been correct since the day the accounting system was installed, and it is on a slide at every partners' meeting.
It is also close to useless on its own, because it is one half of a fraction. Revenue tells you what a client is worth. It says nothing about what he costs, and in a business whose only real input is people's attention, the cost side is where the entire answer lives.
What it actually costs to serve a client
The cost of serving a client is not the recorded chargeable hours at cost rate. That is the part everybody counts. Here is what usually sits outside it.
Grade mix. The same job done by a manager instead of a senior costs the firm more and earns it the same. When this happens once it is resourcing; when it happens every time for a particular client, it is because that client is difficult and the partner quietly assigns his better people.
Rework. The second version, produced because the first went out before somebody senior looked at it, or because the client's records arrived in a state nobody checked. It is real cost, it is rarely recorded as rework, and it clusters on particular clients rather than spreading evenly.
Meetings that were never recorded. The half hour before the meeting, the meeting, the follow-up call, the three internal conversations about what to do next. In most firms none of these have a timesheet entry, and they concentrate around the clients who like meetings.
The partner's own hours. This is the largest omission in almost every practice. Partners are the most expensive people in the firm and the least likely to record their time, and their attention is not evenly distributed — it goes to the clients who ask for it, who are frequently the clients paying the least for it.
Collection effort. The calls, the statements, the reminder that had to be drafted carefully because of the relationship, the partner who eventually had to telephone. A client who pays at ninety days costs more than a client who pays at thirty, and the difference is not only financing cost.
Disbursements never recharged. Government fees, courier, translation, attestation, a filing charge. Small, individually forgettable, and in some firms a meaningful annual number sitting in an expense account with no client attached to it. Whether they should have been recharged at all is a VAT question as well as a billing one.
Add those together and the picture rearranges itself. The client who pays full fee and takes three times the hours is not a good client. He is a client whose revenue is high and whose recovery is poor, and those are two different facts that a revenue ranking cannot separate.
The four positions, and what each one implies
Put revenue on one dimension and recovery on the other and every client sits in one of four positions. This is not a sophisticated model. It is useful precisely because it is not — it turns a conversation about a difficult client into a question with an answer.
The shape of it
Four kinds of client, and only one of them is the one you argue about
Two dimensions. Across: fee revenue — what you invoice them in a year, the number every firm already has. Down: recovery — how much of the standard value of the time you put in came back as cash, the number most firms do not have per client.
High revenue, high recovery
The client you think you have
The scope is clear, the contact answers, the work goes out right the first time and the invoice is paid without a conversation. There are usually fewer of these than the partners believe.
What it implies
Find out what makes it work — it is almost never the client alone — and then ask which of your other clients could be run the same way.
High revenue, low recovery
The one that pays for the office and eats the firm
Full fee, three times the hours. Every partner has one and almost every firm renews it at the old number, because the revenue line is the number in the room and the recovery line is not in the room at all.
What it implies
Reprice, re-scope, or change the grade mix serving it — in that order. Do not renew unchanged for a third year while calling it strategic.
Low revenue, high recovery
Small, clean, quietly profitable
Rarely discussed, because small clients are discussed only when something goes wrong and nothing goes wrong here.
What it implies
Usually under-served rather than marginal. The question is whether it can carry more work, not whether it is worth keeping.
Low revenue, low recovery
The one everybody already knows about
Small, awkward, absorbing a partner's attention out of all proportion to the fee. This is the quadrant firms think client profitability is about, and it is the cheapest of the four to fix.
What it implies
Reprice once, in writing. If nothing changes, stop — nobody will fight you over this one, which is exactly why it is not where the money is.
The quadrant that matters is high revenue and low recovery, and the reason it matters is political rather than analytical. That client is visible. He is on the revenue slide. He is somebody's relationship, often a founding partner's. He was probably the reason the firm hired two people three years ago. Every fact about him that is easy to say out loud argues for keeping him exactly as he is.
The one fact that argues otherwise is not in the room, because nobody computed it.
And the action is almost never to fire him. It is to reprice, re-scope, or change who serves him — in that order, and the third one is the cheapest and the most often skipped. A client whose work is being done a grade above where it was priced does not need a difficult conversation. He needs a resourcing decision.
