Segments

GET /experiments/{testId}/segments/{view}

Returns the experiment's results split by visitor segment, so you can see where a change worked and where it did not. Four views return the same shape with different metrics.

View	Answers
summary	How did each segment convert and earn?
checkout	Where did checkouts stall, and how were discounts used?
browsing	How did visitors move from collections to products to the cart?
subscriptions	How did subscription orders and revenue differ?

Parameters

Parameter	Description
testId	Path parameter, from /experiments
view	summary, checkout, browsing or subscriptions
startDate, endDate, timeZoneName	The date window
Filters	Any filter, applied before the split

Response

Field	Type	What it holds
view	string	The view you asked for
segments[].dimension	string	device, user_type, traffic_source or source_channels
segments[].name	string	The segment, such as MOBILE, new_visitors or google
segments[].visitors	object	control, variations and total, each number | null
segments[].variations[]	object	variationId (integer), isControl (boolean), lift (number | null) and the view's metrics, each number | null

The summary view's metrics are visitors, conversions, conversionRate, orders, revenue, revenuePerVisitor, averageOrderValue, profit and profitPerVisitor. Each other view returns its own set, shown in the example and in the OpenAPI document.

curl https://elevateab.app/api/external/v1/experiments/TEST_ID/segments/summary \
  -H "Authorization: Bearer YOUR_API_KEY"

{
  "success": true,
  "data": {
    "view": "summary",
    "period": {
      "dateRanges": [
        { "startDate": "2026-09-01 08:00:00", "endDate": "2026-09-18 11:58:35" }
      ]
    },
    "appliedFilters": {},
    "segments": [
      {
        "dimension": "device",
        "name": "MOBILE",
        "visitors": 12310,
        "variations": [
          {
            "variationId": 1001,
            "isControl": true,
            "visitors": 6155,
            "conversions": 191,
            "conversionRate": 3.1,
            "orders": 196,
            "revenue": 12600.4,
            "revenuePerVisitor": 2.05,
            "lift": null
          },
          {
            "variationId": 1002,
            "isControl": false,
            "visitors": 6155,
            "conversions": 210,
            "conversionRate": 3.41,
            "orders": 216,
            "revenue": 13500.2,
            "revenuePerVisitor": 2.19,
            "lift": 9.6
          }
        ]
      }
    ]
  }
}

Worth knowing

lift compares different things per view: conversion rate on summary, completion rate on checkout, product view rate on browsing. The subscriptions view has no lift.

• Segments are not filters. This splits one result set four ways; to narrow the experiment itself, use filters.

• Small segments move a lot. Check visitors before reading anything into a segment's rates.

• Unnamed segments are left out, so the segments will not always add up to the experiment's totals.