cookieyes.com on 56 sites: quiet pages, early starts

The light end of the index

The median site here contacted 8 companies over 32 requests, with 1 of them classified as tracking. Against the 88 companies in this index those rank 52nd, 67th and 84th. The median count of companies arriving inside the pre-consent window was 6, ranking 37th and level with 19 others.

Three of those four measures sit in the bottom half, one of them near the bottom outright; the pre-consent count is the exception, mid-table. Whatever else is true of these pages, they are not carrying an advertising auction. The typical shape is a corporate or institutional site with a handful of third parties on it and a banner in the corner.

54 of the 56 still had early contact

That is 96.4%, which ranks 19th of the 88 companies here — near the top, on pages that are near the bottom for weight. The two facts are not in tension: how many companies a page contacts and whether any of them arrive early are different measurements, and a page with six third parties can start just as soon as a page with sixty.

The count is recorded per site, not per company. This corpus knows how many companies arrived inside the window on each page; it does not know which ones, and it does not record whether a banner script's own requests fall inside that window. So there is no supportable claim here in either direction about the banner being outrun on its own sites — only that on 54 of these 56 pages, something was reached before the question could have been answered.

Nothing new arrives after the window

The median number of companies appearing after the consent window closed is 0.

On a typical site here the browser had met everyone it was going to meet before the point where a visitor could have replied. That is a different failure mode from a page that keeps recruiting for another ten seconds: the whole third-party surface of these sites is front-loaded, small, and finished early. It also means that on the median page, no company was first contacted after the window had closed.

The delivery host, and then the usual suspects

cdn-cookieyes.com was present on 55 of the 56 sites (98.2%) — the content delivery hostname that ships the banner script itself. After that: Google on 42 (75%), Cloudflare on 13 (23.2%), jsDelivr and Cloudflare Analytics on 10 each (17.9%), HubSpot, hs-banner.com and visualwebsiteoptimizer.com on 8 each (14.3%), Google Tag Manager on 7 (12.5%) and hsforms.com on 6 (10.7%).

A banner, a CDN, analytics, a tag manager, a CRM and an A/B testing tool. That is what a site with eight companies on it looks like from the network side: no advertising, no auction, just the tools a marketing team uses on its own pages.

All twelve named sites carried the banner

quantummetric.com is the heaviest of them at 25 companies with 20 inside the window, first contact at 897 ms. fyber.com: 11 and 11. usc.edu: 7 and 7 at 204 ms. anydesk.com: 6 and 6 at 341 ms. hugedomains.com: 6 and 6. kargo.com: 6 and 6 at 234 ms. ahrefs.com: 6 and 6, first contact at 182 ms. dribbble.com and smilewanted.com: 4 and 4 each. psychologytoday.com: 3 and 3 at 764 ms.

Two sites did not reach everything early. northwestern.edu contacted 5 companies with 3 inside the window, first contact at 1,994 ms, and markmonitor.com 4 with 3 at 1,189 ms. Both are among the slower starts in the sample. In this group, the pages where fewer companies landed inside the window are the pages that were slower to contact anyone at all.

This is not the same page as the CookieYes banner count

This index measures two different things about consent software and keeps them apart. This page counts sites where the browser contacted the cookieyes.com hostname. The consent-platform page for CookieYes counts sites where a CookieYes banner was detected on screen, which is a different population arrived at by a different method.

On these 56 sites the two overlap heavily — a CookieYes banner was detected on 54 of them — but the figures are not interchangeable and should not be quoted as one number. If you want the same measurement on a page you are responsible for, MurmStack runs it live in the browser and lists the companies contacted with the pre-consent window marked.

Seen on the same pages

Companies most often contacted by the same sites. Co-presence only — the crawl records what arrived, not who called whom.

• cdn-cookieyes.comTogether on 55 of these sites
• GoogleTogether on 42 of these sites
• CloudflareTogether on 13 of these sites
• jsDelivrTogether on 10 of these sites
• Cloudflare AnalyticsTogether on 10 of these sites
• HubSpotTogether on 8 of these sites
• hs-banner.comTogether on 8 of these sites
• visualwebsiteoptimizer.comTogether on 8 of these sites