Jinfeng (劲风), launched February 2020, targets abusive aggregation pages. Baidu defines an aggregation page precisely: a page with no main article content of its own, made of index links to other pages gathered around a particular theme.
The four offences
- Content outside the site’s field, or a site with no focused field at all — mostly scraped and stitched material.
- Content not matching the title or the tags marked on the page. A page labelled with a topic that its links do not actually cover.
- Static search result pages generated by the site’s own search function. This is the one most sites are guilty of without any intent — internal search URLs made crawlable, then made static.
- Empty, thin or expired aggregation pages carrying no effective information.
Why aggregation pages attract abuse
They are free to produce. A template plus a tag generates thousands of pages targeting thousands of phrases, each looking substantial in a link count and containing nothing written. Jinfeng exists because that trade — infinite pages for zero editorial effort — was too attractive to leave unaddressed.
What separates a legitimate aggregation page
Read the four offences backwards and the specification appears: it belongs to your field; its content matches its title and tags; it is not a raw internal search result; and it is not empty. A genuine category page, curated and populated, satisfies all four. A tag page auto-generated from a keyword nobody edited satisfies none.
The practical audit
- Count the aggregation pages the site actually exposes — tag pages, filter pages, internal search results, “related” hubs.
- Check how many carry fewer than a handful of genuinely relevant links.
- Check whether any are static internal search results.
- Block or remove the rest, and keep the ones a person would use.
Where this connects
Baidu’s quality framework already grades internal search pages with no relevant information as having no quality at all, and cross-field content damages domain focus under Hurricane. Jinfeng is where those two meet, and on most large sites the offending pages were generated by a plugin nobody remembers enabling.
How we apply this
Static internal search result pages are the most common version of this we find, and they are always accidental: a plugin that makes search URLs crawlable, plus a caching layer that makes them static. The audit is quick — count tag pages with fewer than five relevant links — and the remedy is deletion, which clients find harder to accept than any technical change.
Related services