Different count for Page Views for Historical Data

[ElinaMirin]

Hello! I have created in Adobe Analytics a segment for URLs (I only need some to contain some specific words) and I extract my data using Report Builder. I've noticed that my counts for Page Views, Visits, Engaged Visits, Entries, Exits, Total Seconds Spent are changing for historical data (ex: may 2023). Yesterday I had a smaller count than today when I have checked again in Adobe Analytics Workspace. Same scenario last week. Counts are different than this week (for historical data, so 2022 and months before december 2023)

Is there any issues with Adobe Analytics or anybody knows why the counts are changing?

[Jennifer_Dungan]

My guess is the problem lies with how you built your segments.... Using the wrong Segment Scope can result in pulling back data that you don't intend... in particular, using a Visitor scope segment will pull back ALL data for Visitors that match the criteria even once... 

For example, if I create the following segment:

Visitor

    Page equals home

Now let's look at a Visitor Journey:

• Dec 1
  • Page A
  • Page B
• Dec 5
  • Page C
  • Page A
  • Page D
• Dec 8
  • Home
  • Page A
  • Page E

When you looked at your Data on Dec 6, this visitor wouldn't appear, since up until this point, the visitor had not been to the "Home Page"

However, when you look at the data on Dec 9th, this visitor has now seen the home page... since your segment is Visitor scope, ALL the pages for this Visitor are now being returned, and will now show data in your reports for Dec 1 and 5.

If you segments are Visitor scope, you also are returning inflated data even if you were to look at Dec 8 in isolation, because Page A and Page E will be returned in your dataset as well...

You need to use a HIT scoped segment, to only return the URLs that match. You can still use your standard Visits and Unique Visitor metrics with such a segment...  the scope of the segment is about what data set you need to return.

[ElinaMirin]

Thank you for your answer!

The segment is created to include "Hit", not Visit or Visitors.

[Jennifer_Dungan]

Hmm that's interesting...

Would you be able to share a screenshot of your segment? (You can hide specific information, or just replace without saving the values with placeholders that are safe to share)

Also, how much of a difference are you seeing? Do you have a mobile app tracking into the suite that has Offline Data enabled? Or are you ingesting external data sources into your Adobe Data? Because these are "delayed" ways to collect data, this could be impacting your historical data....

[ElinaMirin]

The segment is created on a "HIT" level and includes multiple OR statements with: URL contains any of abc OR URL contains any of def OR URL contains any of ghr etc.

So last week the count for Page Views was around 240k, yesterday morning was around 180k and today when I have checked again the count was around 250k (this is for historical data - may 2023). So the drop in the count is really big and I don't understand what is happening.

[Jennifer_Dungan]

Just to be safe, seeing the actual segment (given all the OR conditions) I want to be sure that the segment looks correct (if that is alright).

And it's interesting the data is dropping, not increasing... this could still be a segment issue, but a different kind... perhaps there is some logic check that is now preventing some of that data from being returned.....

[ElinaMirin]

The data is actually dropping and then increasing and again dropping - it makes no sense.

[Jennifer_Dungan]

I cannot see it, what is the scope of your containers? I also hope they are "Hit" level as well?

Also, what is the attribution of your eVar? When capturing data like URL, they should probably be set up with a Hit expiry (this does mean that the value has to be explicitly sent on Actions as well as Page Views), but it prevents the URL from being carried forward to other pages that may not be setting the value properly.

I don't know if that would cause the "yo-yo" effect you are seeing... 

[ElinaMirin]

Because 'contains any of' operator it's really complex for Adobe Analytics I have divided my conditions into small subsegments/containers. The point is that all the conditions have the same logic: URL contains any of OR URL contains any of OR URL contains any of. All are HIT level

[Jennifer_Dungan]

Ok, that segment shouldn't have any issues... and I agree, the "Contains any of" is too hard to manage (see what's there, edit it, add new items, etc) - I also don't use it and do the multiple OR clauses myself...

This will sound strange, but it might be the "contains any of"... I've seen some oddities around those... it shouldn't, but have you tried a similar segment using just a bunch of standard "Contains" for all the variations??

[ElinaMirin]

I have tried with “contains” operator, but I couldn’t add more values in a single condition. With “contains any of” I could add multiple values, delimited by space. But “contains” it will return 0. Do you know if I can use multiple values and what is the delimiter for this operator?

[Jennifer_Dungan]

No, unfortunately, every check would have to be done separately

eg:

URL contains X

or 

URL contains Y

or

URL contains Z

...

But I was helping someone a few months ago, also using Contains Any Of, and they were having issues... it may not be the problem... and given the complexity of your segment, you could try making a smaller sample with only a few pages and see if the same behaviour persists or not... if the test works (i.e the data looks stable), you could try to rebuild the segment for all your pages in full....

[ElinaMirin]

Ok. I will try. Thank you very much for your advices!

[ElinaMirin]

Hi!

Quick update on this issue. I have extracted again my reports feom Adobe Analytics keeping the same conditions and operators: “contains any of”. I have also checked for a few days the counts and now everything looks good. 
Thanks for your support!

[Jennifer_Dungan]

I'm glad you have a solution!