Interlude
A Melbourne barber studio's homepage, built four different ways to find the right brand voice — still unfinished, on purpose.
Interlude is the marketing site for a Port Melbourne grooming studio, and it's a work in progress in the most literal sense — instead of picking a brand voice and building it, I built four fully independent homepages (v1 through v4, each with its own component tree) off a single version-picker, plus a /preview sandbox with seven more section variants under active review. There's no live URL and no public repo yet; this is mid-process, not a finished ship, and I think that's worth showing rather than hiding.
The brief
Interlude's own brief is explicit about sequencing: create brand immersion first, convert to a booking second. The audience is a style-conscious 25–45-year-old evaluating whether this studio matches their standards, not searching a directory — which meant the real open question wasn't "what goes in the hero" but "what does this brand even look like," and that felt worth answering with four real, working candidates instead of one moodboard.
Four bets, one brand
Each variant shares only tokens and fonts, not layout — v1 "Cinematic" is full-bleed hero video and dark editorial chrome, the most complete of the four and the one /services, /about, and /book actually build on. v2 "The Architecture" is light-background, condensed uppercase type, a numbered services list. v3 "Razor" is a stark single-word wordmark hero with a red accent instead of the brand's usual gold. v4 "Suspension" is the newest and most technical: a GSAP-scrubbed floating razor, a canvas-based ambient hair-strand field with five motion modes, and a pinned scrollytelling services section — built from a written spec and plan committed ahead of the code.
Restraint as a rule, in the places it's implemented
The parts that are finished are genuinely disciplined: a single prefersReducedMotion() check gates every GSAP and requestAnimationFrame call across the codebase, not just the obvious ones. But the project is honest about the gaps too — a full theme-toggle system exists but is wired up nowhere, the studio's address disagrees between two pages, and services are priced one way in three variants and another way in the fourth. None of that is dressed up; it's what "in progress" actually looks like mid-build, and I'd rather a case study said so than smoothed it over.
