# Twenty-brand visual benchmark

Requested September 13, 2026. This run applies the same six-dimension presentation rubric used for the Clean Health Labs refined R3 review to one selected product page from each of the operator's 20 brands.

## Selection and evidence

`targets.json` records the precise selected URL and selection reason. These are representative core products, signature lines or products labeled bestsellers by the brand. Selection does not establish that these pages have the highest conversion rate. Redirects and the delivered page state are recorded in the capture manifests.

Captures use 390 × 844 mobile and 1440 × 900 desktop viewports. The capture process scrolls for lazy content and records overlapping readable viewport images through the page, including its footer. Ordinary consent and promotion controls may be dismissed through their visible controls; those actions are logged. Security challenges are not bypassed. A blocked or materially incomplete page is unscored, rather than treated as a poor design.

Each independent judge must use image-viewing tools to inspect the actual rendered frames. Reading HTML, screenshot filenames or a capture manifest is not visual inspection. Each judge records the inspected images and cites frames for deductions. The two judges do not read one another's scores. Capture defects raised by a judge are investigated before they can become deductions against a brand.

## Scoring

Visual craft 25%; brand identity 20%; page flow 20%; trust presentation 15%; offer clarity 10%; mobile presentation 10%. See `SCORING-BRIEF.md` for the fixed anchors. Dimensions use one decimal; the weighted overall uses two decimals, rounded half up. The two overall scores remain separate. Any ordering is only a presentation comparison of these sampled pages under this rubric.

## Limits

These are subjective visual and presentation assessments. They do not measure conversion rates, revenue, speed, accessibility conformance, product efficacy, review authenticity or checkout reliability. Product category changes what buyers need: jewelry does not require a supplement's evidence structure, and a longer page does not automatically have a better journey. Motion, hover states, video playback and unexpanded content are outside still-image scoring unless separately captured and explicitly identified.

CHL's previous 8.26 and 8.63 scores are a historical reference for the same rubric, not a fresh head-to-head score or an assertion that CHL beats any established brand. A useful benchmark should expose both the strengths and the blind spots of the rubric itself.