Why the intuition is unreliable
Human judgement about client profitability tracks memorability rather than cost. A client who escalates loudly twice a year is memorable. A client who consumes forty minutes of somebody’s time every single day is not, because no individual instance is worth remembering.
The second distortion is that resold services feel profitable because they carry an obvious margin, and reactive support feels costly because it is visible. In practice a thin-margin licence resale attached to a heavy support load can be substantially loss-making without any single element looking wrong.
The method
- 1Take one quarter and one client. Start with a client you believe is profitable - the finding is more useful.
- 2Total the revenue: recurring service fees, resold services at your sell price, project fees and chargeable work invoiced.
- 3Subtract direct third-party cost: what you paid for every resold licence, circuit and hosted service for that client in the period.
- 4Add up support time: every hour spent on their incidents. If you do not record it, estimate per incident by category and be deliberately generous - an underestimate defeats the exercise.
- 5Add project delivery time for work in the period, including the overrun nobody billed for.
- 6Add account management time: reviews, meetings, calls, the informal advice. This is the figure most often omitted and it is rarely small.
- 7Cost the time at a fully loaded rate - salary, employment costs, tools, premises, and a share of overhead. Not the hourly rate you charge, and not bare salary either.
- 8Compare. Then do the same for two more clients before drawing any conclusions.
Do three clients before deciding anything. One client in isolation tells you about that client; three start telling you about your pricing.
What providers usually find
- Legacy clients on pricing set years ago and never revisited are the most common loss-makers, by a considerable margin.
- Small clients frequently cost more per pound of revenue than large ones, because account management does not scale down.
- Clients with old or unusual infrastructure consume support disproportionately, and the support load is rarely reflected in the price.
- A high-margin resale line can be entirely consumed by the support the resold service generates - which is only visible if incidents are logged against services.
- The client everybody dislikes is often profitable. The complaint is about behaviour, which is a different and legitimate problem.
What to do about a loss-making client
The reflex is to raise the price, and it is sometimes right. But a client losing money usually indicates a mismatch between what was sold and what is being consumed, and there are typically four options rather than one.
Reprice at renewal, with evidence. Change the scope so what they are consuming is what they are paying for. Fix the underlying cause - an unsupported line-of-business application generating a third of the incidents may be cheaper to replace than to keep supporting. Or, occasionally, part company, which is a legitimate decision that providers defer far too long.
What all four have in common is that they require evidence. A conversation about price that begins with an assertion goes differently from one that begins with a quarter of recorded data.
Making it routine rather than an annual project
The exercise above is a manual approximation of something that should be a by-product of doing the work. If time is booked against incidents at the desk, if services carry cost as well as sell price, and if project time is recorded against tasks, then client profitability is a report rather than an investigation.
That is worth saying without a sales pitch attached: the value is not in the reporting feature, it is in the recording discipline. A provider with disciplined time recording and a spreadsheet will get a better answer than one with an excellent platform and no time entries.