How to create random segments from data extension and send over multiple days (IP Warming)

Question

I have a rather large data extension I want to mail to over the course of a week. I want to use SFMC's built in throttling tool (i'm warming up a new IP). I'd like to take my data extension, break it into 4 or 5 random segments and mail to each of them at different throttled rates. For example, segment 1 at 500 emails per hour for a day. Segment 2 at 1000 emails per hour for a day and so on...

I'm having a hard time figuring out the best way to break my data extension into randomized segments. Is there a tool do this is Marketing Cloud?

Thanks!

Answer 1

while the query definitely works there is a really handy tool for random data extension splitting (we are on a Corporate Edition).

How to go about it:

1. Click Create in your data extension folder
2. Select "Random data extension"
3. Select your source data extension
4. Design your randomized data extension structure with this wizard:

Concerning your dedup issue you should be able to create an exclusion script: https://help.salesforce.com/apex/HTViewSolution?id=000213980&language=en_US

As the marketing cloud is adding subcribers to the All Subscribers List on the first send from a data extension you should be able to use this list (_Subscribers) as target and filter against the EmailAdress field with a Rowcount > 1. Duplicates would then be filtered out at send. Don't forget to set "Deduplicate by Email Adress" when you create the send to avoid duplicate sending within the same data extension (I'm not sure at which send process step the dedup check against the subscriber list comes in).

After all sends are through you can use the source list and the All Subscribers for a join query to create a new deduplicated source data extension.

I hope this works for you.

Have a nice weekend, Stephan

Answer 2

You can select a random sample of a Data Extension using a Query Activity that orders rows by newID(). Like this, where DataExtension is the master:

select distinct top 5000
d._ContactKey as ContactKey
, d.emailAddress
from DataExtension as d
order by newID()
/* target: DataExtension1 */

The other queries could exclude the previous data-sets:

select distinct top 5000
d._ContactKey as ContactKey
, d.emailAddress
from DataExtension as d
left join DataExtension1 d1 on (d1.ContactKey = d._ContactKey)
where d1.ContactKey is null
order by newID()
/* target: DataExtension2 */

The third segment would be something like this:

select distinct top 5000
d._ContactKey as ContactKey
, d.emailAddress
from DataExtension as d
left join DataExtension1 d1 on (d1.ContactKey = d._ContactKey)
left join DataExtension1 d2 on (d2.ContactKey = d._ContactKey)
where d1.ContactKey is null
and d2.ContactKey is null
order by newID()
/* target: DataExtension3 */

Just add additional left joins for any subsequent segments.