Process
Published, because nothing here is improvised.
Most studios keep the method vague on purpose. Ours is scored, written down, and run twice on every project, once before we say yes and once before anything goes live.
The audit, named
The Turnaround Index: what gets checked before we say yes.
Our internal audit framework, run on every prospective project before any design work begins. It pairs automated technical checks with AI-assisted research synthesis across four scored dimensions, combined into one number: below a set threshold, we say so and decline the project rather than take a fee for work we don't believe will move the number that matters.
Load time, Core Web Vitals, and whatever the current theme or platform is costing you before a single visitor reads a word.
Claims, labelling and category-specific risk -- the same pass that once caught a live formulation error on a client's own product page before we'd touched the copy.
Attention ratio, funnel drop-off by step, and whether the page is arguing one thing or five at once.
What the category's own customers actually say, mined from real threads and reviews rather than assumed from a brief.
The scores above are illustrative of how the Index reads, not a result: every real project gets its own report, specific to it, delivered on the paid call. See it applied to a real client in the HealthX Labs case study, where the technical check alone found a fourteen-second load time and the category check was built on roughly ten to fifteen thousand real customer threads and reviews, one set per product. The trust & compliance dimension is written up in full, as a reusable framework rather than a one-brand report, in our taxonomy of Google Merchant Center disapprovals for Indian supplement brands.
The full method
Six stages, the Index run at both ends.
The Index
Every prospective project is scored before we say yes: technical, trust, conversion structure, category and customer, each 0–10. This is the paid call's actual deliverable, and it's why we sometimes decline work after it.
What you get · A scored report, in writing
Thesis & wireframe
A written argument for every section -- why it exists, what it has to prove, what happens if a visitor skips it -- then the wireframe that applies the argument. Nothing is drawn before it's argued.
What you get · A wireframe, with its reasoning attached
Walkthrough
You see exactly what will be built, section by section, and challenge any of it while challenging it is still free. Two revision rounds are included here, not squeezed in after the build starts.
What you get · A signed-off direction
Build
Fifteen working days against the clock we both agreed to, starting the day the wireframe is approved. One revision round is included here too, with a three-day response window on our side.
What you get · A live, working store or page
Launch verification
The Index runs again, this time against what actually shipped: the same technical and trust checks, plus every stated claim traced back to something true. Nothing goes live carrying a flag the first pass would have caught.
What you get · A clean second scorecard
The read
The first two to four weeks of real behaviour, read properly: what the analytics actually show, not what we hoped they'd show. This is the baseline every later Rebuild decision gets measured against.
What you get · A dated baseline report
The fifteen days are ours.
The approvals are yours.
The clock starts the day the wireframe is approved and pauses while a revision sits with you. Two revision rounds are included at wireframe, one at build, with a three working day response window on each. Said plainly here so it is never a surprise later.
The guarantee
In writing, with the exclusions written too.
If we cannot align on direction at the wireframe stage, that stage is refunded. Nothing is sunk yet.
The clock starts at wireframe approval and pauses for revisions. Miss it on us, and you are compensated.