Bounce rate varies for hit segments vs dimension values

[_bel_]

My org has had a few instances of URL paths and/or page names changing globally, so I rely a lot on hit-level segments to replace dimension tables. Something I've noticed is that entries, and therefore bounce rate, does not act as expected in this scenario. When I use a segment that is hit-based with simple (URL = xx or URL = xx2) to combine the pre- and post-change URLs, it gives entries = visits. Looking at the same URLs in a dimension table, the entries is equal to the number of entries on that URL. 

I understand the mechanics behind why this is happening, technically each occurrence of URL = xx did have an entry, but that's not the entry I'm looking for if I'm using a hit-based segment. This behavior feels off to me and I'm wondering if (aside from creating my own bounce rate metric that doesn't rely on entries) there's a solution or workaround.

[vibhatna]

@_bel_ ,

The segment you created is on Hit level but the Entries metric works on Visit level. So let's say the segment you created with definition evar1=Citizen(Hit level). if you apply Entries metric next to this segment it will show the all the values of that evar for that particular visit.

Segment:

Breakdown segment with the dimension used in the segment:

Here you can see all the values of the visit when evar1 was set as Citizen as an Entry.

[_bel_]

Thank you @vibhatna, as mentioned I do understand what is happening. I am saying this does not seem like expected (or desired) behavior for a hit-based segment. A hit-based segment should report the same analytics metrics as that dimension itself in a table. 

I am also asking if there is a solution to report entries/bounce rate for hit segments, given this behavior is not what I need. 

[jcanavan]

Hi @_bel_  (and anyone else struggling with this).
The issue as mentioned is that the total 'entries' to a specific page when using a hit-based segment is replicating the 'visits' figure, which then gives an incorrect bounce rate. 
I have found a solution to calculate bounce rate for a page when using a hit-based segment using the below calculated metric. This is calculating bounce rate from a single page visit view but you could adjust the first part of the calculation to show only visits with 1 recorded hit.

[jeff_bloomer]

@jcanavan Agreed. When it comes to bounce rates, I learned quite a while ago that the out-of-the-box Bounce Rate in AA is not adequate, regardless of Hit- or Visit-based.  You need the Single-Page Visits approach to get what you need.  In fact, the linked video is still one of my favorite references on Single-Page Visit Bounce Rate.

https://youtu.be/cem3n0mgpEA?si=n-7tei4a5ZhQItXZ 

[jcanavan]

Thanks @jeff_bloomer.
Out of interest, have you come across any similar issues where out-of-the-box metrics do not behave as expected?

[jeff_bloomer]

That's a great question.  We talke about some of that during our office hours sesion just the other day:
https://www.youtube.com/@adobeanalyticsusergroups9817