A/B test significance (two-proportion z-test) API

Run a two-proportion z-test on conversion counts: z = (p2 - p1) / sqrt(pPool(1-pPool)(1/n1 + 1/n2)), two-tailed p-value via the normal CDF, absolute & relative lift, and a significant/inconclusive verdict at the chosen confidence level. Answers 'is my A/B test statistically significant','which variant won','what's the p-value'.

Price$0.01per request

MethodPOST

Route/v1/calc/mkt-ab-test-significance

StatusLive

MIME typeapplication/json

Rate limit120/minute

CacheNo cache

calcmarketingab-testsignificancestatisticsz-testexperimentcro

API URLhttps://x402.hexl.dev/v1/calc/mkt-ab-test-significance

Integration docs

Example request

{
  "controlVisitors": 8000,
  "controlConversions": 320,
  "variantVisitors": 8000,
  "variantConversions": 400,
  "confidenceLevel": 0.95
}

Example response

{
  "formula": "two-proportion z-test: z = (p2 - p1) / sqrt(pPool(1-pPool)(1/n1 + 1/n2))",
  "controlRate": 0.04,
  "variantRate": 0.05,
  "absoluteLift": 0.01,
  "relativeLiftPercent": 25,
  "zScore": 3.0509,
  "pValue": 0.002282,
  "confidenceLevel": 0.95,
  "alpha": 0.05,
  "significant": true,
  "winner": "variant"
}

Input schema

{
  "type": "object",
  "required": [
    "controlVisitors",
    "controlConversions",
    "variantVisitors",
    "variantConversions"
  ],
  "properties": {
    "controlVisitors": {
      "type": "number",
      "examples": [
        8000
      ]
    },
    "controlConversions": {
      "type": "number",
      "examples": [
        320
      ]
    },
    "variantVisitors": {
      "type": "number",
      "examples": [
        8000
      ]
    },
    "variantConversions": {
      "type": "number",
      "examples": [
        400
      ]
    },
    "confidenceLevel": {
      "type": "number",
      "description": "decimal in (0,1), default 0.95",
      "examples": [
        0.95
      ]
    }
  }
}

Output schema

{
  "type": "object",
  "additionalProperties": true
}