Single calculator

• Inputs: Visitors V, Conversions C, (optional) Revenue R.
• Conversion Rate (CR): CR = C / V
• Non-converters: V − C  |  % non-converters: 1 − CR
• Revenue per Visitor (RPV): RPV = R / V
• Average Order Value (AOV): AOV = R / C (shown only if C > 0)
• Display precision: percentages use the decimal precision you select (0–3).

A/B test calculator

• Per-variant CR:
  p₁ = x₁ / n₁     p₂ = x₂ / n₂
  where n = visitors, x = conversions.
• Lift (B vs A):
  Absolute: Δ = p₂ − p₁
  Relative: Δ / p₁
• Two-proportion z-test (two-tailed):
  Pooled p̂ = (x₁ + x₂) / (n₁ + n₂)
  SE_pooled = √[ p̂(1 − p̂)(1/n₁ + 1/n₂) ]
  z = (p₂ − p₁) / SE_pooled
  p-value = 2 × (1 − Φ(|z|))
• 95% CI for difference (B − A):
  Δ ± 1.96 × √[ p₁(1 − p₁)/n₁ + p₂(1 − p₂)/n₂ ]
  (Unpooled/Wald interval.)
• Significance rule: “Significant” if p < 0.05; otherwise “Not significant”.

Assumptions & tips

• Outcomes are binary (converted / not converted) and independent within each variant.
• Normal-approximation methods are most reliable when each group has enough conversions and non-conversions (rule of thumb: n·p·(1−p) ≳ 5).
• Very small samples or extreme CRs (near 0% or 100%) can make Wald CIs unstable; consider exact tests or Wilson intervals in such cases.
• The tool uses a two-tailed test; a strictly directional hypothesis would use a one-tailed test.
• No adjustments are made for multiple comparisons or repeated peeking; plan experiments accordingly.