Simply put it is the aggregation of consecutive queries (as you have noticed).
Requirements:
1. The queries should occur within 60s.
2. There was no 'click' in the results list after the first query.
3. They should not be identical.
4. All query terms in the second query should occur in at least 1 document in your index.
/Henrik
Hey,
Thanks for replying. I don't think we are tracking clicks from search results. Will this be causing issues with the accuracy of did you mean?
If you don't track clicks then the 2:nd requirement is always true and might give you accuracy issues. However, if you ar using unified search and .Track() for tracking then click tracking is done automatically (a script is injected on your page in the footer: https://dl.episerver.net/12.7.0/epi-util/find.js that among other things listens to link clicks on your site and if you look at the links returned from your search call they are padded wit _t paramterers used for tracking the clicks).
/Henrik
Thank you very much. I will try and work out how to get that to work with CDN caching of those pages.
Hey,
So I know how to make EpiFind give me did you mean results but I can't find out how that works under the hood. It seems to be based on what people searched for after their initial search. Has anyone found docs that explain how it works?