The reason your search results differ from everyone else’s

Two colleagues at adjacent desks type the same six words into the same search engine within a minute of each other. The first three results match; below that the pages diverge, and one of them contains a local supplier the other never sees. Neither page is wrong. They were assembled for different readers.

What everyone sees

A search engine is imagined as an index with a ranking: type the query, receive the ordered list. Under that model any two people should see the same page, so a divergence looks like an error or an experiment. The list is instead computed per request, from signals that include the query and several things that are not the query.

What is actually happening

Personalisation combines location, language, device, session context and interaction history to estimate relevance. Hannak and colleagues measured the extent of personalisation in web search empirically and found consistent, measurable differences across users for identical queries. Pariser’s filter bubble argument added the second-order point: because history feeds relevance, the results tend to confirm the pattern that generated them.

Why it stays hidden

The tailoring hides because the page presents itself as an answer rather than as a recommendation. There is no visible marker distinguishing a universal result from a personalised one, no control comparison available in the interface, and no reason to suspect a difference unless two people compare screens deliberately. Absence of an alternative view is the most effective concealment available.