Capture multiple values via Classifications?

[mjdavid007]

Hi, I'm working with Classifications, which have proven to be very useful when it comes to give more value to our data. We are facing now a challenge, though: some of the classifications might come with multiple values on the same cell. For instance, let's imagine we have an activity ID, which is captured on an eVar, then we use classifications to give this activity more attributes:

•  Key: activityID
  
  
  • activityName: Webinar on shoe industry in Germany and France
  • activityType: Webinar
  • activityMarket: Germany, France

The problem is that, although I could split that activityMarket into two, I need to allocate them on the same classification level (activityMarket, in this example). I cannot create another classification with the same name, and create a different one, like activityMarket2, would not be useful for us as we need to breakdown only by activityMarket and allocate all values there.

Any thoughts on this?

Thanks a lot.

[RobertBlakeley]

Help me understand.

You cannot do this:

• activityName: Webinar on shoe industry in Germany and France
• activityType: Webinar
• activityMarket: Germany, France
• activityMarket2: Germany

But if you get:

• activityMarket: Germany, France
• activityMarket: Germany, UK
• activityMarket: Germany, Italy

You want a value to see the total allocation for Germany.

• activityMarket: Germany, France     5
• activityMarket: Germany, UK           2
• activityMarket: Germany, Italy         1

So in the above, you want to see Germany = 8, UK =2, etc.

Is that correct?

[mjdavid007]

Hi @RobertBlakeley , it is as follows: imagine we have 3 activities that receive, on their activityMarket classification, the following values in one simple cell:

- activityMarket (for activity 1): Germany, France

- activityMarket (for activity 2): German, UK

- activityMarket (for activity 3): UK, Italy, Germany

When we use activityMarket in Workspace, we would like to have:

- activityMarket:

• Germany: 3
• UK: 2
• France: 1
• Italy: 1

I hope now it's easier to understand.

[RobertBlakeley]

You can always filter the activityMarket report by Germany, UK, etc .and use the report totals value.

I don't think you can use a list variable as that only applies to the parent variable, not the classification.

You could use a sub-classification but if you could do that, you could probably also do an activityMarket2. So that does not seem to be an option.

I am out of ideas.

[mjdavid007]

Yeah, unfortunately I came to the same conclusion, that I could not use List variables either and that I could only use filters. However, that would not allow me to see, for instance, which markets had the highest number of occurrences.

Thank you anyway for your help! Hopefully somebody else can come up with a different approach.

[Josh__Stephens]

I'd probably play with using segments for each of your markets (if activityMarket contains Germany).  Otherwise, might be better to use a list variable.

[mjdavid007]

Hi @Josh__Stephens , thanks for the answer but unfortunately that would only allow me to search data for a market (or some) in particular, but it would not let me see a full ranking.

I guess that Lists will be my only choice, so now my question would be whether it is possible to make a list out of a classification value (and not the other way around, which I know it is possible).

[Sergei_Kalinichenko]

Hi there!

I'm a little bit confused, please allow me to lift up this topic: from the replies here it seems it is not possible to store multiple values in one hit for a single evar.

Like in subject:

evar1 = Germany, France, Italy  | hit = 1

then workspace returns

evar1 | hit

Germany | 1

France | 1

Italy | 1

On the other hand, there is a post Solved: Multiple Value eVar - Adobe Experience League Community - 576401 from '23 where @MandyGeorge and @Jennifer_Dungan were referring to classifications (Hello both!).

So, I'm a little bit confused if classification works or not. Would be helpful before submitting a ticket to my web develipers.

Thank you in advance!

[Sergei_Kalinichenko]

Upd, obviosly: 

Classified evar1 | hit

Germany | 1

France | 1

Italy | 1

[Jennifer_Dungan]

Well, you can and you can't... which I know is confusing...

Classifications work well for different types of information, sent in a specific pattern so that it can be split using Regex.

For instance, I like to use a single dimension to track search results details:

[Page Num],[Number of Pages],[Results per Page],[Sort Order]

So I can use Regex to:

• set the first value into my "Page Number" classification
• set the second value into my "Number of Pages" classification
• set the third value into my "Results per Page" classification
• and set the 4th value into my "Sort Order" classification

Four very distinct classification groups... where I know there will always be 4 pieces of information passed...

In you case, unless your country values are locked to "3 values in a very specific order", where "France" in the first position should be uniquely reported on from "France" in position 2 or 3... this won't work well....

The problem is that each classification acts like a separate dimension... so if your intention is to get all "France" hits counted in one group, you can't, because you might get:

• France,Germany,Italy
• Italy,France,Switzerland
• Belgium,Italy,France
• etc

The first classification would only have France, Italy and Belgium each counting 1

The second classification would only have Germany, France and Italy each counting 1

The last classification would have Italy, Switzerland and France, each counting 1

And this is assuming you are only ever passing 3 values.. if it might pass 2, or 4 or 5 values, this gets even harder, and more split.

You could never get a consolidated "France" counting 3

Unless you are only dealing with a small subset of countries, where you can make a classification for each country, which seems unlikely... I suppose you could do rules like "Contains France" and then count the "France Classification".... 

Each classification can only be set once per hit, you cannot take disparate values of the same thing, and break them into individual values in the same classification.... 

I hope that makes sense?

[Sergei_Kalinichenko]

Perfect explanation, thank you! However, makes me sad...

So my idea won't work. Now, I'm wondering if there is a better way than creating a segments and building a workspace around segments. Example to address:

hit 1: evar1 = France,Germany,Italy

hit 2: evar1 = Italy,France,Switzerland

hit 3: evar1 = Belgium,Italy,France

Then creating a segments:

Jurisdiction: FR = evar1 contains France

Jurisdiction: DE = evar1 contains Germany

Then workspace:

segments | hits

Jurisdiction: FR | 3

Jurisdiction: DE | 1

Good for core markets, but may be tough to maintain for growing interest and small regions.

[Jennifer_Dungan]

If I were you, I would use a List Dimension... 

You have three available to you (so long as they aren't already in use... (if they are, with a few tricks, you can make your list do multiple uses....)

For now, let's keep it simple...

s.list1 = "France,Germany,Italy"

Set up your list with a comma delimiter... this will automatically split your values on the comma (no need for classification rules)

So for example:

Hit 1:

• List1 = "France,Germany,Italy"

Hit 2:

• List1 = "Germany,France"

Hit 3

• List1 = "Belgium,Switzerland,France"
  

In your report, you will see:

		Page View
List 1		3
	France	3
	Germany	2
	Italy	1
	Belgium	1
	Switzerland	1

Lists don't depend on the same number of values to be sent, but each value will be treated as its own value.... 

Now, if you need to see the combination of data sent, I think you can see it in Raw Data, but I don't think you can see it in the Workspace... if that is important to you, you might want an eVar for the "full combination" and a List for the "separate values"