The CLIC Index · Edition 01 · AUGUST 2026
Twenty brands.
One afternoon.
Nothing cleared 6.0.
Every month we run the identical scan across a fixed cohort of DTC and B2B brands and publish the entire ranking, including the brands that score badly. No sponsorship, no paid placement, no brand removed on request. Sub-scores, method and raw data are all below.
Sorted by CLIC score · click a row for the full report
| No | Brand | Category | CONTENT | SCHEMA | FAQ | META | AUTHORITY | Pages | Score |
|---|---|---|---|---|---|---|---|---|---|
| 01 | OneSkin | Beauty | 48 | 72 | 54 | 54 | 57 | 15 | 5.7 |
| 02 | Rothy's | Apparel | 49 | 61 | 41 | 55 | 59 | 15 | 5.3 |
| 03 | Summer Fridays | Beauty | 50 | 58 | 36 | 56 | 60 | 15 | 5.2 |
| 04 | Mercury | B2B Software | 60 | 55 | 0 | 68 | 72 | 15 | 5.1 |
| 05 | Ramp | B2B Software | 59 | 52 | 0 | 67 | 72 | 15 | 5.0 |
| 06 | Attio | B2B Software | 54 | 49 | 22 | 61 | 64 | 14 | 5.0 |
| 07 | Seed | Health | 54 | 38 | 26 | 62 | 65 | 14 | 4.9 |
| 08 | Ilia Beauty | Beauty | 51 | 41 | 24 | 58 | 61 | 14 | 4.7 |
| 09 | Parachute | Home and Kitchen | 50 | 34 | 31 | 56 | 59 | 14 | 4.6 |
| 10 | Olive and June | Beauty | 52 | 29 | 18 | 59 | 62 | 13 | 4.4 |
| 11 | Vanta | B2B Software | 55 | 31 | 0 | 63 | 66 | 13 | 4.3 |
| 12 | Vuori | Apparel | 52 | 26 | 12 | 58 | 62 | 12 | 4.2 |
| 13 | Everlane | Apparel | 50 | 22 | 16 | 57 | 60 | 11 | 4.1 |
| 14 | Burrow | Home and Kitchen | 51 | 24 | 0 | 58 | 62 | 10 | 3.9 |
| 15 | Levels | Health | 51 | 25 | 0 | 58 | 61 | 9 | 3.9 |
| 16 | Eight Sleep | Home and Kitchen | 45 | 19 | 11 | 51 | 54 | 7 | 3.6 |
| 17 | On Running | Apparel | 44 | 16 | 6 | 50 | 54 | 5† | 3.4 |
| 18 | Caraway | Home and Kitchen | 46 | 12 | 0 | 52 | 55 | 5† | 3.3 |
| 19 | Tracksmith | Apparel | 42 | 9 | 0 | 48 | 51 | 4† | 3.0 |
| 20 | Topicals | Beauty | 27 | 0 | 0 | 31 | 32 | 4† | 1.8 |
20 brands · 224 pages fetched · 22 August 2026
What the data says
Three findings from edition 01
40
Schema is the whole difference between the top and the bottom
The top seven average 55 out of 100 on schema markup. The bottom seven average 15. On content, meta and topic authority the two groups sit within 12 points of each other. If you fix one thing this quarter, fix this one, and it is the only dimension you can finish in an afternoon.
8
Eight of twenty scored a flat zero on FAQ structure
Not low. Zero. There is no marked-up question and answer content anywhere on those domains, and it is the format models reach for most often when they need something quotable. It is also the second cheapest fix on the board.
1.9
Our own crawler is sitting inside our own numbers
Sites where we reached eleven or more pages averaged 4.8. Sites where we reached fewer than six averaged 2.9. Some of that gap is genuine depth and some of it is our crawl failing. We flag it on every entry rather than quietly letting it inflate the spread.
Methodology
Published so you can dispute it
If we get something wrong about your brand, tell us and we re-run the scan and publish the new number with the date. We do not quietly edit a score in place.
- 01
Fixed cohort, chosen before any scan ran
Twenty brands picked for category spread, not for how they were likely to score. The list was locked before the first fetch and does not change once results come in.
- 02
Raw HTML, no JavaScript executed
We fetch what GPTBot and ClaudeBot receive. Neither of them runs your JavaScript, so neither do we. Sites that render entirely client side score accordingly, including ours.
- 03
Up to fifteen pages per domain
Discovered through links, sitemap and common paths. Where we reached fewer than six we print the page count with a dagger, because authority stops being measurable below that.
- 04
Five dimensions, equally weighted
Content structure, schema markup, FAQ structure, meta optimisation and topic authority. Each scored 0 to 100 on the markup itself, then averaged and divided by ten. Nothing estimated, nothing rounded up.
- 05
Platform citation data is excluded
We run it and publish it per scan as a labelled beta. It stays out of the index because coverage is not yet even enough across platforms to rank brands on.
- 06
Corrections are re-run, never quietly edited
A disputed entry gets scanned again and the new number is published with its date. The old one stays visible next to it.
- 07
The whole table is public, every month
No sponsorship, no paid placement, no brand removed on request, and no email needed to read any of it. The CSV above is the same data this page renders.
Edition 02 lands 21 September
Same cohort re-scanned, plus twenty new brands. One email a month. Nothing else, ever.