On estimates
Every client asks the same question in the first or second meeting. How long will this take? For a long time I gave them an answer. A real number. Six weeks. Ten weeks. Four. I'd land on a figure that felt confident, backed by the kind of mental math that passes for rigor in the moment.
I was wrong often enough that I stopped.
The problem with estimates in consulting isn't that they're hard. It's that the thing being estimated is still being discovered. The scope on day one is the client's version of the problem — the version they've carried into the room already partly solved. The real problem usually turns out to be shaped differently, which means the real work is sized differently too. An estimate given in the first meeting is a guess dressed as a commitment. The client, understandably, hears the commitment.
What I say now, almost verbatim: I can tell you the shape of the first two weeks confidently. After that, we'll know enough to price what's actually left — and if it turns out to be smaller than we thought, you'll pay for the smaller version.
Three things happen when I say that.
The client relaxes, slightly. They've been burned before by consultants who estimated eight weeks and billed sixteen, or who held to a fixed quote by cutting corners the client only noticed later. Both are worse than being told the truth up front.
The engagement starts honestly. The first two weeks is scoped tightly — I know what I owe at the end of it. The rest is deferred until we both know more. This almost always produces a smaller total engagement than my original gut-estimate would have, which seems bad for me and is actually good: the client leaves pleased, refers me, and comes back for the next thing.
And I stop having to defend numbers I never trusted. There's a small ongoing cost to carrying around an estimate you suspect is wrong. You end up either padding it defensively or racing to meet it. Neither produces good work.
The exception, worth naming: if the work is genuinely a repeat — a diagnostic I've done five times in the same shape, say — the estimate is reliable and I give it. Confidence should match evidence. What I don't do anymore is manufacture confidence to make the client comfortable in a room where I'm not comfortable.
Honesty about uncertainty turns out to be a positioning tool, not a weakness. It is the thing the serious clients, the ones I most want to work with, are listening for.
— K