Discrepancy list and evar count in adobe analytics

[VidhyaAk1]

Hi All,

On article clicked we are storing the card names in evar38 and list2 and having event72 to count the click

When we pulled the report for evar80 with event72  and ist2 with event72.

We can see the difference in event72 count ? is it expected ? why there is different in count although its part of same adobe request .

Network call

Adobe analytics report

evar

list

[pradnya_balvir]

List variables are custom variables, they work similarly to eVars, except they can contain multiple values in the same hit.

You can refer below document link: (go through Video)

https://experienceleague.adobe.com/en/docs/analytics-learn/tutorials/administration/manage-report-su...

[Josh__Stephens]

What are the dimensions settings of your eVar and list var?  I wonder if the persistence and/or allocation is set up differently.  Additionally, list vars are typically set up to capture multiple values in one string.  They'll then be parsed out for reporting.

This link gives more info on allocation & persistence of list variables.  Full allocation gives full credit to each item touched before the success event.  Linear gives equal, partial credit to each touch.

https://experienceleague.adobe.com/en/docs/analytics/admin/admin-tools/manage-report-suites/edit-rep...

[VidhyaAk1]

Hi @Josh__Stephens 

The evar80 setting in AA
evar expires on visit and allocation is most recent

list2
it never expires and allocation is full

event

Could you please tell what is thee exact reason having different count  ?
what would be the recommended approach for this 

[Josh__Stephens]

Hi @VidhyaAk1,

Thank you for sharing those dimension settings.  

With your current settings, for eVar80, all credit for the click will go to the last touch.

For the list var, full credit will go to each touch that's been seen for the lifetime of this Unique Visitor.  For example, if someone has had three touchpoints and then has a click, each of those three touchpoints will be incremented by 1 for the list var.  Only the last touch will be incremented by 1 for eVar80.  This is why you're seeing different results.

The best settings are really based upon your use case and the insights you're looking for.  With the list var, if you selected linear credit, each touchpoint would get partial credit for the click.  So in the example above, each of the three touchpoints would get 1/3 of a click.  When reporting, Workspace rounds to the nearest integer.  

Before making any changes, I'd suggest using Attribution IQ on the metric to look at different models and lookback periods.

[VidhyaAk1]

This list variable is firing on two events one is recommendation event where it stores multiple values and clicked event where it store the single value. and evar80 also store the single value.

If list and evar store single value then how allocation and expiration matters in this case? how it affects reporting basically, I want to understand it.

[Josh__Stephens]

Differing attribution models will yield different results.

Your list variable is set to full attribution and to never expire.  This will result in full credit being given to each value that's ever been recorded for that visitor.

In workspace, you can use Attribution IQ to compare different models.  Here's an example from demo data using orders and page.  

1st column - last touch - gives all credit to the last touch

2nd column - linear 30 days - gives partial credit to every touch seen during the previous 30 days.  (For example, if one person saw 20 pages before placing an order, each page would get 1/20th of an order)

3rd column - participation 30 days - gives full credit to every touch seen during the previous 30 days.  (Same example, one person saw 20 pages before placing an order, each page would get credit for 1 order.)  Note how the breakout values in the 3rd column are much higher than in the 2nd.

Using your eVar and list var, drop multiple copies of your click metric into a table.  Click on the metric settings (gear), then select a non-default attribution model (bottom of the column settings) and you can compare different models.

[bjoern__koth]

Hi @VidhyaAk1 

This is presumably by design. The list var column values are not the actual sum of events, they are the sum of the individual list var _value_ occurrences.

Event X firing twice, one with list var Y set to val1 and once val1,val2 will likely show val1:2 and val2:1, and combined to a column sum of 3.

so, the column sum in a list var context will not match the event instances when looked at them alone.

hope that helps,

Björn