Google Ads MCP: Pull Campaign Data Without the Google Ads UI

The Friction Layer Nobody Needs

Google Ads reporting is powerful but clunky. You log in, set your date range, filter to the campaign you want, export to CSV, open in a spreadsheet — and that is if nothing goes wrong with the interface loading or your session timing out. For one-off checks during a strategy session or while debugging a campaign that is underperforming, it is a friction layer nobody needs.

The Google Ads MCP server puts campaign data into Claude Code. You ask questions in plain English and get structured answers. "Which campaigns have spent more than $5k this month with a ROAS under 2x?" Then act on what you find — update bids, pause keywords, adjust budgets.

What You Get

• Campaign performance metrics: spend, impressions, clicks, conversions, ROAS
• Keyword and ad group performance data
• Bid adjustments and budget modifications
• Audience and targeting insights
• Conversion tracking status and attribution data

The server uses the Google Ads API v17. Configuration requires a developer token and OAuth credentials with appropriate permissions. You need to enable the API in your Google Cloud project and configure the authentication flow — not a five-minute setup, but well-documented.

Where This Actually Helps

For agencies managing multiple accounts, the MCP server enables programmatic bulk operations. Write Claude Code scripts that iterate across accounts, pulling performance data into a consolidated report, or making budget adjustments based on performance thresholds. Instead of clicking through the Google Ads UI for twenty accounts, you let Claude Code handle the data gathering and present you with decisions to make.

Individual advertisers benefit too. Debugging a campaign that is suddenly tanking? Query recent performance directly from your terminal without opening a browser. The context stays where you are working.