Advanced Use Cases

This document describes several advanced features of the Google Analytics Data API v1. For a detailed reference of the API, see the API Reference.

Listing Custom Definitions and Creating Reports

The Data API can create reports on registered Custom Dimensions and Custom Metrics. The Metadata API Method can be used to list the API names of your Property's registered Custom Definitions. These API names can be used in Report Requests to the runReport method for example.

The following sections show examples for each type of Custom Definition. In these examples, replace GA4_PROPERTY_ID with your Property ID.

Event-Scoped Custom Dimensions

Step 1: Query the Metadata API Method with your Property ID.

GET https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID/metadata

Step 2: Find the Event-Scoped Custom Dimension you are interested in creating reports on from the response. If the dimension is not present, you need to register the dimension.

"dimensions": [
...
    {
      "apiName": "customEvent:achievement_id",
      "uiName": "Achievement ID",
      "description": "An event scoped custom dimension for your Analytics property."
    },
...
],

Step 3: Include the custom dimension in a report request. The following is a sample request to the runReport method.

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dateRanges": [{ "startDate": "2020-09-01", "endDate": "2020-09-15" }],
  "dimensions": [{ "name": "customEvent:achievement_id" }],
  "metrics": [{ "name": "eventCount" }]
}

User-Scoped Custom Dimensions

Step 1: Query the Metadata API Method with your Property ID.

GET https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID/metadata

Step 2: Find the User-Scoped Custom Dimension you are interested in creating reports on from the response. If the dimension is not present, you need to register the dimension.

"dimensions": [
...
    {
      "apiName": "customUser:last_level",
      "uiName": "Last level",
      "description": "A user property for your Analytics property."
    },
...
],

Step 3: Include the custom dimension in a report request. The following is a sample request to the runReport method.

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "entity": { "propertyId": "GA4_PROPERTY_ID" },
  "dateRanges": [{ "startDate": "7daysAgo", "endDate": "yesterday" }],
  "dimensions": [{ "name": "customUser:last_level" }],
  "metrics": [{ "name": "activeUsers" }]
}

Event-Scoped Custom Metrics

Step 1: Query the Metadata API Method with your Property ID.

GET https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID/metadata

Step 2: Find the Event-Scoped Custom Metric you are interested in creating reports on from the response. If the metric is not present, you need to register the metric.

"metrics": [
...
    {
      "apiName": "customEvent:credits_spent",
      "uiName": "Credits Spent",
      "description": "An event scoped custom metric for your Analytics property.",
      "type": "TYPE_STANDARD"
    },
...
],

Step 3: Include the custom metric in a report request. The following is a sample request to the runReport method.

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dateRanges": [{ "startDate": "30daysAgo", "endDate": "yesterday" }],
  "dimensions": [{ "name": "eventName" }],
  "metrics": [{ "name": "customEvent:credits_spent" }]
}

Conversion Rate Metrics For One Conversion

Step 1: Query the Metadata API Method with your Property ID.

GET https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID/metadata

Step 2: Find the Conversion Rate Metric For One Conversion you are interested in creating reports on from the response. If the conversion event is not present, you need to set up the conversion event.

"metrics": [
...
    {
      "apiName": "sessionConversionRate:add_to_cart",
      "uiName": "Session conversion rate for add_to_cart",
      "description": "The percentage of sessions in which a specific conversion event was triggered",
    },
...
],

Step 3: Include the conversion rate metric in a report request. The following is a sample request to the runReport method.

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dateRanges": [{ "startDate": "30daysAgo", "endDate": "yesterday" }],
  "dimensions": [{ "name": "eventName" }],
  "metrics": [{ "name": "sessionConversionRate:add_to_cart" }]
}

Event-Scoped Custom Metric Averages

Step 1: Query the Metadata API Method with your Property ID.

GET https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID/metadata

Step 2: Find the Event-Scoped Custom Metric Average you are interested in creating reports on from the response. If the metric is not present, you need to register the metric.

"metrics": [
...
    {
      "apiName": "averageCustomEvent:credits_spent",
      "uiName": "Average Credits Spent",
      "description": "The average of an event scoped custom metric for your Analytics property.",
      "type": "TYPE_STANDARD"
    },
...
],

Step 3: Include the custom metric average in a report request. The following is a sample request to the runReport method.

