Download websites that use Baidu Analytics (百度统计)

An analytics tag is the clearest statement a site makes about who it thinks its visitors are, which is why this one carries more meaning than most measurement scripts.

Baidu Analytics (百度统计) is the web analytics platform operated by Baidu Inc., offered since 2007 on a Proprietary basis and free to use, installed as a JavaScript snippet in the page. Sites add it for a plain reason. The operators want to see traffic the way Baidu Inc. sees it, because the search engine behind the analytics is where a meaningful part of their audience begins looking. As of October 2026 we detect the tag on 9,044 sites.

The population is not simply Chinese websites, and reading it that way loses most of its commercial value. A large part of the set is the opposite: international brands, hotels, universities, manufacturers, luxury retailers and travel operators running a China facing version of their site and measuring it with the analytics that match the search engine they are optimising for. Presence of the tag is a statement about intended audience rather than about hosting, ownership or language. A company that installed it has already decided that a specific market is worth building and measuring for.

For a commercial researcher that makes it one of the cleaner intent markers available in a technology dataset. Cross border ecommerce services, translation and localisation vendors, payment providers, content delivery and hosting firms with mainland presence, and marketing agencies working that corridor all want the same list: organisations that have committed budget to the decision rather than organisations that merely translated a landing page. The tag qualifies them, and the rest of the detected stack on the same domain tells you how far along they are.

Detection is snippet based and deliberately conservative. We record the tag where the tracking script appears in a page we crawl, which misses sites loading their analytics through a tag manager we cannot resolve, sites proxying the script through their own domain, and China facing subdomains or separate country domains that never link back to the page we reached. Reading the count as a floor is the safe approach, and the absence of a version breakdown for a hosted script means install age is not a question the data can answer here.

None of the questions worth asking can be handled one domain at a time in any case. Which companies in an industry run a China facing measurement stack, how that set overlaps with a Western analytics vendor on the same page, which country code domains carry it, and who added or removed the tag between two of our snapshots are all population questions by nature. Our export is that population, in a single file, with the rest of each site's detected stack next to it.