Total Time Spent on Site Metrics Use Cases

[SusiePo1]

Hi, our team gets a lot of questions regarding the time spent on site OOTB Adobe metrics and which one to use for which circumstance. Can someone help explain the differences between them? The Adobe article is not very helpful.

Which metric would I use in the following circumstances?

1. What is the total time spent on average at a sitewide level?
2. What is the total time spent on average on Page A?
3. What is the total time spent on average for segment #1 All Visits (OOTB adobe) vs. visits that make a purchase? (segment = VISIT container, Orders exists).

If we could understand which time on site metrics to use where, that would help us a lot in improving our website and conversion rate. Thanks-

[Jennifer_Dungan]

Yeah, I agree with the assessment that the documentation isn't all that helpful...

So before we dive into the metric, let's start by looking at how Adobe collects the time spent... because this is not a heartbeat system, Adobe really just calculates the time spent based on the timestamp of the hits:

Example:

• Page A - timestamp 10:03:45
• Page B - timestamp 10:05:05
• Page C - timestamp 10:06:56

The time spent on Page A is calculated by taking the timestamp of Page B and subtracting the timestamp of Page A... so 10:05:05 - 10:03:45 resulting 1 min 20 seconds

Page B is the same, taking the timestamp from Page C and subtracting Page B (10:06:56 - 10:05:05 = 1 min 51 seconds)

Page C.. well, that one is interesting, because without any more timestamps, Adobe has nothing to calculate with... so there is no time collected for Page C in this instance.

Now, onto the metrics, are you looking for "Total" time, or "Average" time?

There is a Total Second Spent metric, and a bunch of Average Time metrics all based off of the same calculated times.

All of these can work with different breakdowns, though some do work better than others (or at least used to.. it may have been fixed, and the issues only came up under certain circumstances that I can't recall at the moment)... without any filtering, you will see that the metrics come out to the same value, except for the Time Spent per Visitor (this one stitches together multiple visits for a visitor and it will look higher. (the last is just converted into second instead of showing as a timestamp):

(The column headers are showing the daily average)

When you break these down by pages, you will see that the values still match, but that the column headers are now an average based on the page)

Now, when you start using Segments, the column headers become "absolute total" so it's no longer averaging across your segments:

So if you have Segment A and Segment B and Segment C, and you need to average them, you will have to create a calculated average (or Mean) for the rows.....

But really, you can use either depending on how you want to use the data. If you are interested in Visitor averages, the you should use the Time Spent per Visitor (seconds), the Time Spent per Visit (seconds) is focused more on the Visit level...  the other two, Average Time on Site and Average Time Spent on Site (seconds) usually match.

Now, on top of this, if you don't like the data to be displayed in seconds, there is an easy fix for this. You can create a calculated metric and just change the display of the data:

I hope this helped

[kayawalton]

The easiest way I used to determine the difference between Average Time on Site and Time Spent Per Visit is that Average Time on Site is calculated based on the number of unbroken sequences linked to a specific dimension while Time Spent Per Visit is a function of the entire visit that includes that dimension (as long as it is not a single-page visit).

Therefore, what visualization and dimension you use to analyze "average time on site" matters, especially for question #3.

If you use a Key Summary Metric visualization, Average Time on Site and Time Spent Per Visit will mostly be the same unless you use minutes or hours for granularity. That is because visits will likely span two (or more) minutes or hours so a single visit can be broken into different sequences, increasing the denominator for the Average Time on Site.

If using Freeform Table to compare those segments, I prefer to use Time Spent Per Visit. (As @Jennifer_Dungan mentioned, you will need to create a calculated metric if you prefer to display that in hh:mm:dd format.). Any segment used as a dimension in a freeform table will be assessed based on the dimensions used in the segment, so sequences will play a big role in the denominator for Average Time on Site. If you look at the tables below, the Average Time on Site and Time Spent per Visit are significantly different when the segment is used as a dimension because Average Time on Site is now a function of the dimensions associated with the segment, which can create more sequences than non-bounce visits. In the second table, these two metrics are similar because it is not likely that many visits are broken up into two sequences because they span two days.