Your achievements

Level 1

0% to

Level 2

Tip /
Sign in

Sign in to Community

to gain points, level up, and earn exciting badges like the new
Bedrock Mission!

Learn more

View all

Sign in to view all badges

SOLVED

Adobe Target A/B testing not split equal

Avatar

Level 1

 

Hi,

 

We've had an A/B test with 1 variant running at Manual (50:50) traffic split for the last 7 days. However we're seeing more of a 65% 37% split in Visit and for unique visitor it is coming 64% 42%. Is there any reason for the test not being closer to 50%:50%?

We are using Manual Splitting in Target.

 

1 Accepted Solution

Avatar

Correct answer by
Employee Advisor

@Arnab3 In AB test , the traffic distribution is random in beginning which means even-if you have set the  traffic distribution to 50/50 , still you may see 20/30 , 60/40 split in initial days . Although later with the increased traffic the numbers would move towards the allotted traffic split. 

There are few factors that could cause this discrepancy : 

 

1. Traffic on the activity qualification page.

2. Reporting source (Target or Analytics) : In these cases , unstitched hits may lead to variance in reporting numbers.

3. Modification in Activity setup : Scenario's wherein the traffic distribution was initially set to 40/60 but after few days of activity being live , the distribution was changed to 50/50.

 

Hope the details above are helpful.

View solution in original post

2 Replies

Avatar

Level 7

Hi @Arnab3  how is the traffic on the page where activity is defined.

Usualy if traffic is less then AB test takes longer time to normalize the split.

Avatar

Correct answer by
Employee Advisor

@Arnab3 In AB test , the traffic distribution is random in beginning which means even-if you have set the  traffic distribution to 50/50 , still you may see 20/30 , 60/40 split in initial days . Although later with the increased traffic the numbers would move towards the allotted traffic split. 

There are few factors that could cause this discrepancy : 

 

1. Traffic on the activity qualification page.

2. Reporting source (Target or Analytics) : In these cases , unstitched hits may lead to variance in reporting numbers.

3. Modification in Activity setup : Scenario's wherein the traffic distribution was initially set to 40/60 but after few days of activity being live , the distribution was changed to 50/50.

 

Hope the details above are helpful.