> If there is an opportunity here, it is for search to understand the difference and know how to present the two kinds of results differently.
Tag documents into categories (using ML diceroll, up to one category per document), with a strong bias towards "none" when in doubt. When a search term's results include some categories more frequently than the global average, offer them as filters.
Would its utility survive the resulting wave of SEO innovations?
Tag documents into categories (using ML diceroll, up to one category per document), with a strong bias towards "none" when in doubt. When a search term's results include some categories more frequently than the global average, offer them as filters.
Would its utility survive the resulting wave of SEO innovations?