← Index

A timed benchmark — reference URL to deployed structure in 30 minutes.

A timed benchmark, not a deliverable. The question on the table: how fast does the practice go from a reference URL to a deployed structural reimplementation.

WebCapability Build
Duration · Single session · ~30 minutes

~30m

zero to deployed

reference URL → live alias

213 KB

production bundle

gzipped

2

commits, single session

no team, no rework

reference site matched

tier-1 European agency build

Hero panel of the deployed reimplementation — modernist residential photography with overlaid German-language CTA card in copper.
— Hero panel of the deployed reimplementation — modernist residential photography with overlaid German-language CTA card in copper.

— The Brief

A peer asked the question every prospective client eventually asks: "If I gave you a reference site I liked, how close could you get and how fast?" The honest answer is a stopwatch and a URL.

— Diagnosis

A tier-1 European residential developer site was selected — a tightly-art-directed brand artifact built by a flagship global agency. The brief: match the structure, the color system, the typography hierarchy, the section composition, with original copy and stock imagery. Single session, no team, full deploy.

— How I Read It

First-person · operator's note

A peer asked me a stopwatch question and I refused to answer in words. I picked the hardest reference URL I could find — a tier-1 European agency build — because if I beat the reference and the answer was honest, every subsequent timeline question got cheaper. The discipline was order: tokens first, structure second, copy substitution third, deploy last. I was not optimizing for closeness to the original; I was optimizing for non-fakeability. The benchmark holds because I didn't stop the timer, didn't go back to polish, didn't cherry-pick the screenshots. The clean alias is the receipt.

— Process

  1. 010–5 min

    Token extraction

    DOM inspection. Color tokens, type scale, spacing rhythm.

  2. 025–22 min

    Structural rebuild

    Vite + React + Tailwind. Component-by-component, top to footer.

  3. 0322–27 min

    Substitution

    Original German copy. Inter for the proprietary face. Unsplash imagery.

  4. 0427–30 min

    Deploy

    Hooks pass. Build clean at 213 KB. Live alias.

— The Build

01 · Token extraction

DOM inspection of the reference site to capture the exact color tokens, type scale, weight stack, spacing rhythm, and layout grid.

02 · Structural rebuild

Vite + React + Tailwind, component-by-component: header, panel-style hero, feature trio, two project grids, four alternating CTA bands, contact, footer.

Project grid · structural reimplementation
— Project grid · structural reimplementation

03 · Original substitution

Original German copy in identical content slots. Inter as substitute for the proprietary display face. Unsplash imagery in identical card frames.

04 · Deploy

Pre-commit hooks passed, build clean at 213KB, deployed to Vercel on a clean alias. Zero to live URL inside the session window.

Match a tier-1 agency in 30 minutes, and every other deadline question gets cheaper.

"Tier-1 European structure, deployed inside the meeting that asked the question."

Project grid · structural reimplementation
— Project grid · structural reimplementation

— Artifacts Delivered

  1. 01Deployed site at a clean Vercel alias
  2. 02Private GitHub repo, two commits
  3. 03Side-by-side comparison screenshots (desktop + mobile)
  4. 04Component inventory with reuse notes

— Outcome

The benchmark holds. Same playbook applied to a pitch deliverable would land inside one working day with substantial polish budget remaining.

— Receipts

  • live URL inside the session window
  • pre-commit hooks passed on first run
  • ~1dprojected for full polish on a pitch deliverable