A benchmark you will not apply to yourself is marketing. So we ran ours on this site first, published the weak parts, and fixed what we could in the same week. This is the changelog.
Where we started
On web design and usability (22 criteria, 440 points) we scored 304 and ranked 11th of 772. Overall we ranked 80th. Strong on speed and accessibility, thin on proof, and with a handful of real defects.
What we changed
- Fixed ten color-contrast failures. White text on bright colored labels now uses dark ink. The automated accessibility scan went to zero violations on every page we checked.
- Stopped the stat counters from showing wrong numbers mid-animation, and aligned “four products” with “eight apps”.
- Added a find-your-fit tool, an audience strip, a services-and-prices table, a five-step timeline, and industry pages with FAQs.
- Wrote down how we build: an Engineering page with dated measurements, an AI page, and technical notes.
- Replaced a screenshot that showed an unrelated site with a real one of our own preview form, and made the follower owl speak only when hovered so it never covers text.
Where we are now
Re-scored with the same rubric and judges, the site ranks 1st of 772 on web design and usability and 5th overall. Looks and feel and the technical measures are 1st. The weakest area is still proof and trust, which no amount of design fixes: it needs real client evidence, and we are collecting it.
How far to trust this
- The judges are AI models, and one of them helped build the site. We cross-checked with a second vendor, which agreed on the broad bands but not on exact ranks.
- We re-scored a local build, so the speed score reuses the production measurement.
- The ranking is one run at one moment. We plan to re-run it and share the change.