How the data is made

Every listing starts from something a website publishes itself. Estimates are built on top of that evidence and always labelled as estimates.

Sources

ads.txt and app-ads.txt: the files sites publish to declare who may sell their ad space.
sellers.json: the files ad networks publish to list the sites and companies they pay.
Public page code: a site's home page and a few of its public pages, read as any browser would, without logging in or filling in forms.
Public link data: affiliate links on those pages, and where short links redirect to.
Public popularity data: the Chrome UX Report's country rankings, the Tranco list and Common Crawl's web graph, used to find sites and to estimate traffic.

AdSense reverse lookup

When a site counts as an AdSense site

A site is listed as AdSense when its own pages carry AdSense ad code with a publisher ID. It is listed as possible when its ads.txt file or Google's sellers.json point to an AdSense account but no ad code was seen. Placeholder IDs copied from tutorials are ignored.

Ads serving says whether the publisher ID appears in Google's public sellers.json. In our tests, sites whose ID was missing there almost never received an ad (1 in 27), so their revenue estimate is zero and we show what they could earn instead.

Niche

Each site's title, description and headings are classified into one of 28 niches with a free-text sub-niche, using a language model, then mapped onto about 1,300 market clusters. Samples are checked by hand.

Revenue estimate

Monthly revenue = visits × pages per visit × page RPM ÷ 1,000. Page RPM starts from a US figure for the niche, anchored on Google's public AdSense revenue calculator and 2025–26 publisher benchmarks, and is scaled by the site's country mix (US 1.00; UK, Canada, Australia 0.80; Western Europe and Japan 0.50; Brazil, Mexico, Turkey 0.20; India, the Philippines, Indonesia 0.11). We show a low, mid and high figure; a site's real figure is typically within x0.4 to x2.5 of the mid one.

Traffic estimates

Monthly visits come from a model that combines a site's Chrome UX Report rank in each country, its Tranco rank and its Common Crawl rank. The model is fitted on a sample of sites with independently measured traffic. On sites with reliable measurements, 56% of estimates fall within 2x of the measured figure; small sites are the hardest to estimate. Sites that appear in no popularity list get no traffic estimate.

Network Scout

We read the sellers.json files of 170+ ad networks and the ads.txt files and home pages of the sites they list. A site is verified on a network when its ads.txt has the network's own DIRECT line or a managerdomain record, or its page loads the network's tag. A site listed in sellers.json without either is shown as listed but not verified. Joins and leaves come from comparing each new snapshot of sellers.json and ads.txt with the last one.

Tool and download pages come from the site's home-page links and sitemap. Owner clusters join sites that share an analytics, tag-manager, AdSense or ads.txt owner ID; IDs shared by very many sites are treated as a template and ignored. Revenue = visits × 1.85 pages per visit × a typical premium-network RPM for the site type, shown as a range.

Affiliate Scout

We read a site's home page and up to five commercial pages its sitemap points to (reviews, comparisons, best-of lists), resolve its short links and cloaked links, and match every link against 100+ affiliate programs. A site is confirmed by one proven link (an Amazon link with a real tag, a network click link with a publisher ID, or a publisher-side affiliate script), likely from two strong signals such as an affiliate disclosure and sponsored links, and a candidate from commercial markers only. Links in comments never count. Sites that share an owner ID are grouped; media companies are labelled.

The revenue band = visits × commercial share × a 10% click-out rate × typical earnings per click for the program type × the country mix. It is uncalibrated and meant only to rank sites against each other.

How we fetch

See our crawler. Site owners can ask for their site to be removed at any time.