Marketing ID via Data Warehouse



Goal: Determine the entry of an evar/prop against a list of Marketing Cloud IDs.

I have a list of Marketing Cloud IDs that I would like to understand a specific evar associated with them.

What I've tried:

  1. Connecting to Power BI -- but I get unspecified, non uniques exceeded, etc.
  2. Create a segment that looks for a list of 100 MIDs by setting the MID dimension to "contains any of" and lists the MIDs in the criteria separated by spaces. From there I requested the data via the warehouse -- but the data that is returned isn't the Marketing ID but a list of numbers:


I eventually want to do this for upwards of a 10,000 MIDs or so -- is method #2 the best way to do this?

If so, why does the marketing ID return what looks to be just some string of numbers that are not MIDs?


Accepted Solutions (1)

Accepted Solutions (1)



Why not just do a data warehouse pull for Visitor type with MIDs present at time of evar capture? If you must have just a certain set of MIDs then create a segment for those IDs only.

OR... If you want to really kill it...

Your marketing ID use SAINT. Break down a sub category of MID,  Visitor type(Customer, non customer).

When you upload assign each ID its state.

MID                    MID visitor type

10000000x     customer

10000000xx     noncustomer

100000y     customer.

You could then pull you MID report by each sub category or compare them. Or if a segment is visit based use it to see other events by MID visitor types. when you update teh ID(once only needed) then anything new would be unspecified. The saint method would only need updates as frequent as you need them to reflect customer type changes.


Answers (5)

Answers (5)



I have a follow up question on the topic.

We ran a test using Adobe Target a couple months ago and saw interesting behavior for a portion of the treatment traffic (roughly 2,000 visitors). Test has since been turned off. The team is looking to create an AAM audience for this portion of traffic to identify this propensity against other models.

What has been done:

  • List of MIDs exported from Data Warehouse
  • Segment Created with ~2,000 MIDs (scary I know)
  • Segment Shared with AAM

What comes through:

  • Analytics - the segment does eventually load but it's very slow
  • AAM - audience still shows zero visitors (highly unlikely there have been no repeat visitors)

Based on the 50-100 entry limitation you speak of, and given the assumption that the audience can only be identified with MIDs today, is there a better way to do this?

I'm wondering if the solution has to do with the "Select the window for audience creation" option (i.e., create a segment with the Target activity and tell it to look back into the timeframe where the test was running?):





I start to run into performance issues creating a segment around 50-100 values. Pablo is 100% correct that if you're looking to filter lots of unique items like that to use a classification.

Create a new classification for eVar75, download the template, organize each of them into customer/noncustomer, then upload the file back into analytics. Rinse and repeat as you get more MID's that you know are customers/noncustomers. You can also classify only one and leave the other blank.

Once you have eVar75 classified, create a segment based on the classification. Then you'll never need to update the segment; only the classification as new values come in.



This solved the first part for me, thanks! The data is there but have to Open the csv via File, not simply open the csv.

Second question -- what is the limit for the amount of MIDs I can include in a segment before something breaks? 1000? 10000?



That is totally expected... 2 reasons

Your users are

1 accessing it from a different PC/Device

2 They have since cleared the cookie with the original MID

Also if you are trying to open these IDs in excel via open file, beware the cell field needs to be opened specially(via import menu) as it will truncate MIDs.(they are to long)



Thanks for the reply. That is what I did with method #2, yet the MIDs that returned in the data pull are NOT the MIDs that were listed in the segment (or even look like MIDs for that matter).