Engineering · Aug 2025 · 6 min read
Performance Is a Design Decision
Speed is not a technical afterthought bolted on at the end. It is shaped in the first sketch — by layout, media and motion choices.
Ask a room of designers when performance happens and most will point to the end of the project — the optimisation pass, the compression step, the developer's problem. That mental model is exactly backwards. By the time a site reaches engineering, the majority of its performance budget has already been spent, invisibly, in design decisions.
A 4K hero video nobody chose to compress. Twelve typefaces loaded for one headline. A layout that shifts for two seconds because the ad slot did not reserve its space. None of these are code bugs. They are design decisions — and they were made in Figma, weeks before anyone wrote a line of code.
Where speed is actually decided
Typography: one well-chosen family with two weights will almost always look more intentional than five families competing for attention — and it ships a fraction of the bytes. Media: art-directed, correctly sized, modern formats, loaded only when they enter the viewport. Motion: transform and opacity only, everything else is a promise your users' devices cannot keep. Layout: stable dimensions reserved up front, so nothing jumps while anything loads.
This is why we put a performance review inside the design review, not after it. Every mock-up gets the same questions: what does this cost, when does it load, what does it move? If a section cannot answer, it does not ship. The discipline sounds restrictive until you notice what it produces: pages that feel instant, which is the most luxurious sensation a website can offer.
A slow beautiful site is a contradiction. The beauty ends the moment the user waits.
The payoff is measurable. Every 100ms of load time measurably dents conversion; a site that responds under 100ms feels physically good to touch. Search engines agree — speed is a ranking signal. But the real argument is emotional: performance is respect. It says the visitor's time and device matter more than our opportunity to show off.
So the next time a project plan slots “performance” into the final sprint, move it into the first workshop instead. Decide to be fast while the canvas is still blank — that is the only moment the decision is actually yours to make.