50/50 Random Sample for A/B testing Flow Sequences

  • 15 September 2022
  • 1 reply

Badge +1


I recently put a 50% random sample conditional split to test two different flow setups however my “No” has not received any emails yet. . . I looked at the waiting group, and they seem to not be scheduled until December. Do I need to back-populate the flow? Or, is there another solution to this? 



Best answer by Dov 16 September 2022, 15:55

View original

1 reply

Userlevel 7
Badge +61

Hi @John Parker,

Thanks for sharing this question with us.

There could be a few things happening to explain why users are not processing through the “no” path. Did you only recently add in the A/B split? Did you happen to add in the “No” path of the A/B flow test after the “Yes” path emails were already live? Did you recently update the email status from “draft” to “live” for the “No” path emails? Or have you previously backpopulated the flow? Any of these actions are potential reasons why you’d see users only showing up down the “yes” path.

With that said, the 890 users you have in “waiting” at the top of the flow in the 50/50 random split will remain in “waiting” status until those 90 days have passed - at which point they’ll be distributed between the “yes” or “no” paths. That also explains why they are queued-up for December (because of the 90-day time delay the precedes the split).

These 890 users will process on their own through the flow organically (and be evenly distributed). If your goal is to re-queue these users who “missed” the “No” path - let us know and we’ll work through a solution!

I hope that helps clear things up.