The customer you should let go.

You already know who they are. You knew before you finished reading the title. There is a customer whose name on your phone changes the temperature of your morning, who queries every invoice, who moves the goalposts and then asks why it took so long, and who has been doing this for years while you tell yourself the work is worth having.

The question is not whether they are annoying. Annoying is survivable. The question is whether they are profitable, and the honest answer for most owners is that they have never worked it out — because the cost of that customer does not appear on any report. It appears as evenings, as rework nobody logged, as a quote you rushed because you were dealing with them instead, and as good people quietly deciding they would rather work somewhere else.

Why the accounts never show you this

Your profit and loss shows revenue by month, not profit by customer. Unless someone has deliberately built customer-level costing, every customer in your business looks equally good on paper, because the only thing you can see is what they paid you.

That is how a customer can be your third biggest by turnover and your worst by contribution at the same time. Turnover flatters them. The cost sits in three places nobody counts: unbilled time, working capital, and the opportunity you did not take because your capacity was already spoken for.

What one customer actually costs — the arithmetic

Here is an illustrative case. The arithmetic is real; the business is not.

A service firm has a customer billing £18,000 a year. Direct cost of delivering their work is £12,600, so the gross margin is £5,400 — a 30% margin, against a house average of 42%. Already below par, but not obviously a problem.

Now count what nobody bills. This customer takes about six hours a month of owner and admin time that is not in the scope: the calls, the rework, the meeting to re-explain what was agreed, the chasing. That is 72 hours a year. At an internal cost of £45 an hour, that is £3,240.

• Revenue: £18,000
• Direct delivery cost: £12,600
• Unbilled time, 72 hours at £45: £3,240
• Real contribution: £2,160 — a 12% margin.

Compare that with an average customer of the same size: 42% margin is £7,560, and one hour a month of unbilled time costs £540, leaving £7,020. The gap between the two is £4,860 a year on identical turnover.

Then the cash. This customer pays on 74 days against your 30-day terms, so an extra 44 days of their annual billing is permanently tied up in your business: £18,000 × 44 ÷ 365 = £2,170 of working capital funding one relationship, every day of the year. Our guide on profit is not cash covers why that number hurts more in a growing business than a flat one.

And the 72 hours. That is nearly two working weeks of owner and senior time. Redeployed to the customers you actually want, or to winning one more of them, it is worth considerably more than the £2,160 the relationship generates.

How do you find yours?

You do not need a costing system. You need one afternoon and a willingness to write down uncomfortable numbers.

• List every customer with what they billed in the last twelve months. From the sales ledger, no estimates.
• Take off the direct cost of delivering their work. Materials, subcontractors, the hours of the people who did it.
• Estimate the unbilled hours honestly. Ask whoever deals with them most, not yourself. They will know, and their number will be higher than yours.
• Check average days to pay for each of the top twenty. Most accounting software reports this in two clicks.
• Rank by contribution, not by revenue. Then look at the bottom five.

Most owners are surprised twice. Once by who is at the bottom, and once by the fact that the customer they were sure was the problem is fine, while a quiet, polite, undemanding one is being served at a loss because the price was set in 2019 and never touched. Our owner time audit is the fastest way to get the unbilled-hours figure if nobody is tracking it.

Fix it before you end it

Letting a customer go should be the third option, not the first. Two things come before it, and both often work.

Reprice. A 12% margin customer at a 25% price increase becomes a perfectly decent customer. Owners flinch at that number because it sounds enormous, but on this example it is £4,500 — less than the unbilled time they already consume. Some will say no, and that answer is useful too. Our guide to pricing for profit works through how to put the number in front of someone without apologising for it.

Reset the scope. Often the customer is not expensive, the arrangement is. Written scope, defined response times, a named contact instead of everyone ringing the owner, changes quoted rather than absorbed. A great many "difficult customers" are simply operating inside a vacuum you created by never saying where the edges are. That conversation is easier than the exit conversation and it fixes the same money.

If you have genuinely tried both and the arithmetic has not moved, then it is a decision rather than a grievance — and decisions are easier to carry out calmly.

How do you actually end it?

Badly done, this costs you a reputation. Done properly, it costs you almost nothing.

• Give notice, not an ultimatum. Enough time for them to find someone else — a month, a quarter, whatever is fair for the work.
• Keep the reason short and about fit. "The way we work has changed and we are no longer the right firm for you." True, unarguable, and not an accusation.
• Do not itemise their sins. You are ending a commercial arrangement, not winning a case. Nothing good comes from the list.
• Finish the outstanding work properly. How you leave is what they will describe to other people.
• Offer two names. Firms that genuinely suit them better. It is generous, it is useful, and it turns the conversation from rejection into referral.
• Never do it in temper. If today was a bad day with them, have the conversation next week instead.

One caveat worth taking seriously. If the customer you want to lose is also large enough to matter — more than about a fifth of your turnover — do not start there. Replace the dependency first, over two or three quarters, then have the conversation from a position where their answer does not decide your year. Difficult client conversations goes further into how to hold that line without escalating it.

What this is really about

The money is the smallest part of it.

A business that will not let a customer go is a business that does not believe it can replace them, and that belief shapes everything else: what you charge, what you agree to, what you tolerate, how you speak about your own work. It is why discounting is a mindset problem before it is a pricing problem, and why the owners who fix this stop being the cheapest option in their market within about a year.

Your team notice too. Every hour spent placating someone who treats them badly is an hour that teaches them what this business accepts. Letting that customer go is one of the loudest things a leader can say without making a speech, which is why it turns up so often in leadership coaching as the moment something shifted.

And you get the capacity back. Seventy-two hours, two thousand pounds of working capital, and the mental space that comes from not bracing when a name appears on your phone. That capacity is what growth is made of — you cannot take on a better customer if you are full of a worse one. Our guide to improving margins is largely about spending that recovered capacity well rather than letting it fill straight back up.

What to do this week

• Rank your top twenty customers by contribution, not revenue. One afternoon.
• Ask whoever deals with them most how many unbilled hours the bottom five consume. Believe their number.
• Pick one. Decide whether it is a price problem, a scope problem or a fit problem.
• If it is price or scope, put the conversation in the diary this month.
• If it is fit, write the notice letter now and send it when you are calm.
• Before you release the capacity, decide what it is for. Otherwise it fills itself.