I have a question with regard to analyzing A/B test results in Adobe Analytics. Thanks to the attributes 'target activities' and 'target experiences', I'm able to do extensive analysis on my default and different variant groups. One of the important KPI's I would like to check: what's my number of 'users with an order'? I have the idea to make a segment with following condition 'orders exits' - visitor container. After applying this filter, I know the answer on my question by looking at 'unique visitors'.
But taking a close look at the data, I see unexpected results. How to explain that my number of orders is lower than my number of unique visitors? I would expect that this number is higher, as a unique visitor can place several orders.
Using this same KPI as main KPI, I would also like to deep dive my data by making a distinction between new and returning visitors. Therefor, I use the following segment to breakdown my experiences:
Return visits = visit number > 1
New visits = visit number = 1
Will my segment 'users with an order' make sense doing this deep dive? Will I receive reliable data?
Many thanks in advance!
I would say that the ideal numbers should be : Unique Visitors = Orders or Unique Visitors < Orders and should not be Unique Visitors > Orders.
Can you revise your Segmentation by Orders is greater than or equal to 1 and validate the number?
Regarding deep dive, you can append as many segments as you can. That would definitely help you to get the answer!
Unfortunately yes, I still see the following if I take a look at my figures on the level of 'target activities':
# of orders + 25% = number of unique visitors (after applying the segment).
Seems that I have to look for another KPI to validate my hypothesis (A/B testing)?
This is fine. Still your Unique Visitor number is greater than Orders?
Can you share your Segment(As a screenshot)? Need to see the same before advising you.
Thank you for your answer. I've adapted my segment according to your feedback, but that unfortunately doesn't change my figures. Maybe it has something to do with my 'window for audience creation'? I'm revising data from April and kept the default of 90 days.
What's your opinion?
Thank you in advance!