Inconsistent outcomes for a custom metric

[kirkramble]

I am seeing some inconsistent outcomes for a custom metric I have created.

Custom metric is called ‘Progression’ and it was created to measure how efficiently our visitors are moving down the purchase funnel.

The combined ‘Progression’ metrics is the sum of multiple sub metrics reflecting visitors moving down the step of our purchase funnel:

• Top of funnel progression
• Middle of funnel progression
• Bottom of funnel progress
• Conversion initiation progression

When reviewing the output for all sessions, this metric seems to function correctly.

Note:  The metric counts only when a session views a page lower in the purchase funnel than the session entry page.  As such, not all sessions are eligible for each of the sub metrics.

Ex:  A session starting on a bottom of the funnel landing page is not eligible to be considered “progression” if they visit a top of funnel page.

The issue I am having is that I see a zero for the final output for sessions not eligible for every sub metric. 

Note:  I see zeros in the UI for these sub metrics, but when I export, I see blank or missing values.

My theory is that Adobe Analytics is reporting a null or blank when a session is not eligible for a specific metric, and when the combined metric attempts to add up the component values it is not able to do so.

Can anyone confirm this is likely case?  If so, is there a way around this issue as it is understood not all the sub metrics will have values for all sessions?

[Jennifer_Dungan]

Hi @kirkramble,

I am not sure if I fully understand what you are trying to do... events in Adobe are stored as a list of event ids... if an event is not fired, then it won't be included in the list.

Such as "Event 1" is "200", "Event 2" is "201", etc.

So when you send:

s.events = "event1,event2"

In Adobe it's actually stored as "200,201"

Clearly events 3-1000 are not triggered above, and they wouldn't be included in how Adobe stores the data. So you are correct, no events would be empty in the actual storage, but this shouldn't impact adding values together... if not part of the events_list, then it will be treated as 0 in the interface.

In your custom metric, what is the logic?

If it's simply a "event1 + event2 + event3 + event4" this should work fine...

If you have some advanced functions involved then maybe there is an issue with that?

If you can share some more information this would help us to help you.