Why a firm cannot fire a client it cannot measure
There is a particular kind of paralysis that comes from having only half the fraction.
A partner suspects a client is unprofitable. He cannot prove it. The revenue number says the client is important. Any proposal to reprice is therefore a proposal to risk a visible number on the strength of a feeling, in front of colleagues who will remember it if the client leaves.
So nothing happens, and the fee is renewed at the old figure. Then it is renewed again. This is not weakness — it is the correct response to an argument with no evidence in it, and it will continue until somebody produces the other half of the fraction.
Which is why this is a systems question and not a management-courage question. The conversation the partners cannot have is one that costs nothing once the number exists.
| What most firms have | What actually changes a decision |
|---|---|
| Revenue by client, by year | Recovery by client, against standard value |
| Chargeable hours at cost | All hours, including the partner's and the unrecorded meetings |
| Write-offs as one total | Write-offs by reason, by client |
| Disbursements in an expense account | Disbursements attached to the client who caused them |
| An average collection period for the firm | Days to pay, per client, with the effort it took |
| A feeling about who is difficult | A position on two dimensions, per client, updated monthly |
The honest limit
Everything above depends on recorded time, and recorded time in most firms is reconstructed.
That has to be said plainly rather than buried, because the failure mode here is worse than having no report at all. A client profitability model built on Friday timesheets produces a precise-looking figure with a specific, systematic bias: the clients whose work is fragmented into short tasks look cheaper to serve than they are, because short tasks are the first thing recall loses. The client who telephones four times a day is the one most likely to be under-costed by a reconstructed week.
So a firm in that position is not measuring profitability. It is producing a well-formatted opinion with a bias that runs in exactly the wrong direction, and acting on it confidently is worse than acting on instinct, because instinct at least knows it is instinct.
The order of work is therefore fixed and it is not the order most firms want. Fix the timesheet first — which means making entry cheaper rather than enforcing it, and recording non-chargeable and partner time as a matter of routine. Then compute recovery. Then rank clients. A profitability report is a conclusion drawn from a record, and there is no version of this where the conclusion is more reliable than the record.
What to make a vendor show you
On a live system, on real clients, not on slides.
- Show revenue and recovery for one client on one screen, for a period you choose.
- Show all time on that client including partner time and internal meetings, not only the chargeable entries.
- Show disbursements attached to the client that incurred them, with which were recharged and which were not.
- Show write-offs for that client by reason code.
- Show days to pay for that client against the firm's average.
- Rank every client on both dimensions at once, and let a partner sort it himself.
- Show the same client's figures for last year beside the fee being proposed for next year, on the screen where the fee is set.
Item 2 is the one that decides it. Every system will report chargeable hours by client, because that is what the invoice is built from. A system that can show you the unbilled attention a client consumes — the partner's hours, the internal meetings, the collection effort — is the only one answering the question this article is about, and it is the capability most likely to be quietly missing from a demo that otherwise looks complete.
Item 7 is the one that makes any of it matter. A profitability figure produced in March, about last year, that is not on the screen when this year's fee is agreed, is a report. On that screen, it is a price.
The short version
Revenue per client is the number every firm has and the one that decides least. Cost of service is the number that decides everything, and it includes several things nobody is recording — grade mix, rework, meetings, the partner's own hours, collection effort and disbursements that were never recharged.
Plot the two together and the client who worries everybody turns out to be small, and the client nobody worries about turns out to be the one absorbing the firm. The action is usually repricing or resourcing rather than resignation, and it is a conversation that becomes easy the moment the second number exists.
But the second number is computed from timesheets, and a reconstructed timesheet is biased in a specific direction — against exactly the clients whose work comes in short pieces. So fix the record before you trust the ranking. A confident decision on bad data is worse than an honest instinct, because it stops anybody asking.
The value that never gets billed at all is unbilled work in progress; how it leaks between standard value and cash is the realisation bridge; and where the extra work came from is usually scope.
If you want an arithmetic starting point before any of this exists, the cost of chaos calculator works from your own figures rather than from a benchmark. What it looks like built once is the professional and business services page and ServX.