Google Analytics 4 MCP Workflow Setup Tutorial

A comprehensive guide to connect Google Analytics 4 data to AI tools using Model Context Protocol (MCP).

Overview

This tutorial walks you through setting up a workflow to query your Google Analytics 4 (GA4) data through AI assistants using Model Context Protocol (MCP). By the end, asking questions like “Show me user trends over the past month” will provide real GA4 data results.

Prerequisites

Google Analytics 4 property with data collection active
Google Cloud Platform account
Python 3.8+ installed
Basic command line proficiency

Part 1: Google Cloud Platform Setup

1. Create or Select a GCP Project

Go to the Google Cloud Console
Create a new project or select an existing one from the dropdown

2. Enable the Correct APIs

Critical: Enable "Google Analytics Data API" (for GA4)

In GCP Console, go to APIs & Services → Library
Enable Google Analytics Data API
Do NOT use "Google Analytics API" (legacy)

3. Create a Service Account

APIs & Services → Credentials
Create Credentials → Service Account
Name the account (e.g., ga4-mcp-server)
Click “Create and Continue”, skip role assignment, then click “Done”

4. Download Service Account Key

Select your new service account
Go to "Keys" tab, then "Add Key" → "Create New Key"
Choose "JSON" format and download the key
Keep the JSON file secure (you'll need its path)

Part 2: Google Analytics 4 Configuration

5. Add Service Account to GA4

Copy the client_email from your JSON key file
Go to Google Analytics → select GA4 property
Click Admin (gear icon) → Property Access Management
Click "+" → Add users → paste service account email
Select Viewer role → uncheck email notification → Add

6. Get Your GA4 Property ID

In GA4 Admin → Property details, copy your Property ID (numeric only)
Note: This is different from the Measurement ID (starts with G-)

Part 3: Install and Configure MCP Server

7. Install the GA4 MCP Server

Option A: Install via pip (Recommended)

                    pip install google-analytics-mcp
                  

Option B: Install from source

git clone https://github.com/surendranb/google-analytics-mcp.git

cd google-analytics-mcp

python3 -m venv venv

source venv/bin/activate  # On Windows: venv\Scripts\activate

pip install -r requirements.txt

8. Test Your Setup

import os

from google.analytics.data_v1beta import BetaAnalyticsDataClient

# Set credentials path

os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "/path/to/your/service-account-key.json"

# Test connection

try:

    client = BetaAnalyticsDataClient()

    print("✅ GA4 credentials working!")

except Exception as e:

    print(f"❌ Error: {e}")

Part 4: Workflow Scripts Setup

9. Download the Workflow Scripts

Schema Extraction Script
Request Execution Script
Request Validator Script

10. Configure the Scripts

mcp_config = {

    "mcpServers": {

        "ga4-analytics": {

            "command": "ga4-mcp-server",

            "env": {

                "GOOGLE_APPLICATION_CREDENTIALS": "/path/to/your/service-account-key.json",

                "GA4_PROPERTY_ID": "your-property-id"

            }

        }

    }

}

Part 5: Using the Workflow

11. Extract GA4 Schema

# This gets all available tools, metrics categories, dimension categories,

# and parameter information

schema_data = get_complete_ga4_schema()

12. Generate Natural Language Queries

"Show me user growth over the past month"
"What are my top traffic sources by country?"
"Compare bounce rates between mobile and desktop"
"Display daily page views for the last two weeks"

13. Execute and Validate Queries

Convert your natural language query to GA4 API calls
Validate parameters (e.g., check non-empty dimensions)
Execute against your GA4 data
Return structured results

Part 6: Understanding the Results

Data Format Example

{

  "data": [

    { "date": "20240730", "newUsers": "10" }

  ],

  "metadata": {

    "total_rows": 15,

    "returned_rows": 15,

    "applied_optimizations": ["Intelligent sorting applied"]

  }

}

Common Data Limitations:

Only includes data collected since GA4 tracking started
Recent data (24-48 hours) may be incomplete
Large queries may be automatically optimized

Troubleshooting

Issue	Solution
"No module named ga4_mcp_server"	Ensure you're in the correct Python environment with the required package installed
"Dimensions list cannot be empty"	The request validator will auto-fix this by adding a default dimension
"Permission denied" errors	Make sure the service account has "Viewer" role access to your GA4 property
"API not enabled" errors	Double-check "Google Analytics Data API" is enabled

Verification Steps

Test API access with the credential test script
Run a diagnostic script to confirm your data range
Validate queries using the request validator before execution

Advanced Usage

Custom Dimensions & Metrics

All standard GA4 dimensions (date, country, device, etc.)
Custom dimensions and metrics
Custom conversion events

Query Optimization

Automatic data volume warnings and controls
Server-side aggregation for efficient queries
Intelligent sorting for relevance

Security Best Practices

Store your service account JSON file securely
Use the minimum "Viewer" access required
Prefer environment variables for sensitive data
Rotate service account keys regularly

Conclusion

You now have a robust workflow for querying Google Analytics 4 with natural language using AI and MCP. The system simplifies GA4 API usage, empowering intuitive analytics and reporting.

Next steps:

Try different query styles
Integrate with your favorite AI tools
Set up automated reports
Explore advanced GA4 features

For support and advanced details, review the original MCP server documentation and the inspiring Medium article.

Created with Frevana.com