The short version
- A wrong-fit project does not just earn a poor margin. It occupies the capacity a right-fit project needed and degrades the work you do for everyone else that quarter.
- Four to refuse: the client who has fired three agencies, work depending on someone not in the room, scope the client cannot describe, and anything priced against a cheaper firm's floor.
- The hardest refusal is interesting work at the wrong time, because every reason to say yes is real.
- The rule: if saying yes requires assuming nothing goes wrong on existing accounts, say no. Something always goes wrong.
Every profitable year we have had came partly from work we turned down. That is not a line about discipline. It is arithmetic, and it took me two bad years to see it.
A wrong-fit project does not just earn a poor margin. It occupies the delivery capacity that a right-fit project needed, it degrades the work you do for everyone else that quarter, and it usually ends in a way that costs you a referral you would otherwise have received.
What should you refuse?
| Signal | What it usually means | Cost of ignoring it |
|---|---|---|
| Has fired three agencies, no self-reflection | The constant is the client | You become the fourth |
| Key person not in the sales conversation | An invisible dependency | Delays you absorb and get blamed for |
| Cannot describe a good outcome in a sentence | No agreed goal | The engagement becomes a search |
| Price anchored to a cheaper quote | You are being asked to be them | Cheaper-firm work at your cost base |
| Great fit, wrong quarter | Capacity you do not have | Existing accounts get less |
The client who has fired three agencies. Sometimes the previous three were bad. Usually the constant is the client. Ask what went wrong each time. If the answer never includes anything they would do differently, you are the fourth, and the fourth is not the one who fixes it.
Work that depends on a person who is not in the room. A founder who will “definitely be involved” but is not in the sales conversation, an approver you have not met, a subject expert whose time is theoretical. Every dependency you cannot see is a delay you will absorb and be blamed for.
Scope the client cannot describe. If they cannot say what a good outcome looks like in a sentence, no proposal will fix that. The engagement becomes a search for the goal, billed at your rate, and it ends when they conclude you did not find it.
Anything priced by someone else’s floor. If the number is set by what a cheaper firm quoted, you will do cheaper-firm work at your cost base, and you will resent it by month two. The resentment shows up in the output before it shows up in a conversation.
The one that is hardest to refuse
Interesting work at the wrong time. A brand you would like in your portfolio, a problem you would enjoy, arriving in a quarter that is already full.
This one is hard because every reason to say yes is real. It is also the one that does the most damage, because taking it means the accounts already running get less attention, and those accounts are the ones renewing.
If saying yes requires assuming that nothing goes wrong on the existing accounts, say no. Something always goes wrong.
How do you refuse without burning it?
Say no early and give the real reason in one sentence, without dressing it as a capacity problem if it is not. Most prospects handle a direct no better than a slow yes, and a straight answer preserves the relationship better than a vague one.
Refer them to someone who is a good fit, if you know one. This costs nothing and has returned more than any marketing we have run.
Leave the door open on specifics: not now, and here is the condition under which it would be a yes. A surprising number come back later with the condition met.
What it buys you
The year we started refusing properly, revenue was roughly flat and margin improved by about a third. More usefully, the work got better, because the accounts that remained got the attention they were paying for.
Capacity is the only real asset a services business has. Spending it badly is the whole failure mode, and it is why we protect the weekly cadence as hard as we do.
Frequently asked questions
How do you know when to turn down client work?
Four reliable signals: the prospect has fired several agencies and describes nothing they would do differently, delivery depends on a person who is not in the sales conversation, they cannot describe a good outcome in one sentence, or the price is anchored to a cheaper firm's quote rather than to the work.
What does taking wrong-fit work actually cost?
More than the margin. It occupies the delivery capacity a right-fit project needed, degrades the work you do for existing accounts that quarter, and usually ends in a way that costs a referral you would otherwise have received.
Should you take an interesting project when you are already full?
Usually not. The rule we use is that if saying yes requires assuming nothing goes wrong on existing accounts, the answer is no, because something always goes wrong. Existing accounts are also the ones renewing.
How do you turn down work without damaging the relationship?
Say no early, give the real reason in one sentence rather than hiding behind capacity, refer them to someone who fits, and leave the door open on a specific condition. Most prospects handle a direct no better than a slow yes.
What is the warning sign in a prospect who has fired several agencies?
The absence of self-reflection. Sometimes the previous agencies were genuinely bad. If the account of what went wrong never includes anything they would change, the constant is the client and you will be the next one fired.
Does refusing work reduce revenue?
In our case revenue stayed roughly flat the year we started refusing properly and margin improved by about a third. More usefully the remaining accounts got the attention they were paying for, which is what drives renewals.
Sources
- Chua Network delivery data across 8 client accounts (internal fact bank)
