Complex underneath. Obvious on the surface

Requirements arrive as a list. Products are not a list — they are a sequence of decisions somebody makes under time pressure, usually while doing something else. The first job is turning one into the other.

So the thing gets mapped before it gets drawn: what the product is for, what a person is actually trying to finish, and the order those steps happen in rather than the order the spreadsheet had them. Hierarchy falls out of that. What deserves the top of the screen is whatever the next decision needs.

Skipping this is what produces products that look considered and feel exhausting. A confused structure can be styled indefinitely and never improve, because the surface was never the problem.

People arrive at a screen with a question — where am I, did that work, what happens if I press this, how much longer. An interface earns its place by answering the one being asked at that moment and staying quiet about the rest.

That makes feedback part of the design rather than a later ticket. Loading, saving, empty, partial, expired, failed: people spend real time in those states, and leaving them to be handled during the build is how a confident product turns confusing at exactly the wrong moment.

Friction is worth spending deliberately. A few steps should be slow — deleting something permanent, spending money. Most should not, and the ones people repeat every day deserve the most attention and usually get the least.

A flow that reads well as a diagram can still be wrong to use. The cheapest place to find that out is something clickable, not something built.

So it goes in front of people early, with real content, real lengths and real edge cases, at the widths it will be used at. Placeholder copy hides every problem worth finding: names run longer than the box, lists are emptier than the mockup, and the interesting screens turn out to be the ones nobody designed.

What comes back changes the design while changing it is still cheap. That is the whole argument for it — after the build starts, the same finding costs a fortnight and an awkward conversation.

Products keep going. Features arrive, teams change, and the design has to survive people who were not in the room when it was made. That is a systems problem rather than a screens problem.

So what gets handed over is components with their states defined, spacing and type that hold at every width, and rules specific enough to actually follow. Brand character lives inside that rather than on top of it: a product can be distinctly yours and still behave the way people expect, and where those two genuinely conflict, usability wins.

It stays honest by staying close to the build — the same components, in the browser, at real widths. The gap between the design file and the shipped product is where most of the quality is usually lost.