The first four weeks decide the other twenty

We keep a private document listing every engagement that has gone badly since 2015, with a one-line cause. It is not a long list, but it is a monotonous one. Nearly every entry resolves to the same thing: we started making before we had agreed what we were arguing.

The brief is a symptom report

A brief is usually an accurate description of a symptom and a confident guess at a cause. 'We need a rebrand' is very often a true observation that something is wrong and an untested theory about what. Roughly a third of the time the cause is in the brand. The rest of the time it is pricing, hiring, a product that does not do what the sales team says it does, or an internal disagreement that nobody has been willing to have.

You cannot design your way out of an argument that two directors are still having in private.

So the first two weeks are spent trying to disprove the brief. We interview the five or six people who actually decide things, separately and at length, and we ask each of them the same closing question: what would have to be true for this project to have been a waste of money? The answers rarely match. That mismatch is the project.

Write one sentence somebody could argue with

By the end of week four we hand over a single claim. Not a purpose statement, not a set of values — a claim, phrased so that a reasonable person in the same industry could disagree with it. 'Northmoor should publish its build record in full, quarterly, including the slippage' is a claim. 'Northmoor is a trusted partner in the energy transition' is a wish.

The test is simple: if the opposite of your sentence is obviously absurd, you have not said anything. Nobody is arguing for being an untrusted partner.

Agree what is out, in writing

The scope document we send at the end of week four spends as much space on what is excluded as on what is included. Every painful conversation we have had about money since 2019 has been traceable to an assumption that lived in someone's head and never made it onto that page.

• Name the decision-maker. One person, named, for the whole engagement.
• State the number of routes. Two, developed properly, beats five sketched.
• State the feedback window. Ours is four working days, consolidated.
• State what happens when the claim changes. It sometimes should — but it is a new project, priced as one.

None of this is exciting, and none of it is design. It is also the only reliable difference we have found between the engagements we are proud of and the ones we finished.