The IAB TCF consent banner, measured on 846 sites
The label "IAB TCF consent tool" does not name a product. TCF stands for Transparency and Consent Framework, a technical standard published by the IAB: a shared way for a banner to record what a visitor chose and hand that record to the advertising companies a page works with. Many separate vendors build banners that speak it, so what a crawler detects here is a protocol running on the page, not one company's software. It was found on 846 of the 5,294 sites measured, 16% of them, and no other of the nine consent platforms grouped in this index sits on heavier pages — though two of them match it, Iubenda on companies contacted and Klaro on requests.
A standard rather than a vendor
Most of the other entries in this section are companies you could buy from. This one is a specification. A publisher picks a banner vendor, that vendor implements TCF, and the resulting page exposes the framework's JavaScript interface so that advertising partners can read a consent string instead of each asking separately. Detection therefore groups together banners from many different suppliers that happen to share a protocol.
By the number of sites it appears on, TCF is second of the nine consent platforms measured here, on 846 pages. That is a different kind of second place than a vendor's: it counts adoption of a standard across an industry, not the customer list of one business.
No consent platform here sits on heavier pages
The median site carrying a TCF banner contacted 8 companies over 45 network requests. Neither figure is exceeded anywhere among the nine consent platforms in this index, and neither is unique to it: Iubenda's sites also contact 8 companies at the median, and Klaro's also make 45 requests. The median count of companies classified as tracking on those pages was 2, second of the nine behind Klaro's 3 and level with Iubenda.
That ordering is not surprising once you know what TCF is for. The framework exists to pass consent signals to advertising vendors, so a page that implements it is usually a page with advertising vendors to pass signals to. The measurement records the association; it does not show the banner causing the load, and the same commercial posture plausibly produces both.
Where the median page's companies arrive
On the median TCF site, 6 companies had already been contacted by the time the banner could have been answered. The median number of companies contacted after that window closed is 0.
That second figure is the one worth sitting with. It is computed per site and then taken as a median, so it says that on a typical page in this group the third-party set was essentially complete before the window ended. Early contact of some kind was recorded on 805 of the 846 sites, 95.2%. This is a site-level measurement: it records that a company was contacted inside the window, not which company, and it does not place the banner's own requests inside or outside that window either.
The sites are largely ad-funded publishing
The sample of TCF sites in this index reads like a list of the ad-supported web: msn.com, forbes.com, weather.com, bbc.com and bbc.co.uk, nytimes.com, t-online.de, tumblr.com, twitch.tv, researchgate.net, springer.com and nature.com.
The spread inside that list is wide. weather.com contacted 13 companies, all 13 inside the early window, with a first contact at 393 ms. msn.com contacted 12, with 11 inside the window. At the other end, nytimes.com contacted 4 companies with 1 inside the window and a first contact at 1,997 ms, and the two science publishers, springer.com at 1,055 ms and nature.com at 1,244 ms, also start slowly. A shared consent framework does not imply a shared page design.
How quickly these pages start
The median time to first third-party contact across the 846 sites was 536 ms, which ranks eighth of nine on that measure — a fast start relative to the other consent platforms here. The fastest single first contact recorded on a TCF site was 64 ms.
536 ms is roughly the time it takes to read the first word of a headline. Nothing about that number is unusual for a commercial page — scripts in the document head resolve long before a person has decided anything — but it is the reason a consent window measured in the first second is a narrow one. If you want to see the same timeline for a single page you visit, that is what MurmStack records.
| Site | Companies | Before consent | First contact |
|---|---|---|---|
| msn.com | 12 | 11 | 417 ms |
| nytimes.com | 4 | 1 | 1997 ms |
| tumblr.com | 2 | 2 | 770 ms |
| forbes.com | 10 | 10 | 387 ms |
| t-online.de | 2 | 2 | 360 ms |
| bbc.com | 6 | 6 | 258 ms |
| twitch.tv | 5 | 3 | 721 ms |
| researchgate.net | 5 | 5 | 415 ms |
| bbc.co.uk | 6 | 6 | 146 ms |
| weather.com | 13 | 13 | 393 ms |
| springer.com | 4 | 4 | 1055 ms |
| nature.com | 4 | 4 | 1244 ms |
| Tool | Sites |
|---|---|
| IAB TCF consent tool | 846 |
Questions
What does "IAB TCF consent tool" mean when it appears in a report?
It means the page exposed the IAB Transparency and Consent Framework interface. TCF is a standard, not a product, and many different banner vendors implement it, so the label identifies the protocol rather than the supplier.
Do sites with a TCF banner contact fewer companies?
Not in this index. The median TCF site contacted 8 companies over 45 requests, and no other of the nine consent platforms measured has a higher median on either — Iubenda matches the 8, Klaro the 45. That is selection rather than an effect of the banner: a site adopts TCF because it works with advertising vendors in the first place.
How many TCF sites had a company contacted before the banner could be answered?
805 of 846, or 95.2%. Measured against the other consent platforms in this index that share ranks eighth of nine, so most of the other groups sit higher.
How much of the crawl does this cover?
846 sites, 16% of the 5,294 measured. By site count that makes TCF the second-largest consent platform grouping in the index.