We’re smarter together. Learn from this collection of community knowledge and add your expertise.

How do I make sense of Customer Intelligence Metadata?

by SAS Employee ESuttonCI on ‎12-15-2015 03:46 PM (214 Views)

If you regularly use SAS Customer Intelligence Studio (SAS CIS), you will see that some of the data items in your information map can be selected from a drop-down list and others can’t. For example, Gender is a good example of when you see a list and Address is one for which you never see a list.

 

This list is what we call Customer Intelligence Metadata (not to be confused with Metadata Servers or SAS Metadata Repository.)

 

SAS CIS has this metadata functionality for two reasons:

 

1 – To make it quicker and more accurate for you to select fields. Wouldn’t you agree it is easier to select the word FEMALE from a drop down list than to have to type it in all the time?

 

2 – To help you approximate how many items fit in certain groups, which may help you make better selections when designing your campaigns.

 

For example, when you have metadata set on Gender, you may see something like the screenshot below, an approximate count and percentage of how many subjects (Customers) fall into each category. This can help you verify that the counts you are expecting are valid.

 

Customer_Intelligence_Metadata.png

 

Metadata Generation is the SAS process that updates the tables that store this CI Metadata. You decide how often you run this process, which is purely for usability. When the campaign/nodes are run, the actual database is accessed to get the counts, so there’s no need to update every time the database is refreshed.

 

Metadata exists to give the campaign builder an estimate of the data behind the scenes. While it doesn’t have to be 100% accurate, it can be if the data is refreshed daily and you generate metadata daily. Is that necessary? If the system can go a week and still be usable from a campaign-building perspective, then don’t create extra work for yourself. Just update it weekly.

 

Bottom line: Don’t worry the next time you run a count on the number of females you have and the count you get in the node differs from the count in the properties of the node. This is normal. It probably just means your data has been refreshed and the metadata generation process hasn’t yet been updated!

 

Top Tip: When you think about fields on which you want to generate metadata, only use data items with a small number of unique values.

 

Examples of fields that are great for metadata include:

 

STATE – as in Texas, Florida etc

TITLE – Mr, Mrs, etc

STATUS – Married, Single etc

 

Never use these fields:

 

Zip code

Address

Email address

Telephone number

Contributors
Your turn
Sign In!

Want to write an article? Sign in with your profile.