Download websites that use Marfeel

Editorial teams work to deadlines, and working out which stories earned their audience and which merely filled a slot is not something that fits between them. Marfeel sells that problem as a hosted suite of analytics, testing and reporting tools built for publishers, and our index files the product under Advertising, with Marfeel as the owner, a Proprietary license and a first date of 2011.

A detection here works as a vertical filter. Technology of this kind is bought by sites that publish continuously and monetise attention rather than transactions, so the fingerprint separates media properties from the rest of the web without any content classification being involved. As of November 2026 we detect Marfeel on 34 sites, and for anyone assembling a publisher segment that is a cleaner starting point than guessing from URL shapes or article counts.

The counting caveat is specific to this category and deserves to be taken seriously. Publisher side scripts are frequently injected through a tag manager, proxied behind the publisher's own domain, or loaded only on article templates rather than on the home page. A crawler that fetches one page per domain sees whatever that page happens to load, and for a news site the home page often carries a different tag set from the article pages where the audience actually lands. Undercounting is the expected direction of error, and the sites we miss are not a random sample of the population.

Publishers change measurement and audience tooling far more readily than they change content management systems, because a reporting layer can be swapped without a migration project. Competing vendors read a list like this as a displacement map. Agencies and consultants read it to size the media segment in a given country before pitching. Researchers studying the economics of online news read it because tooling choice is one of the few observable signals of how a publisher runs its business, in a trade that discloses almost nothing about its own operations.

In each of those cases the useful object is the whole detected set with country and top-level domain columns attached, plus the difference between two refreshes. A per-domain lookup confirms a suspicion you already had about one site. A population tells you which national markets a vendor has actually won, whether its footprint is concentrated among a few large groups or spread across independent titles, and which properties dropped the fingerprint recently, since a publisher who has just changed analytics vendor is in the middle of exactly the decision most readers of this data care about.