Can you 50/50 test a whole email?

[Michaela_Morale]

I want to set up an A/B test for an email send. I have 2 versions of the email. I want to send it 50/50 to everyone. Can this be set up through the A/B test option? I don't want the winner to be sent out since I want it to all go out at one time and then see which is possible.

[Josh_Hill13]

Sure.

Just set the Sample to 50/50 and there will be no winner sent. Do you mean something else?

Probably best to use Random Sample method.

[Michaela_Morale]

I guess I'm unclear how to do that since the slider options are Receives Test and Receives Winner

[Valerie_Whitin2]

Hi Michaela Morales​

You would set the test sample size to 100% for it to work that way. That means it will evenly split your test among 100% of the audience you selected. Which would in turn be a true 50/50 split. It will still make you select a time to "send the winner" but really nothing will happen other than you will get an email with the results of the test. This is how I've almost always run my tests and have never had any issues. I hope this helps answer your question

[Michaela_Morale]

Yes it does, thank you Valerie!

[Natascia_Almeid]

Hi Valerie,

great idea! But can you also track which version "wins"?

Thanks

[Nithish_Bhat]

you can add a new recipient in the list, after the test completes. Winning mail goes to the new recipient.

[Valerie_Whitin2]

Hi Natascia Almeida​ - yes, it will still let you know which one wins. You will get the normal email program dashboard results with the A/B test results broken out.

[Ankit_Dua]

You can send the email via Random sample 50/50, by this the email will be sent to all the people who qualifies for the campaigns and you can run a report to check which email performed better.

You can do it via A/B testing too, as you have shared in the snapshot.

[Alex_Baumgarten]

Be careful doing a 100% test. If you have a clean database with no duplicates you should have no issues. But, because of how Marketo handles A/B tests, the winning email may be sent out to duplicate contacts once as a test and then again once a winner is declared. You could avoid this by making the winner declaration manual rather than automatic.