Instead of designing one perfect trading screen, I let users build their own
Vistapex · XyraLabs · Solo design + shipped front-end
For three weeks I tried to design one trading screen that worked for everyone. I failed every time. The breakthrough wasn't a better screen — it was admitting the perfect one doesn't exist, and handing the decision to the people actually using it.
Two users who want opposite things
A trading interface serves two people who could not be more different. One is opening their first position — every number is a small threat, and too much on screen makes them freeze and close the tab. The other has traded for years and wants density; show them a clean, airy layout and they get suspicious, because a trading screen with too little on it reads as a toy, and nobody trusts a toy with their money. Then stack a second problem on top: everyone is on a different screen — a 13-inch laptop, an ultrawide, a cramped window. There is no single size to design for.
I assumed it was a craft problem
I did what most designers do first — I assumed that if the layout felt wrong, I just hadn't found the right one yet. So I iterated: multiple design languages, pushed the spacing out, pulled it in, recolored everything, re-weighted the hierarchy twice. Each version was fine. Each was also missing something I couldn't name. And I was doing it alone, with no design peer to tell me whether I was circling the answer or the drain.
Clarity didn't soften it
Eventually I stopped trusting my eye and opened Microsoft Clarity. The dense, complete version I was proud of was burying first-time traders — they were leaving the trading screen at around 76%. So I simplified it. The simplified version was worse: too thin for experienced traders, and it showed so little it read as untrustworthy. Two versions, each one fixing the problem by breaking it for the other half. That's when it landed: there was no right design. I'd been trying to make a decision that was never mine to make.
Stop choosing for the user
I quit guessing the niche and let each trader tune the screen to their own eyes and seriousness — Overall Appearance, Buy/Sell Position, Color Mode, Accent, Contrast, Font Size, Corner Radius (None–XL), Spacing, Notifications. This wasn't a feature I bolted on; it was the answer to a persona problem I couldn't solve top-down. A beginner can push spacing and type up until the screen breathes; a pro can crush the density down and trust what they see. The real design work moved to fencing the freedom — sane defaults, a fixed radius scale instead of a free slider, spacing in steps — so the job became "design the range of good screens, and make every point in it defensible."
You can't mock your way out of multiplication
Every setting multiplies the possible screens — I couldn't hand a developer a Figma file per combination. So I built it myself: took the existing design system plus the solution and prototyped the front-end in Claude Code, which returned a production-ready repo. The AI accelerated execution; the architecture calls, the bounds on the controls, and the decision to do this at all were mine. I closed the design-to-code gap instead of drowning a dev team in screens.
Both personas built their own product
Around 68% of traders changed at least one setting — they didn't tolerate the controls, they reached for them. The split told the rest: beginners drifted toward bigger spacing and type; pros packed the density to the floor. First-time-trader drop-off fell from ~54% to ~31% — the exact failure Clarity showed at the start.
