Time spent per visit Vs. Average time on site Discrepencies

[Yohan_khan00]

We are trying to understand the difference we are seeing here just by transposing the table. 

The "Time spent per visit" is the same when we have the table transposed however the "Average time on site" changes. Why is that?

From adobe - the delta between the two metrics is:

"Time spent per visit (seconds) metric and Average time spent on site are similar, but have several key differences. Both metrics use ‘Total seconds spent’ as the numerator. However, ‘Average time on site’ uses the sequences that include a dimension item as its denominator. Time spent per visit uses visit count as its denominator.

As a result, these metrics yield similar results at a visit level, but are different at a hit level."

But it does not explain why transposing the table is changing the numbers for "Average time on site" from 1:40 -> 1:29

[anuk926]

Transposing a table does not change the underlying data or the calculation method of the metrics. Therefore, the change in the "Average time on site" metric from 1:40 to 1:29 when the table is transposed is likely due to a difference in the way the data is being aggregated and presented.

When a table is transposed, the rows become columns and vice versa. This can affect the way the data is being grouped and displayed, which can in turn affect the average time on site metric. For example, if the original table was displaying data for each page on the website, transposing the table would group data by visit instead. This could result in a different average time on site calculation, as visits may have different numbers of page views and different lengths of time spent on the site.

It's also possible that there is an error in the way the data is being presented in the transposed table, such as incorrect values or missing data, which could affect the accuracy of the average time on site metric.

In summary, while transposing a table should not change the underlying data or calculation method of metrics, it can affect the way the data is being grouped and presented, which may in turn affect the results of certain metrics.

[Yohan_khan00]

Thank you for your input - I can understand that changing the way the data is being visualized can account for the change. This begs the question "what is the correct way to visualize this?" which number is the correct one? If I do this with other metrics, I get the same results and the breakdowns are comprehensive in both cases so that question never comes up. Does adobe have a best practice for how to visualize time-related data? ex...metrics should ONLY be in Columns or some?

[cr_17]

Adding metrics and dimension in columns is the best practice. Even in the adobe training videos metrics and dimensions are added in as columns. Also, that is the best way to populate the data. Let's say for an example, breaking down the data by daily for a month, for that adding 30/31 rows visualized better than 30/31 columns right? So, adding metrics and dimensions in columns would be better.

[cr_17 · Accepted solution]

I did tried your example, and yes, the transposing the table is changing the values of average time on site. I think as anuk926 mentioned below i do agree that transposing the table is changing the way of data is being grouped or presented. But it shouldn't happen like that.

[svromero]

I agree that it shouldn't be doing this based on how the table is transposed. This has been giving me headaches for about a week now. My problem is somewhat related, but because of a Month dimension added to my Freeform table. If broken down by month, overall Avg. Time on Site bloats to 8 minutes, but by taking out the Month dimension from the table, the overall Avg. Time on Site becomes something like 2 minutes

[svromero]

Okay, turns out I misunderstood the way Avg. Time Spent on Site works. If there's a dimension attached, calculates it based on that dimension. I got confused because without a dimension (just using a segment for rows), instead of getting the overall Avg. Time Spent on Site for a whole visit for every person in the segment, it calculates it based on something else (occurrences? hits? pages? I'm not entirely sure).