Timeseries

GET /experiments/{testId}/timeseries

Returns each variation's metrics split into time buckets, plus the visitors across all variations. Use it to draw a chart of how an experiment moved, or to check when a result started to hold. For the current numbers rather than their history, use /summary or /results.

Parameters

Parameter	Description
testId	Path parameter. The experiment's ID, from /experiments.
mode	cumulative (default) for running totals up to each bucket, or interval for each bucket on its own.
startDate, endDate, timeZoneName	The date window. The time zone also decides where each bucket starts.
Filters	Any filter, applied before the buckets are built.

Response

Field	Type	What it holds
mode	string	cumulative or interval, as asked for
interval	string	The bucket size as an ISO 8601 duration, P1D today
timeZoneName	string	The time zone the buckets are aligned to
period.dateRanges[]	object	The running periods the buckets cover
totalVisitors[]	object	start (string) and visitors (number | null), one per bucket
variations[].variationId	integer	The variation
variations[].name	string | null	Its name
variations[].isControl	boolean	Whether it is the control
variations[].points[]	object	start (string) plus the metrics below, each number | null, in the same order as totalVisitors

Each point's metrics: visitors, conversions, conversionRate, revenue, shippingRevenue, revenuePerVisitor, averageOrderValue, addToCarts, checkoutsStarted, checkoutsCompleted, completionRate, subscriptionOrdersPerVisitor, collectionPageViewRate, productPageViewRate, collectionToProductRate, productToCartRate, cartToCheckoutRate, exitBeforeProductRate.

curl "https://elevateab.app/api/external/v1/experiments/TEST_ID/timeseries?mode=interval&timeZoneName=America/New_York" \
  -H "Authorization: Bearer YOUR_API_KEY"

{
  "success": true,
  "data": {
    "testId": "3f2a9c1e-5b7d-4e8a-9f10-2c4d6e8a0b12",
    "mode": "interval",
    "interval": "P1D",
    "timeZoneName": "America/New_York",
    "period": {
      "dateRanges": [
        { "startDate": "2026-09-12 09:52:05", "endDate": "2026-09-18 11:58:35" }
      ]
    },
    "appliedFilters": {},
    "totalVisitors": [
      { "start": "2026-09-12T00:00:00-04:00", "visitors": 23 },
      { "start": "2026-09-13T00:00:00-04:00", "visitors": 13 }
    ],
    "variations": [
      {
        "variationId": 53397,
        "name": "Control",
        "isControl": true,
        "points": [
          {
            "start": "2026-09-12T00:00:00-04:00",
            "visitors": 10,
            "conversions": 2,
            "conversionRate": 20,
            "revenue": 107.1
          },
          {
            "start": "2026-09-13T00:00:00-04:00",
            "visitors": 7,
            "conversions": 1,
            "conversionRate": 14.29,
            "revenue": 139.4
          }
        ]
      }
    ]
  }
}

Worth knowing

• Read interval rather than assuming a day. Longer experiments may get larger buckets later, and the shape will not change.

• The last bucket is usually partial, ending at the end of the period or at the current moment for a running experiment. In interval mode its counts look lower for that reason; its rates are unaffected.

• A visitor counts in every bucket they were active in, so in interval mode the buckets add up to more than the experiment's unique visitors. cumulative mode counts each visitor once.

• Paused periods are not buckets. They are missing from the series, which is what period.dateRanges describes.

• Personalizations return only the personalized variation.