Hacker News new | past | comments | ask | show | jobs | submit login

I once read that 85% of Google searches are repeat searches, queries that Google has "seen" before. If this is false, please ignore the rest of this comment.

If 85% of searches are repeated queries, does this mean Google will treat a query differently if it is a repeated one?

If a query looks like a repeat query is it funnelled into a retrieving a set of predetermined results?

(To be clear, I am referring to queries that do not match exactly.)

No doubt this would be much faster than a "dynamic" search where any similarities to prior queries are ignored.

To keep things "fast" it might be necessary to treat the 85% differently from the 15%.

It might be beneficial for the search engine provider to encourage users to use repeat searches rather create than new ones.

As we know, Google is not transparent regarding the series of steps they use in providing search results from their web cache thus these questions are likely to remain unanswered.





Queries might be the same, but the result will differ. eg: localisation, personalised result etc.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: