Search Behaviour in Adult Content Has Nothing in Common with Anything Else

Search like everybody, but not

The conventional wisdom about search behaviour does not survive contact with this market. Everywhere else, users search, compare, evaluate, and commit. Here they search once, browse for eleven minutes, and leave having watched something they did not search for.

The search box is not a search box. It is a suggestion mechanism that happens to accept typing. The user knows this. They type an approximation, expect an approximation back, and then do the real work with their eyes on the results grid. The query is the opening move, not the answer.

Design for that and you get a site that works. Design for the standard e-commerce search metaphor and you get a site that behaves exactly like the one people abandoned in 2019.

What the data knows about you

The behavioural signal in this market is unusually rich precisely because nobody performs for it. There is no social layer, no wishlists, no reviews. The only thing a platform has is raw behavioural exhaust: what was searched, what was clicked, what was abandoned halfway, what was searched again immediately after.

That last one is the most valuable signal in the entire dataset. A search followed immediately by a slightly different search means the first result set failed in a specific, correctable way. Sites that learned to read that signal improved their relevance by multiples. Sites that did not are the ones nobody visits anymore.

The uncomfortable thing about being good at it

The adult porn tube site that understands this is the one where your third search works better than your first — because the site learned something from searches one and two and quietly changed what it showed you. Nothing about that is visible. It is felt, as a vague sense that this site gets it, and that vague sense is worth more than any amount of design polish.

Which is a strange thing to say about an industry where almost nobody reads anything.