Do Funnelback search results improve over time?

I’m having some discussions with my colleagues and I just want to clarify before I start making untrue assumption. For this new collection (not live yet), when we search for a term (e.g.“chemistry with science”) the top result is probably not what users will be looking for for that term, however the 2nd and 3rd are good matches. To my understanding, Funnelback will learn based on how users click through from search result to page so long as we’re using the clickTrackingUrl rather than the page URL alone. So in time the 2nd and 3rd result, if that is indeed what users are looking for, might bump up to 1st and 2nd.

So am I correct to assert this to my team? Does Funnelback learn over time and improve its own search results? I feel like I’m asking obvious questions about search engine fundamentals but just want to be sure :slight_smile:

Hi Bizt,

Yeps. You are correct. Funnelback uses the click data to influence the relevancy ranking over time.

You can find more information using the following link:

Hope this helps.

~Gioan

Thanks Gioan,

Actually, seems we are enabled by default - I can see the option is ON in the web UI and there is a location of a log file. So I guess all good for improving search results.

Yes, click tracking is enabled by default and will operate if you are using the default simple.ftl that ships with Funnelback.

However if you do your own implementation you must make sure you use the clickTrackingUrl when displaying links for your search results as this includes the callback to Funnelback to log the click.

The link that Gioan sent also includes some options that allow you to vary the period of click data that is used to influence the ranking.

Yes, we have our own implementation but will be using clickTrackingUrl for links on the live site. Thanks