Category: I discussed this with Jakob Nielsen at CHI last year, and Lou Rosenfield mentioned it at this year’s IA Summit (slide 26):
Result 10 gets more clicks than position 9.
Lou’s data looks a bit wonky, but the pattern replicates across a slew of different search experiences.
This is interesting because it’s non-rational! If you believe in the ranking algorithms of web search, it should be the case that result 10 is less likely to answer your query than result 9.
I have two hypotheses to explain this finding:
- User’s rebound from the “next” pagination link, choosing the last link rather than opting in to wait for the next page to load and re-orienting to the page.
- The visual clarity created by additional spacing and layout changes following the last result, as well as the additional fixation time created by the scroll stopping, make the 10th item much more likely to be considered.
The infinite scroll that debuted briefly on web results during last summer’s Live.com beta (but was pulled before launch), normalized the click curve post result 7 to a normal descending function. Alas, this doesn’t help distinguish between the two alternatives!
Tags: 
Maybe this is a recall effect where the first and last items in a series are preferred over other options.
[...] I’ve written a lot about how pagination affects the user behavior in search, and done much more proprietary investigations thereof. The process of result evaluation is super optimized by the user, though less so than commonly understood given the 30% navigational query percentage. [...]
[...] written a lot about how pagination affects the user behavior in search, and done much more proprietary [...]