On making decisions you can't reverse.
How to tell which design decisions are one-way doors and which aren't — and why teams that confuse the two ship slower, not faster.
Most design decisions are reversible. The colour of a button, the copy on a hero, the layout of a settings page — if any of these turn out to be wrong, you fix them in the next sprint and life moves on. The cost of being wrong is small: a redeploy, a Figma update, maybe a copy review.
A few decisions are not. Renaming a primary noun in your product (account → workspace → tenant) ripples through your URL structure, your support docs, your customer's muscle memory, and your help-centre SEO. Picking a primitive in your design system (is a Field a wrapper or is it the input?) compounds across thousands of components.
The teams we work with that ship slowest are the ones that treat both kinds of decisions the same way. They run a six-week design exploration on the colour of a button, then ship a fundamentally wrong primitive in a pull request on a Tuesday afternoon.
The test
Ask: if this turns out to be wrong six months from now, what does it cost to undo? If the answer is “an afternoon,” ship it. If the answer involves a customer-comms plan, you have a one-way door. Slow down.
What slowing down looks like
Slowing down isn't more meetings. It's a written proposal, a 48-hour comment window, and a named decider. Most teams skip the writing because they think it's overhead. The writing is the point — half the bad decisions die in the act of being articulated.
The other half survive the writing and come out better for it.