How to calculate
Divide conversions by visitors in each group. Use their combined conversion rate to calculate the null standard error for the difference.
Worked example
Use Visitors or trials in A 1000 · Conversions in A 100 · Visitors or trials in B 1000 · Conversions in B 130 · Alternative hypothesis Two sided · Significance level 5%. The result is ≈ 0.03548845047.
Input and result notes
Use independent assignments and one outcome per trial. The direction compares A minus B.