The Art of Saying No — Why Selective Focus Is a Growth Strategy
Every yes is a no to something else. Business owners who learn to say no well grow faster than those who don't.
The word "no" is one of the most powerful tools a business owner has — and one of the least used. In the early stages of a business, saying yes to most things makes sense. You need revenue, you need clients, you need to understand the market. But at some point, continued yesses become a drag on the business rather than a driver of it.
Nobody can tell you in the abstract where that point sits. So this article gives you the calculation instead: a way of pricing a yes before you give it, using numbers you already have.
Every yes is a purchase
A yes does not cost nothing now and possibly pay off later. It costs capacity immediately, and capacity in a small business is fixed in the short run. So a yes is a purchase: you are buying an outcome with hours you could have spent on something else. The only question worth asking is whether the thing you bought is worth more than the thing you gave up.
Framed that way, the problem becomes obvious. Most owners have never declined a purchase they never priced.
The number that makes the decision
You need one figure: contribution per delivered hour. Take a piece of work, subtract the direct cost of delivering it, and divide what is left by the hours it consumed. That is the only sensible currency for comparing one commitment against another, because hours are the thing you are actually short of.
Direct cost means the people doing the work, costed properly — salary, plus employer's National Insurance at 15% on everything above the £5,000 secondary threshold for 2026-27, plus any software or licences that exist because the work exists. Not premises, not your own drawings. Those do not change with one more job.
A worked example
Take an illustrative six-person creative agency turning over £420,000, and the retainer client every agency has.
The client pays £1,200 a month — £14,400 a year, which feels like real money. The retainer is written for eight hours a month. The agency time-records the account properly for three months and the true average is 21 hours a month: 252 hours a year. On top of that, the owner spends around three hours a month on the client's escalations, which is another 36 hours. Call it 288 hours of firm time.
The agency's blended delivery cost, loaded as above, is £32 an hour. So 288 hours cost £9,216, and the client contributes £14,400 less £9,216 — £5,184 a year, or £18 per hour of firm time.
Across the rest of its book, the same agency earns £62 of contribution per delivered hour.
What the yes actually cost
Put those 288 hours on work that performs like the rest of the book and they produce £17,856 of contribution instead of £5,184. The gap is £12,672 a year. That is not a lost opportunity in the vague sense. It is the price of the yes, payable annually, and it has been paid every year since the retainer started.
The payment behaviour makes it worse, but not in the way owners expect. The client's terms are 14 days and it pays around day 68. Statutory interest on a late commercial debt runs at 8% above the Bank of England base rate, which the Monetary Policy Committee held at 3.75% on 17 June 2026 — so 11.75%. On a £1,200 invoice 54 days late, that is £20.85, or roughly £250 across a year. The Late Payment of Commercial Debts (Interest) Act 1998 also allows a fixed £70 in recovery costs on each debt between £1,000 and £9,999.99, which would be £840 a year if the agency ever charged it, and it never does.
Note the scale, though: about £1,090 of late-payment entitlement against £12,672 of misallocated capacity. The cash is not the problem. The hours are.
The obvious conclusion is usually the wrong one
The instinct at this point is to fire the client. Before doing that, look at where the 288 hours came from.
The retainer was written for eight hours a month — 96 hours a year, plus perhaps a dozen owner hours. At 108 hours and the same £32 cost, the contribution would be £14,400 less £3,456: £10,944 a year, or £101 an hour. On the agency's own numbers, that would make this the best client in the book by some distance.
So the client was never the problem. The scope was, and the yesses that quietly widened it — a favour here, a quick extra there, none of them ever priced. This is the ordinary way a good client turns into a bad one: not one bad decision, but two years of small ones.
Three ways to say no that are not "no"
A flat refusal is rarely the best move available, and it is the one owners dread, which is exactly why they say yes instead. There are three softer instruments and between them they cover most cases.
Price it. The fee that makes this work worth doing at the agency's own benchmark is 288 hours at £62 of contribution plus £9,216 of cost — £27,072 a year, or £2,256 a month. Quote it. Either it is accepted, in which case the problem is solved, or it is declined, in which case the client has made the decision for you and nobody has fallen out.
Scope it. Go back to what was actually bought: eight hours a month, defined, with anything beyond it quoted separately before it is done. This is the version that usually works, and it costs one uncomfortable conversation rather than a client. Our guide on difficult client conversations covers how to open it without a row.
