Process
How philosophy becomes products
A belief is easy to state and expensive to keep. These four steps are how it actually gets kept — on the fortieth revision, on the deadline nobody moved, on the detail no one else would have checked. None of them are a formality; skipping any one of them is exactly how a considered product turns into a merely finished one.
Think
Every engagement starts before any design tool opens. We sit with the actual problem — not the feature list someone assumed would solve it — until we understand it well enough to recognize which of the obvious ideas are wrong. Most of them are. Saying no early is cheaper than saying no after something's built.
In practice: a week of watching how something is actually used today, or a plain conversation about what's failing and why, before a single wireframe exists.
Design
Design here isn't a visual pass at the end. Every decision — a layout, a color, a piece of copy — gets reviewed twice: once for whether it looks right today, and once for whether it will still make sense after the fifth revision, the new stakeholder, the feature nobody planned for. Decisions that only survive the first review don't ship.
In practice: fewer polished mockups shown, more real conversations about what happens when the content is longer than expected, or the user is on a five-year-old phone.
Build
What gets built is what ships — not a prototype that later gets "properly" rebuilt, which is where most of a project's real bugs and real technical debt quietly get introduced. Production-quality code from the first commit means there's no expensive second build hiding at the end of the timeline.
In practice: real data, real edge cases, and real performance budgets from the first working version — not added later as a cleanup pass that never quite happens.
Refine
The gap between "technically done" and "actually good" is almost entirely in the last ten percent — the loading state nobody specced, the error message that actually helps, the transition that makes an interface feel considered instead of assembled. This is the stage most projects cut when the deadline gets tight. We don't.
In practice: a dedicated pass, after everything already works, whose only job is finding the parts that merely function and making them feel intentional.