The studio
Brand architecture that survives the pitch.
Two founders, one method, and a framework we applied to ourselves before we sold it to anyone.
Founders
Real photographs only. Until they are taken, the frames stay empty.
Mohammed
Second founder
The method
We measure a site the way a surveyor measures a plot.
Eighteen parameters, values read off the wire: the hex, the faces actually rendered, the radii and how often each appears, the button geometry and its label contrast, the gap tokens, the transition durations, the count of images with alt text. Each value gets the mechanism from the manual, the source, and the grade the source earns.
We did this for 23 sites before we did it for our own. The portfolio taught us where the failures cluster: unused font families, primary buttons under 4.5:1, spacing off any base, one metric snapshot copied across every case study, reduced-motion never written.
Then we turned the framework on ourselves. The result is Case 00, and it is measured, not asserted.
How we work
Tokens are the contract.
Versioned, not drifted
The system changes by version.
A value change bumps the version and writes its reasoning and grade into the system document. A pull request that adds an accent somewhere new is a design change, not a tweak.
Bound per project
Every field resolves per item.
Before launch we verify that every dynamic field on a template resolves for each record and every themed token is overridden per project. The two documented failure modes of the reference system.
Payload discipline
A face that renders on no page ships on no page.
Fonts subset and self-hosted under 160 KB; Core Web Vitals budgets gate the build in CI; alt text is required at the CMS field.
Hiring
No open roles.
When a role opens it is published here first, with its scope written the way we write a parameter: the rule, the reason, and what it costs. Unsolicited notes are read. Write to hello@brentwelldigital.com.