I'm firing two values into prop20 from the same location to measure an a/b test:
I'm getting nearly identical counts of unique visitors exposed to each category, but when I compare registration paths, i only have 47 people completing the 'with-edu-emp' variation, and 3800 participating in the 'new' variation.
Could you provide me some more info on how are you doing the a/b test? Are you using Adobe Target? Some things to take into consideration would be how you've implemented the A/B test and your segment definition. It also may be work looking into if your with-emp value is correctly being defined.