A/B Test Result Analysis Tool

Sampling ok: ☑️ ✖️
Statistically Significant: ☑️ ✖️
Practically Significant: ☑️ ✖️

Conclusion: Sampling looks biased - invalid experimental results Experiment outperformed - keep new flow Run experiment longer to get practical significance or make judgement call Control outperformed - keep old flow Fill values to get experimental results



Control Conversions (Xcont)

Control Visitors (Ncont)


Minimum Detectable Effect (dmin) ex: "5" for 5%

Experiment Conversions (Xexp)

Experiment Visitors (Nexp)

Control CTP = {{ pCont }}

Experiment CTP = {{ pExp }}

p̂pool = {{ pPool }}

SEpool = {{ sePool }}

d̂ = {{ dHat }}

m = {{ m }}