Hi, is it possible calculate Dwell Time in Adobe Analytics?
I created a metric with time spent per visit divided by visits but it doesn't work.
Dwell Time is the amount of time that a searcher spends in search engine's result pages starting from the search result, clicks on a page, and returns back to the SERPs. (the page with the results).
Ahh.. so basically this would also constitute (in most cases), a single page visit on your site?
The problem with this is that Adobe isn't a heartbeat tracker... single page visits don't have any time associated to them, as there is no time stamp to calculate against....
I mean, I suppose IF you want to try and do something custom, I suppose you could add a listener for "Browser Back Button Used" and only IF the page's referrer is from a known Search Engine....
The you could get the raw data from Adobe, look at the page view time stamp and subsequent "back button" used, then calculate out the time that way... but this will require a lot of custom code (making sure that you don't track all uses of the back button on your site), and processing your Dwell Time outside of Adobe Analytics....
Just to clarify, if the user does this:
12:00pm: runs a search for keyword "foo".
12:03pm: clicks a result from the "foo" search results to open page B.
12:07pm: returns to the "foo" search results page by pressing the browser's Back button.
12:08pm: clicks another result from the "foo" search results to open page C.
12:10pm: returns to the "foo" search results page by pressing the browser's Back button.
12:12pm: runs another search, this time for the keyword "bar".
12:15pm: clicks a result from the 'bar" search results to open page X.
12:22pm: returns to the "bar" search results page by pressing the browser's Back button.
Then is your Dwell Time calculated as follows?
Dwell Time for "foo" = (12:07pm - 12:00pm = 7 minutes) + (12:10pm - 12:00pm = 10 minutes) / 2 (because he returned twice for the same keyword search) = (7 + 10) / 2 = 17 / 2 = 8.5 minutes
Dwell Time for "bar" = (12:22pm - 12:12pm = 10 minutes) / 1 (because he returned once for the same keyword search) = 10 / 1 = 10 minutes
I think the time to consider is from 12:00 to 12:07
(12:00 to 12:07) + (the time spent in the result clicked) / one session.
Hmm, so you just want the dwell time for the first search results click for "foo" only? You don't care about the other clicks for "foo", nor do you care about "bar" at all?
If the same user enter in a new page form the same result, it will be part of the analytics of the new page clicked becasue the main purpose of this metric is mesure the quality of SEO.
You can try this:
In a freeform table, include your eVar or prop that tracks your search term. Then add the Average Time on Site metric. Despite its "on Site" name, this metric really reports the average time based on the dimension in your table, i.e. your search terms. See https://experienceleague.adobe.com/docs/analytics/components/metrics/average-time-on-site.html?lang=...for more information about how it is calculated.
Question on this - i thought avg time on site doesn't include bounces/single page visits - so from what the user is trying to achieve, would this work if people are going straight back off the site?
You're correct @AbbySc .. that is why I suggested earlier that they would need to add a custom event when the user went back to Google... but then the freeform option that yuhuisg suggested won't work... since it won't calculate the time spent at all... or rather, it won't account for the time the user left the site... if the use goes:
Google > Site Page A > Google > Site Page B (and only if they came to the site again this second time)
The time spent metric would calculate the time of A as "Page B - Page A" and include all the time spent on Google...
This has to be completely custom, and likely needs to be taken outside of Adobe to calculate (like using Raw Data Feeds where you get the complete timestamp value)
I wonder if yuhuisg was thinking this was dwell time on an internal site search, as opposed to Google > Site Page > Google?