Get ready! An upgraded Experience League Community experience is coming in January.
How to see the time spend on specific page in workspace interface?
Hi all,
I received several time the question about 'How many seconds is my page viewed before going to a different one?'
My end users always said it's easy to get it in Google Analytics but not in Adobe Analytics.
That is true that there is a lot a information about 'time spent' in Adobe Analytics interface : https://experienceleague.adobe.com/en/docs/analytics/components/metrics/time-spent.
And so it could be confusing for them.
And even more when you read this :
Even me, I start to get confused but I think I found a solution that I want you to challenge.
This is my result:
What's the logic behind:
- I created a segment only about Page on hit level as I'm only interested in the time between pages, I don't need the other events:
- I used the metric 'Average time on site' filtered by the segment described above. It will calculate the difference between hits in my segment.
- I selected the page dimension about the page I'm interested to know about.
- I added the Exits metric because no time is recorded for the last hit of a visit. As there is no info for those cases, good to know if I have some of them.
Do you think this make sense or do you see any issue or improvement needed?
Thanks for your help
Robin
Your segment in this case really isn't doing anything...
Aside from some minor calculation rounding, you can see that most of my pages will show the exact same time whether I apply the same segment as you or not:
Basically, you can use the Average Time as is, you don't really need to do the extra segment as you have in your report.
The definition that you pulled up specified the "Page" dimension because they wanted to be clear that you did not have to pair that dimension with "Page"... it can be paired (broken down, or a breakdown of) other dimensions like custom dimensions (such as Page Type, if you want to see time by the all pages that are the same type, or by Site Section, etc), or you can pair it with standard time based dimensions like Day, Week, Month, etc.
But I agree, the definitions of the time metrics are confusing and could use some improvements.
I do have to disagree with GA time spent though... the engaged time in GA4 is horribly undercounting on all our properties... while I prefer looking at time spent on a heartbeat system (something that actively checks if the user is still on the page, stops counting when the tab or window is inactive, etc), I have basically told people in my org to ignore GA... my top page has an average of 1 min 57 seconds in Adobe... and 7 seconds in GA... the same page in our roll-up property in GA says it has a 1 second average... I don't trust either number....
You are correct, single page visits and last page of the visit have no time (no time stamp to calculate with). So, I agree, adding Exits is a good idea, to monitor those "last pages" in the visit.
Agreed with Jen I also don’t see any major changes either, But I did notice a slight difference in the numbers though very marginal. See right is with the segment :--
This topic is extremely interesting, and responding to this post helped me revisit underlying fundamentals.
Reference document:
https://experienceleague.adobe.com/en/docs/analytics/components/metrics/time-spent
I recommend reading this document multiple times to fully understand the nuances involved.
Based on my review of the documentation, the original post, the comments, and my own experimentation, here are my key takeaways:
Understanding Adobe Analytics Time Spent Metrics and Segment Implications
The Time Spent and Average Time Spent metrics in Adobe Analytics are highly nuanced. Personally, I avoid using these metrics when analyzing a single piece of content. They are more meaningful when comparing two or more pieces of content.
When reporting on a single item, it often becomes difficult for stakeholders to understand the underlying nuances of how Time Spent is calculated. Additionally, outliers can significantly impact averages—for example, when a user leaves the application open in the background for 20 minutes and then returns within the same session (up to the 30-minute visit timeout). This behavior can artificially inflate Time Spent values.
Segment Experimentation and Key Observation
While experimenting with the Time Spent metric, I noticed an interesting behavior when applying the following segment filters:
HIT where Page View exists
HIT where Page exists
In essence, both filters restrict the dataset to Page View hits only. At first glance, one might assume this would not materially affect Time Spent calculations. However, due to Adobe’s internal calculation logic, the results can differ significantly—depending on how tracking calls are implemented.
How Adobe Calculates Time Spent: “Spread Forward” Logic
Adobe uses a “Spread Forward” approach when calculating Time Spent for a given dimension. This means that time between hits is attributed to the most recent instance of the dimension until a new value replaces it.
Example User Journey:
User lands on Page A
Spends 5 seconds with no interaction (no tracking calls fired)
User performs 3 clicks at 5-second intervals
These are action (link) calls
Action calls do not capture the Page dimension
User navigates to Page B from the third click
User bounces from Page B
Total session duration: 20 seconds
Adobe Metric Calculation
Total Seconds Spent: 20 seconds
The “Spread Forward” logic attributes time from subsequent action calls back to Page A
This logic applies to Page and Prop dimensions (similar in concept to persistence for eVars, but specific to Time Spent)
Average Time on Site (for a dimension):
Total Seconds Spent for the dimension ÷ Number of dimension sequencesKey notes:
The last hit (or bounce hit) is excluded from the sequence count
Page A persists across all intermediate hits until Page B is encountered
If Page A reappears without being interrupted by Page B, it is counted as the same sequence
Impact of Applying HIT-Level Segments
When applying segments such as:
“HIT where Page View exists” or
“HIT where Page exists”
…the Spread Forward logic no longer applies in the same way. As a result, subsequent action hits are excluded from the Time Spent calculation, which can significantly reduce the Total Seconds Spent metric.
This explains why Time Spent values can differ when a segment is applied versus when it is not—even though the dimension itself appears unchanged.
Recommendation
In my experience, Time Spent per Visit at the dimension level is often easier for stakeholders to understand and interpret, instead of the Time Spent on site considering the nuances of the sequence being used as the denominator.
Most importantly, we must be very cautious when applying HIT-level segment filters while reporting on Time Spent metrics, as they can fundamentally alter how time is attributed.
