Hi,
I'm trying to build a sequential segment that includes onsite search exits.
I was expecting to achieve the result with this segment:
VISIT container
hit 1: internal search exits
THEN within 1 pageview
exit page exists.
However, the data is far too high and if I build a fallout report, I can see a much greater share of visitors who search to then navigate to a range of site sections.
Do you have any tips?
Thank you.
Solved! Go to Solution.
Views
Replies
Total Likes
Maybe it's a problem with "within 1 pageview"? (since exits are hits, but not PVs)
Have you tried "within 1 hit"?
You also have to be careful what metric you use with the segments, since the segment is "visit" level
Here is what I tried (note that I track "page type" values, and I used that as my identifier):
And when I used it, I had to make sure I paired it with the "exit" metric, or the numbers were huge if I used "Page Views" or "Occurrences", but if I looked only at "Exits" in this scenario, this looks correct (based on a quick correlation):
Not knowing how you have "internal search" setup, I can't replicate an exact rule for your datasets, but hopefully seeing this will help you build something equivalent.
Good luck.
Views
Replies
Total Likes
Maybe it's a problem with "within 1 pageview"? (since exits are hits, but not PVs)
Have you tried "within 1 hit"?
You also have to be careful what metric you use with the segments, since the segment is "visit" level
Here is what I tried (note that I track "page type" values, and I used that as my identifier):
And when I used it, I had to make sure I paired it with the "exit" metric, or the numbers were huge if I used "Page Views" or "Occurrences", but if I looked only at "Exits" in this scenario, this looks correct (based on a quick correlation):
Not knowing how you have "internal search" setup, I can't replicate an exact rule for your datasets, but hopefully seeing this will help you build something equivalent.
Good luck.
Views
Replies
Total Likes
Thanks @Jennifer_Dungan really appreciate your response.
I've tried all and the data still looks off and I wonder if it's because we have multiple versions of the search results page and there are some elements shared with non-search pages so I'll need to dive deeper into it at some point.
For now, I've decided to use a work-around - instead of building a calculated metric I'm using the fallout report (with the 2nd step being any page other than the search results page). Fallout data is a lot more closely aligned with the flow report than any of the segments I've tried (they all return the same number of visits but must be due to our naming conventions).
Many thanks again.
Views
Replies
Total Likes