Multiple sends of the same email message, whether it be A/b Testing or individual sends, can cause problems for your intended audience generation.
To lay the groundwork, our system will not resend the exact same email message to the same recipient. To resend, the message itself has to be slightly altered to be sent out again to the same people.
With that in mind, if you send multiple copies of the exact same message to various and overlapping groups, it could affect the overall audience calculations. This can include normal email sends and A/B testing. You might even see the following filter in place on your email statistics: 'Has been sent this message already'.
If you schedule the message the same day as delivery, the audience will be generated at delivery in chronological order. Where things get tricky is if you schedule the delivery a day or more in advance. For these scheduled sends, the audience is generated the night before, regardless of the time in which they're sent the next day. These audience generations are independent of each other. Even though the message has not yet been sent, the audience has already been generated. This is what can cause one or two of the messages to have an audience of zero before the delivery even occurs.