

THE TAKEAWAYA client who cannot describe a layout can always choose one. Building the variants up front turns an open-ended revision loop into a decision.
A design system and twelve-screen storefront for a diamond jewellery brand — built so the client could choose between variants instead of describing what they wanted.
Facetz sells certified diamond jewellery. The category has a visual default — black, gold, serif, heavy — and that default is exactly why one brand in it looks like every other. The positioning we landed on was deliberate opposition: quiet luxury, but in a cool blue identity instead of the black-and-gold cliché. The tagline it had to carry was "Diamonds, Boldly." My Role: I designed the brand application and the full front-end UI/UX. Not the back end — there is no commerce engine behind this, no payment integration, no inventory. What I built is the design: the system, the screens, and the interaction decisions. The design system came first, as a written source of truth rather than a mood board. Five brand colours with a fixed role each and a 60/30/10 split so nothing drifts; Deep Sky Blue as the only colour allowed on a primary button, so a second "brand blue" can never creep in; Midnight instead of pure black for every dark surface; a gold accent permitted only on certified-diamond and carat badges, explicitly so the design cannot slide back toward the look it was built to avoid. Sky Blue Mist is barred from body copy on light backgrounds because the contrast fails accessibility — written into the rules, not left to whoever opens the file next. Then twelve screens: home, category, product, cart, checkout, payment, order confirmation, account, login, signup, password recovery, and the 404. A complete purchase path, not a homepage with sample pages behind it. The part that was a decision rather than a deliverable: I built the variants. Six homepage layouts. Multiple product-zoom behaviours. Search presented several ways. Shop layouts with the filter left, right, across the top, and in a drawer. Collection list against collection grid, at two, three and four columns. Several product-card styles. Blog and FAQ page treatments. That looks like over-production until you have sat in the revision loop it replaces. A client asking for "the sidebar on the left, actually on the right, actually can we see it without one" is not being difficult — they are being asked to describe a layout in words, which is a thing almost nobody can do. Showing them the options converts an open-ended conversation into a decision they can make in a minute by pointing. The variants were the method, and the method was the point.
Walk the storefront — every screen and layout variant the brief covered
facetz-storefront.vercel.app
NEXT PROJECT ↘Casuall — Rebuilding a Catalogue Nobody Would Buy From