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dateRanges": [{ "startDate": "2020-11-01", "endDate": "2020-11-10" }],
  "dimensions": [{ "name": "eventName" }],
  "metrics": [{ "name": "averageCustomEvent:credits_spent" }]
}

Cohort Report Examples

Cohort reports creates a time series of user retention for the cohort. For detailed documentation of each API field, see the REST reference for CohortSpec.

Creating a cohort report

Here is a sample cohort report where:

  • The cohort is users with a firstSessionDate of 2020-12-01; this is configured by the cohorts object. The dimensions and metrics in the report response will only be based on the cohort's users.
  • The cohort report will show three columns; this is configured by the dimensions and metrics objects.
    • The dimension cohort is the cohort's name.
    • The dimension cohortNthDay is the number of days since 2020-12-01.
    • The metric cohortActiveUsers is the number of users still active.
  • The cohortsRange object specifies that the report should contain event data starting from 2020-12-01 and ending at 2020-12-06 for this cohort.
    • When a granularity of DAILY is used, the dimension cohortNthDay is recommended for consistency.

The report request for the cohort is:

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dimensions": [{ "name": "cohort" }, { "name": "cohortNthDay" }],
  "metrics": [{ "name": "cohortActiveUsers" }],
  "cohortSpec": {
    "cohorts": [
      {
        "dimension": "firstSessionDate",
        "dateRange": { "startDate": "2020-12-01", "endDate": "2020-12-01" }
      }
    ],
    "cohortsRange": {
      "endOffset": 5,
      "granularity": "DAILY"
    }
  },
}

For this request, an example report response is:

{
  "dimensionHeaders": [
    { "name": "cohort" }, { "name": "cohortNthDay" }
  ],
  "metricHeaders": [
    { "name": "cohortActiveUsers", "type": "TYPE_INTEGER" }
  ],
  "rows": [
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0000" }],
      "metricValues": [{ "value": "293" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0001" }],
      "metricValues": [{ "value": "143" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0002" }],
      "metricValues": [{ "value": "123" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0003" }],
      "metricValues": [{ "value": "92" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0005" }],
      "metricValues": [{ "value": "86" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0004" }],
      "metricValues": [{ "value": "83" }]
    }
  ],
  "metadata": {},
  "rowCount": 6
}

From this report response, a chart for this Cohort report follows. An insight from this report is that the largest drop in active users for this cohort is between the first and second day.

Visualization of cohort users over time

Multiple cohorts and User retention fraction

User acquisition and retention are ways to grow your Website or App. Cohort reports focus on user retention. In this example, the report shows this property has improved its 4 day user retention by 10% over the course of two weeks.

To create this report, we specify three cohorts: the first with a firstSessionDate of 2020-11-02, the second with a firstSessionDate of 2020-11-09, and the third with a firstSessionDate of 2020-11-16. Because the number of users on your property will be different for these three days, we compare the cohort's user retention fraction metric of cohortActiveUsers/cohortTotalUsers rather than use the direct cohortActiveUsers metric.

The report request for these cohorts is:

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dimensions": [{ "name": "cohort" },{ "name": "cohortNthDay" }],
  "metrics": [
    {
      "name": "cohortRetentionFraction",
      "expression": "cohortActiveUsers/cohortTotalUsers"
    }
  ],
  "cohortSpec": {
    "cohorts": [
      {
        "dimension": "firstSessionDate",
        "dateRange": { "startDate": "2020-11-02", "endDate": "2020-11-02" }
      },
      {
        "dimension": "firstSessionDate",
        "dateRange": { "startDate": "2020-11-09", "endDate": "2020-11-09" }
      },
      {
        "dimension": "firstSessionDate",
        "dateRange": { "startDate": "2020-11-16", "endDate": "2020-11-16" }
      }
    ],
    "cohortsRange": {
      "endOffset": 4,
      "granularity": "DAILY"
    }
  },
}

For this request, an example report response is:

{
  "dimensionHeaders": [{ "name": "cohort" },{ "name": "cohortNthDay" }],
  "metricHeaders": [{
      "name": "cohortRetentionFraction",
      "type": "TYPE_FLOAT"
    }
  ],
  "rows": [
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0000" }],
      "metricValues": [{ "value": "1" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_1" },{ "value": "0000" }],
      "metricValues": [{ "value": "1" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_2" },{ "value": "0000" }],
      "metricValues": [{ "value": "1" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_2" },{ "value": "0001" }],
      "metricValues": [{ "value": "0.308" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_1" },{ "value": "0001" }],
      "metricValues": [{ "value": "0.272" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_2" },{ "value": "0002" }],
      "metricValues": [{ "value": "0.257" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0001" }],
      "metricValues": [{ "value": "0.248" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_2" },{ "value": "0003" }],
      "metricValues": [{ "value": "0.235" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_2" },{ "value": "0004" }],
      "metricValues": [{ "value": "0.211" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_1" },{ "value": "0002" }],
      "metricValues": [{ "value": "0.198" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0002" }],
      "metricValues": [{ "value": "0.172" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_1" },{ "value": "0003" }],
      "metricValues": [{ "value": "0.167" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_1" },{ "value": "0004" }],
      "metricValues": [{ "value": "0.155" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0003" }],
      "metricValues": [{ "value": "0.141" }]
    },
    {
      "dimensionValues": [{ "value": "cohort_0" },{ "value": "0004" }],
      "metricValues": [{ "value": "0.118" }]
    }
  ],
  "metadata": {},
  "rowCount": 15
}

From this report response, a chart for this Cohort report follows. An insight from this report is that the 4 day user retention increased by 10% over the course of two weeks. The later cohort with firstSessionDate of 2020-11-16 exceeds the retention of the earlier cohort with firstSessionDate of 2020-11-02.

Chart of multiple cohort retentions

Weekly cohorts and Using cohorts with other API features

To remove day-to-day variance in user behavior, use weekly cohorts. In weekly cohort reports, all users with firstSessionDate in the same week form the cohort. Weeks begin on Sunday and end on Saturday. Also in this report, we are slicing the cohort to compare the users with activity in Russia with users with activity in Mexico. This slicing uses the country dimension and a dimensionFilter to only consider the two countries.

The report request for these cohorts is:

POST https://analyticsdata.googleapis.com/v1beta/properties/GA4_PROPERTY_ID:runReport
{
  "dimensions": [
    { "name": "cohort" },
    { "name": "cohortNthWeek" },
    { "name": "country" }
  ],
  "metrics": [{ "name": "cohortActiveUsers" }],
  "dimensionFilter": {
    "filter": {
      "fieldName": "country",
      "inListFilter": {
        "values": [ "Russia", "Mexico" ]
      }
    }
  },
  "cohortSpec": {
    "cohorts": [
      {
        "dimension": "firstSessionDate",
        "dateRange": {
          "startDate": "2020-10-04",
          "endDate": "2020-10-10"
        }
      }
    ],
    "cohortsRange": {
      "endOffset": 5,
      "granularity": "WEEKLY"
    }
  },
}

For this request, an example report response is:

{
  "dimensionHeaders": [
    { "name": "cohort" },
    { "name": "cohortNthWeek" },
    { "name": "country" }
  ],
  "metricHeaders": [
    { "name": "cohortActiveUsers", "type": "TYPE_INTEGER" }
  ],
  "rows": [
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0000" },{ "value": "Russia" }
      ],
      "metricValues": [{ "value": "105" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0000" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "98" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0001" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "35" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0002" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "24" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0001" },{ "value": "Russia" }
      ],
      "metricValues": [{ "value": "23" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0004" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "17" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0003" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "15" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0005" },{ "value": "Mexico" }
      ],
      "metricValues": [{ "value": "15" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0002" },{ "value": "Russia" }
      ],
      "metricValues": [{ "value": "3" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0003" },{ "value": "Russia" }
      ],
      "metricValues": [{ "value": "1" }]
    },
    {
      "dimensionValues": [
        { "value": "cohort_0" },{ "value": "0004" },{ "value": "Russia" }
      ],
      "metricValues": [{ "value": "1" }]
    }
  ],
  "metadata": {},
  "rowCount": 11
}

From this report response, a chart of this Cohort report follows. Based on this report, this property is doing better at retaining users with activity in Mexico than users with activity in Russia.

Chart of country comparison cohorts