Reports
When is a client not worth keeping?
Small firms keep bad clients for the same reason they keep bad habits: the cost is spread thin and the exit feels dramatic. But the exit is not dramatic, it is one polite sentence at the right moment, and the cost of not saying it compounds every year the client stays.
QScope Team·18 August 2026·5 min read
One boundary first, because getting it wrong converts a commercial decision into a legal one: walking away means declining future work, not abandoning current work. Downing tools mid-contract without a lawful ground is breach, whatever the provocation, the safe ways to stop have their own strict machinery. Everything below is about the client whose next job you are free to refuse, and you are always free to refuse the next job.
Do the arithmetic before the feelings
A bad client’s cost hides in pieces. Add them for the last relationship year, honestly: the late payments and what financing them cost; the discount extracted at every signing and every closing; the free extras their expectations normalised; the retention still stranded; and the soft costs, the evenings on their arguments, the good client you could not serve that month. Divide the year’s margin from them by that total. For genuinely bad clients the answer is routinely near or below zero: you have been paying to keep them.
Hard is fine, harmful is not
Do not fire your demanding clients: demanding clients with money who pay on time are the trade’s backbone, and their standards make firms better. The line is elsewhere:
- Hard: negotiates firmly, checks everything, expects the programme kept, escalates real defects. Uncomfortable, profitable, keep.
- Harmful: pays late as policy, disputes as a discount strategy, moves scope verbally then denies it, treats your margin as an error to be corrected, or lies. The distinguishing mark is that paperwork does not fix them: a hard client respects a clean process; a harmful one treats every process as the next thing to game.
Leaving without a war
Finish the current obligations properly, defects period included, collect what is owed using the ordinary ladder, and then decline the next job the way the tender page declines: brief, bland, blameless, “we cannot give it the attention it deserves”. No exit interview, no scorecard, no honesty dividend: the harmful client hears criticism as an opening bid, and your goal is a closed account, not a won argument. Two cautions: time the exit so nothing of yours is hostage, ideally after the final retention lands, or with a clear-eyed decision that the stranded remainder is the price of leaving; and stay factual with third parties, the referral network that sent you this client will send you the next one, and bad-mouthing travels.
What to do this week
1. Run the arithmetic on your most draining client, one evening, from the records.
2. If they land in harmful, run the two-cycle test consciously from today: best practice, documented, and watch what happens, it either fixes the relationship or finishes the file for the decision.
3. Map what of yours they still hold, retention, unbilled work, live obligations, and sequence the exit so the leaving costs what you choose, not what they choose.
Where the information stops
A client who is your turnover’s backbone is a different problem wearing the same behaviour, and firing them without a replacement pipeline is a cash flow event this page cannot bless: that version of the decision belongs in a planning conversation with your accountant, with the forecasts on the table.