Seven lines I keep coming back to.
Not rules. The shape of how I think when the work gets hard. Most of them are on the rest of the site in scattered form — this page collects them.
- 01
Slow clarity
Most decisions get better when you slow down enough to notice the underlying question.
Speed is a real virtue when the question is correctly named. It is a quiet liability when it is not. Slow clarity is the discipline of refusing to design until the question is honestly stated. It usually compresses the rest of the work.
AppliesWhenever a brief lands faster than the team can defend its assumptions.
NotWhen the situation is genuinely time-bound. Slow clarity is not slow execution — it is fast execution on the right thing.
- 02
It is almost never one big problem
It is three or four small ambiguities, each tolerable on its own, all of them quietly compounding.
The teams that come to me describing "a redesign problem" almost always have something earlier going on — quiet disagreement on what the product is for, drift in who it is for, a roadmap that has stopped saying anything specific. Going earlier is usually the move.
AppliesWhen the same kind of symptom keeps reappearing in different parts of the product.
NotWhen something specific and bounded is genuinely broken. Sometimes a fix is just a fix.
- 03
Most redesigns are reactions to a feeling
That is not wrong. But the work after is different from the work after a real fault.
The feeling is usually accurate; the diagnosis attached to it usually is not. Sitting with the feeling for a week — without designing — is one of the hardest and most useful interventions I know. The feeling usually thins, sharpens, or relocates. Whatever is left after the week is the real brief.
AppliesWhenever someone says "the product feels dated" without naming a specific failure.
NotWhen the broken thing is named, specific, and small enough to fix targeted.
- 04
The team agrees in the room and disagrees in the work
Quiet disagreement is the hard kind. It leaks out as design friction, slow decisions, and a roadmap nobody fully owns.
When the work tells a different story from the meetings, the underlying question has not been asked plainly. The move is naming the question for the team — and being willing to sit in the discomfort of the answer. Most teams skip this because everyone agrees on the words.
AppliesWhen design reviews keep relitigating the same decision, or when scope arguments recur in slightly different shapes.
NotWhen the disagreement is genuinely about tactics. Some arguments are healthy and want to stay.
- 05
Fog reduction, not feature addition
The product gets clearer when something is removed, not when something is added.
Most product backlogs are inverted. The interesting work — the work that compounds — is the work of removing, renaming, and re-scoping. New features rarely change the shape of the product. Cuts almost always do.
AppliesWhenever the backlog grows faster than the team can defend each item.
NotIn genuinely greenfield work where the surface area is still being established.
- 06
End when the answer is in
Engagements that go past their honest end date stop being useful — for both sides.
A listening week that produces a clean diagnosis sometimes makes the rest of the engagement unnecessary. The dignified move is to say so out loud, even when there is appetite to keep going. I price the diagnostic separately from the design work for exactly this reason.
AppliesWhen you have what you came for and the calendar still has time on it.
NotWhen the answer reveals deeper questions that genuinely want the time.
- 07
Mentorship is a discipline, not vibes
Real mentorship has scope, diagnostic, path, and a checkpoint. Encouragement alone is something else.
The most useful mentorship I have seen — and tried to give — is structured. A clear scope of what we are working on, an honest diagnostic of where the mentee is, a small ordered path of next moves, and a checkpoint to look again. Without that, mentorship drifts into reassurance.
AppliesIn any one-to-one relationship where someone is asking you to help them grow.
NotWhen what someone needs is not mentorship but a friend, a therapist, or a paid coach.
Most of these started as a sentence in a real engagement and outgrew it.
If you have a counter-example or a sharper version of any of them, write to me — I will swap them out.