ga4-conversions-to-bq

This repository contains a Python script that uses the Google Analytics Admin API to pull conversion data from Google Analytics 4 (GA4) and load it into BigQuery.

Read my full blog post here for more information.

Prerequisites

1. A Google Analytics 4 (GA4) property.
2. A Google Cloud Platform (GCP) project with billing and the Google Analytics Admin API enabled.
3. A BigQuery dataset to store the conversion data (or the script will create it for you).

Setup

1. Enable the Google Analytics Admin API in the Google Cloud Console.
2. Add the App Engine default service account (<your-project-id>@appspot.gserviceaccount.com) as a user in the GA4 property with the "Viewer" permission (assuming using a Gen 1 function).
3. Add a .env.yaml file to include the following environment variables:

GCP_PROJECT_ID: <your-project-id>
GCP_TABLE_ID: <your-project-id>.<your-dataset-id>.<your-table-id>>
GA4_PROPERTY_LIST: "123456789, 345678901" # Comma-separated list of GA4 property IDs to fetch the conversion metadata from

4. Deploy the main.py script to a Cloud Function using the /clf/deploy-clf.sh script (using Gen 1):

gcloud functions deploy your-function-name --runtime python39 --trigger-http --env-vars-file .env.yaml --region your-region --entry-point main --timeout 540s --ingress-settings all --gen2

5. Schedule the Cloud Function to run at regular intervals via Cloud Scheduler using the /cls/deploy-cls.sh script:

gcloud scheduler jobs create http your-job-name --schedule "0 0 * * *" --uri "https://<your-region-your-project-id>.cloudfunctions.net/<your-function-name>" --http-method GET --time-zone <your-timezone> --location <your-region>

Usage

The Cloud Function will run at the scheduled intervals and pull the conversion data from GA4 and load it into BigQuery. The resulting dataset will contain the following columns:

• date(DATE) - The date of the conversion.
• property_id(STRING) - The GA4 property ID.
• conversion_api_name(STRING) - The API name of the conversion.
• event_name(STRING) - The event name of the conversion.
• custom_event(BOOL) - Whether the conversion is a custom event.
• deletable(BOOL) - Whether the conversion is deletable.
• create_time(TIMESTAMP) - The time the conversion was created.
• counting_method(STRING) - The counting method of the conversion.
• default_conversion_value(FLOAT) - The default conversion value.
• default_conversion_value_currency_code(STRING) - The currency code of the default conversion value.

This table can be joined with your GA4 event data to analyze the performance of your conversions. See the /bq/conversions.sql file for an example query. Please note that the counting_method column affects how the conversion data is counted and should be taken into account when analyzing the data.

Troubleshooting the Initial Setup

1. Make sure to add the App Engine default service account (<your-project-id>@appspot.gserviceaccount.com) as a user in the GA4 property with the "Viewer" permission. This is necessary for the Cloud Function to access the GA4 property.

2. Cloud Functions Gen 1 vs. Gen 2: The code provided in the repository is for Cloud Functions Gen 1. If you want to use Gen 2, you have to adjust the Cloud Scheduler intefration and the Service Account used for the Cloud Function. Gen 2 Cloud Functions use the Compute Engine default service account, whereas Gen 1 Cloud Functions use the App Engine default service account. The code for the Cloud Function itself should work without any changes.