Allow Classification Rule Builder to work on more than 1 level deep

Avatar

Avatar

yuhuisg

MVP

Avatar

yuhuisg

MVP

yuhuisg
MVP

27-12-2020

Description - When a variable has nested classifications (i.e. classifications that are more than 1 level deep), Classification Rule Builder is unable to operate on those nested classifications.

Why is this feature important to you - It is common to have classifications that are nested. Furthermore, the Rule Builder's interface implies that it can operate on nested classifications, when in reality, it cannot. This is the biggest bug.

How would you like the feature to work - When there is a rule that operates on a nested classification, then that rule should work in the same way as with top-level classifications.

Current Behaviour - Nested classifications cannot be classified with the Rule Builder, even though the Rule Builder interface implies that it can be done.

Adobe Analytics Analytics classification classification rule builder classifications
5 Comments

Avatar

Avatar

jkm-disco

Avatar

jkm-disco

jkm-disco

31-12-2020

Hi @yuhuisg ,

As an alternative, one can create a regex with nested capturing groups. Is the idea to make it simpler for users less comfortable with these expressions? In theory, the computation of your suggestion would have the potential to be simpler if users were using only "starts with" and/or "end with" conditions on nested classifications as versus regex.

Avatar

Avatar

yuhuisg

MVP

Avatar

yuhuisg

MVP

yuhuisg
MVP

02-01-2021

@jkm-disco , the problem isn't to do with making life simpler for AA users. It is about making Classifications work "as advertised".

Given this setup:

eVar1:

  -- Classification A

  -- Classification B

      -- Classification B1

      -- Classification B2

  -- Classification C

Classification Rule Builder works with Classifications A, B and C, but not Classification B1 nor Classification B2, i.e. classifications that are not at the top level.

That is the problem that my idea attempts to address.

Avatar

Avatar

jeff_bloomer

Avatar

jeff_bloomer

jeff_bloomer

12-02-2021

@yuhuisg,

I believe I am a bit confused by your ask.  I have personally done this successfully in Adobe.  In our interface for Classification rules, the nested classification shows as a choice in the drop-down with the primary value followed by the caret and then the nested value.

Using your example: Classification B^Classification B1.  As a result, you would have a minimum of five rules to manage your Classification.
I have already seen this done successfully on more than one variable in AA.  Our example was an ID in a string that had a secondary ID and we wanted to put those into nested values as well.
The expression looked like this:
^(.*)\|(.*)\|(.*)\|(.*)_(.*)\|(true|false)\|(.*)\|(.*)\|(.*)$
The IDs were in the 4th and 5th positions.
In my Classification, I had a total of 11 rules to manage that string.  One rule addressed missing data, which assigned "not available" to one of my primary values, but I had two rules for position 4 - one for a primary ID, and then I also assigned it to the nested value.  I also assigned position 5 to a nested value.
Is that how you're expecting your Classification Rules to work?

 

Avatar

Avatar

yuhuisg

MVP

Avatar

yuhuisg

MVP

yuhuisg
MVP

13-02-2021

@jeff_bloomer Yes, that is how I'm expecting my Classification Rules to work. But they don't. So every time there is a new Product, then even though I have valid Classification Rules to assign values to the nested classifications, those nested classifications never get set, even several days after the Product has been tracked. So I end up having to use Classification Importer to get the classifications in place.

Avatar

Avatar

jeff_bloomer

Avatar

jeff_bloomer

jeff_bloomer

01-03-2021

@yuhuisg That's a fascinating problem.  I would highly recommend speaking to your TAM or contacting support to see why it's not working correctly.  That sounds like a legitimate technical problem that needs to be evaluated.  Hope they can help you figure that out.