Spliting a query into two random lists

Michael Dugan

Hi.  I am looking to run a pre-call postcard test on a select group of our alumni. I want to send a pre-call postcard to only half of the group and be able to test the results at the end of the appeal. What I am having trouble doing is breaking the list into two random lists that I can then export.  I can export the whole group and use excel to randomly split the group but then I have no way of getting that data back into RE, as far as which group the individuals were included on.  Does anyone have any suggestions?

Mark Guncheon 2

If you are using a special appeal for this, have two packages, one for each group.

You could divide the group by last names alphabetically.

Quick thoughts on a Thursday.

Kylea Johnson

You could import an attribute designating A or B list, then delete it once you're done (or keep it).

Christine Adams

You could create a query based on the IDs in the export or a regular query and go to query options and use the apply output limits and do random sampling that way

Christine Cooke 2

as mentioned you could import a different appeal code for each group or you could import in Action for each of the two groups.  I would lean toward action myself

Dariel Dixon 2

Michael Dugan:

Hi.  I am looking to run a pre-call postcard test on a select group of our alumni. I want to send a pre-call postcard to only half of the group and be able to test the results at the end of the appeal. What I am having trouble doing is breaking the list into two random lists that I can then export.  I can export the whole group and use excel to randomly split the group but then I have no way of getting that data back into RE, as far as which group the individuals were included on.  Does anyone have any suggestions?

Why can't you just import the list back in via constituent code?  I'm sure you're exporting the consID out into your spreadsheet/.csv anyway.  I would just export the list, and then use a random sampling to separate the list in Excel, and use the one of function to create a query to import the action/appeal back into the database.  Then you can use a global add to add the appeal codes you want to use (or package, or whatever).  Using the SmartPaste plugin will help with getting over the "one of" limit.  I think it's a good idea to run an A/B test like you're doing. Hopefully you'll get some good data from your experiment.

Russell Presgraves

I usually chop the lists in Excel and use different package codes, but there is a random sampling option under Query Options.

https://kb.blackbaud.com/articles/Article/71197