Deciding which unprofitable customers to release, carefully

5 days ago
3 min read

Updated: 2 days ago

Introduction

Every business carries customers it loses money on. They are usually long-standing, frequently demanding, and often the ones who were there at the beginning. The analysis that identifies them is straightforward; what to do about it is not, and the advice to "fire your worst customers" is glib enough to be dangerous.

The reason it is dangerous is that unprofitability has several distinct causes, and only one of them justifies ending the relationship. A customer might be unprofitable because you priced badly, because your process suits them poorly, because they behave unreasonably, or because they absorb capacity that would otherwise sit idle. Each calls for a different response, and releasing all four is expensive.

1. Deciding which unprofitable customers to release starts with the cause

Not with the ranking.

Produce the list, then establish for each one why the number is what it is. A customer you underpriced three years ago and never reviewed is a pricing failure, not a bad customer, and the fix is a conversation rather than an exit.

2. Measure contribution, not full-cost profit

The distinction is critical.

A customer covering their direct costs and contributing something towards overhead is worth keeping if the alternative is idle capacity. Allocated overhead can make a genuinely useful customer look unprofitable, and releasing them leaves the overhead exactly where it was.

3. Separate the price problem from the behaviour problem

Two different remedies.

Low price is correctable at the next renewal or the next job. Excessive demands, late payment, abuse of your team, constant revisions and unreasonable expectations are behaviour, and behaviour rarely improves because you asked politely once.

4. Try repricing before releasing

Most of the value is here.

A properly explained increase, or a change in what is included, resolves a large share of these cases. Some customers accept it, some leave on their own terms, and both outcomes are better than a termination you had to deliver.

5. Change the scope rather than the price where that fits

Often more acceptable.

Fewer visits, longer response times, minimum order quantities, delivery on a fixed day, work only within your normal hours. The customer keeps the price they value and you remove the cost that made them unprofitable.

6. Count what leaves with them

The part that gets forgotten.

Referrals, reputation in a sector, a reference you rely on, volume that supports a supplier discount, or work from an associated business. A customer who is unprofitable in isolation is occasionally profitable through what they bring.

7. Make sure you can replace the capacity

Timing matters.

Releasing revenue you cannot replace turns a low-margin customer into no customer and the same fixed costs. Do this when your pipeline is strong, and release one at a time rather than conducting a purge.

8. Handle the exit properly

They will talk about it.

Reasonable notice, a clear and non-accusatory reason, help with transition, and no dispute over final invoices. In small markets the manner of the ending is remembered far longer than the reason for it.

9. Fix the process that let it happen

Otherwise you rebuild the same list.

Annual price reviews, defined scope in writing, minimum order values, payment terms enforced from the start. Every customer on the list arrived there through a gap in one of these, and the gaps are still open.

Check your contractual position before ending anything. Notice periods, minimum terms and any obligations to complete work in progress all apply, and they vary by agreement and jurisdiction.

Conclusion

Establish why each customer is unprofitable before deciding anything, because most cases are repriceable rather than terminal.

Measure contribution rather than fully allocated profit, distinguish a pricing failure from a behaviour problem, try a properly explained increase before considering an exit, offer a reduced scope where the customer values the price more, count the referrals and reputation that would leave with them, only release capacity you can replace and do it one at a time, handle the ending well because it will be discussed, and close the process gaps that produced the list in the first place.