Refer it. Work that is wrong for you is often right for somebody else. Passing it on with a genuine recommendation costs nothing, keeps the relationship intact, and has a habit of coming back.
The decision rule
Before any significant yes — a client, a service line, a partnership, a speaking slot — do two things.
First, estimate the hours honestly, then add half again. Owners underestimate new commitments with remarkable consistency, and in the example above even that would have been optimistic: eight contracted hours a month turned into 21.
Second, divide the expected contribution by those hours and compare it with what the same hours earn on your best existing work. If the new thing is lower, the yes is a decision to earn less, and it needs a reason beyond not wanting to disappoint anybody.
There is one honest exception. A strategic yes — the first client in a sector you intend to enter, work that builds a capability you need — can be worth taking below your benchmark. The discipline is to say so out loud, write down what it is buying, set the date you will review it, and price it properly once it has served its purpose. An unreviewed strategic yes is just a cheap client with a story attached.
What to do this week
- Pick your three largest clients by revenue and time-record them properly for four weeks. Not estimates.
- Calculate contribution per delivered hour for each, using salary plus 15% employer's National Insurance above £5,000 as the cost base.
- Do the same across the whole book to get your benchmark figure.
- List every client below half the benchmark. For each one, decide now: price it, scope it, or refer it. "Monitor it" is not on the list.
- Put one of those conversations in the diary for this month, with a date rather than an intention.
If the exercise shows almost nothing clears your benchmark, the problem is pricing rather than selection, and pricing for profit is the better place to start. If it shows good work being crowded out by a handful of bad accounts, the same arithmetic runs through the leaky bucket.
Common questions
How do I say no to a client I have already said yes to for three years?
You are not withdrawing the yes, you are repricing or rescoping it, and that is a normal commercial conversation rather than a betrayal. Lead with the facts and skip the apology: the retainer was written for eight hours a month, the last three months averaged 21, and here are the two ways to put that right — the fee moves to match the work, or the work moves back to match the fee. Give them the choice and a date. Long-standing clients are usually less surprised than owners fear, because they can see how much they are asking for. The ones who react badly to being asked to pay for what they use are telling you something you needed to know.
What if I need the revenue and can't afford to lose anyone right now?
Then rescope rather than reprice, because rescoping keeps the full £14,400 and hands back the capacity. In the worked example, holding the account to its contracted eight hours a month takes it from £18 to £101 of contribution per hour without losing a penny of turnover. That is the move when cash is tight. What does not work is doing nothing on the grounds that you cannot afford the risk, because the £12,672 gap is being paid anyway — quietly, out of the capacity that would otherwise be selling and delivering better work. Cash pressure is an argument for acting sooner, not for waiting until you feel comfortable.
I don't have time records. Can I estimate the hours instead?
You can start there, but expect to be wrong by a wide margin and in a predictable direction. In the example the contracted figure was eight hours and the recorded figure was 21 — nobody estimates their way to that. Four weeks of rough recording is enough: name of client, task, minutes, nothing more elaborate. Include the invisible time, which is where the error usually lives — the calls, the chasing, the escalations that land on the owner, the rework nobody logs because it feels like a mistake rather than a job. If four weeks is genuinely impossible, record two, and treat the answer as a rough direction rather than a decision.
Isn't turning work away risky in a quiet market?
It would be if saying no meant idle capacity, but it rarely does — the capacity goes back onto work that already earns more per hour, or onto winning it. The genuine risk case is different: a business with real slack, no pipeline and fixed costs to cover. There, marginal work at low contribution is defensible for a defined period, because £18 an hour beats £0 an hour when the alternative is people sitting idle. Say that out loud, put a review date on it, and reprice when the pipeline recovers. The failure is not taking the work; it is still charging 2023 rates for it in 2027 because nobody went back.
How do I stop scope creeping back once I've reset it?
Make the boundary visible to the client rather than only to you. Three things do most of the work. State the included hours in every invoice or monthly summary, so usage is a shared fact rather than your private grievance. Quote extras before doing them, however small — the habit matters more than the amount. And give whoever handles the account permission to say "that sits outside the retainer, I'll get you a price" without checking with you first, because scope creep is mostly junior people being helpful under pressure. If your team has to escalate every small extra, they will simply absorb it instead, and you are back where you started within a quarter.
