Try to figure out how Time Spent per Visit is calculated, but I found this
When I bucketed the metric, I found that it might even lower the lower boundary of the interval. For ex., for 5 to 10 mins bucket, the Time Spent per visit was 2.25 mins. this deosn't make any sense to me. Could anyone help me to understand this? Many thanks.
Solved! Go to Solution.
Views
Replies
Total Likes
The only explanation I could think of is this:
2.34 is the average time spent for this prop value, by 584,965 visits.
When you correlate the prop with "Time Spent per Visit - Bucketed" which is a dimension, you are asking Adobe which values of "Time Spent per Visit - Bucketed" and "Prop8's this value" were seen together, and what was the Average time spent per visit for these values seen together.
The Bucketed values are not "bucketing" the 2.34 of the prop's value, they simply letting you know what the Average Time Spent per Visit (Minutes) was for these Bucketed values and were also seen with the prop.
In other words, I could be in the "1 to 2 hours" value and saw this prop, my Average Time Spent per Visit (Minutes) was 5.72, and for this prop I spent 2.34 minutes average.
In other other words, the metric "Average Time Spent per Visit (Minutes)" for Bucketed values is not dependent on the prop's values.
This should be verifiable by looking at the "Average Time Spent per Visit (Minutes)" for these bucketed values without prop8 and they should be same as this screenshot.
Views
Replies
Total Likes
I think I have figured out. Let's assume Prop8 (Department) is "Dresses".
This report is telling you that People who spent 5 to 10 minutes on your site (and viewed dresses) spent around 2.25 minutes in the dresses department.
As an FYI - you might want to use the "Average Time On Site" metric - it's alot easier to read.
Views
Replies
Total Likes
Dear hs_water,
Time Spent per Visit is a calculated metric and it is already an average i.e. Total Seconds Spent / (Visits - Bounces). Now, if you breakdown the same by another dimension, you can creating an average for an average which is not ideal.
Try to take Total Seconds Spent, Visits and Bounces as your metrics, you should understand it easily.
Thank You, Pratheep Arun Raj B (Arun) | NextRow Digital | Terryn Winter Analytics
Views
Replies
Total Likes
Thanks, @David_Jerome and @PratheepArunRaj . Even though I remove the department dimension, the numbers still don't match up. Sorry I can't share the screenshot at that level. I also used Total Seconds Spent / (Visits - Bounces) to verify but got a different set of numbers. There could be some kind of conditions involved, like removing bots or outlier, or the system might use approximation to calculate. it's hard to understand.
After all, shouldn't it be that the buckets of a metric has same condition as the metrics does? It seems "Time Spent per Visit - Bucketed" and "Time Spent per Visit (minutes)" have different "Time Spent per Visit".
Views
Replies
Total Likes
The only explanation I could think of is this:
2.34 is the average time spent for this prop value, by 584,965 visits.
When you correlate the prop with "Time Spent per Visit - Bucketed" which is a dimension, you are asking Adobe which values of "Time Spent per Visit - Bucketed" and "Prop8's this value" were seen together, and what was the Average time spent per visit for these values seen together.
The Bucketed values are not "bucketing" the 2.34 of the prop's value, they simply letting you know what the Average Time Spent per Visit (Minutes) was for these Bucketed values and were also seen with the prop.
In other words, I could be in the "1 to 2 hours" value and saw this prop, my Average Time Spent per Visit (Minutes) was 5.72, and for this prop I spent 2.34 minutes average.
In other other words, the metric "Average Time Spent per Visit (Minutes)" for Bucketed values is not dependent on the prop's values.
This should be verifiable by looking at the "Average Time Spent per Visit (Minutes)" for these bucketed values without prop8 and they should be same as this screenshot.
Views
Replies
Total Likes
Views
Likes
Replies
Views
Likes
Replies