I'm running some A/B testing in Target. I have set up the test to have a 50:50 split for the selected audience, i.e. 50% of users will see the control and 50% of users will see the test experience.
However, looking at the first few weeks of results, I am seeing a split that is more like 47% vs 53%. I wondered if anyone could help me out with how Target allocates audiences to experiences? Can I expect this figure to level out? Does Target have allowable tolerances which it works within?
Any help you could give would be hugely appreciated.
This is possibly due to how visitors come back to the experience after they convert. They are not being counted again by default. You may have set it up so that they are being counted again, which you can change. Check out the article on Success Metrics in the Adobe Target Product Documentation - “Increment count, release user, and allow reentry after a user encounters the goal metric" could cause returning visitors to skew the visitor metric.