Features: Google Analytics MCP for ChatGPT

Core features

1) Marketing source discovery

Use marketing_get_sources to list accessible GA accounts, properties, web/app streams, Google Search Console sites, Bing Webmaster sites, and mapped analysis surfaces.
This makes setup reliable because you can fetch the correct propertyId, appStreamId, measurementId, firebaseAppId, or siteUrl before running reports.

2) Holistic marketing reporting

Use marketing_get_report to run canonical or GA4-native reports with:

• Flexible date ranges (today, yesterday, NdaysAgo, or explicit dates)
• Normalized metric and dimension aliases across GA4, Search Console, and Bing Webmaster
• Filter expressions for precise slicing
• Sorting and row limits for focused analysis
• Optional all-source extraction from mapped sources by default

This is the core feature for daily analysis in ChatGPT: ask, refine, and query again without exports.

3) Funnel reporting for drop-off analysis

Use marketing_get_funnel to analyze conversion funnels with:

• Holistic acquisition-to-conversion bundles across mapped sources
• GA4 ordered multi-step funnel definitions
• Open or standard funnel behavior
• Breakdown dimensions for segment-level visibility
• Next-action and visualization support

This is the fastest way to understand where users drop and which step needs attention first.

4) Analysis surface mapping

Use marketing_manage_surface to get or update mappings for:

• GA4 properties
• Search Console sites
• Bing Webmaster sites
• Multi-selector surfaces

This lets holistic reports pull from the right connected systems without repeating selectors in every prompt.

5) Security-first access model

• OAuth2 authorization flow
• Read-only analytics scope intent
• No write operations to your GA setup
• Zero GA report payload storage messaging
• EU/Germany hosting posture

Example use cases

SEO and organic growth prioritization

Goal: identify pages where traffic is growing but business outcomes are weakening.

Prompt: “For the last 8 weeks, show landing pages with increasing organic sessions but declining conversion rate. Rank top 10 pages by opportunity size and suggest what to test first.”

Landing page optimization queue

Goal: build a practical test backlog for CRO.

Prompt: “List top landing pages by sessions for the last 30 days, including engagement rate and conversion rate. Highlight the 5 weakest pages by expected revenue impact and provide one test hypothesis per page.”

Weekly executive reporting

Goal: produce an action-oriented weekly narrative for leadership.

Prompt: “Create a weekly report: last week vs previous week for sessions, users, conversion rate, and revenue by channel. Summarize key changes, likely causes, and three actions for next week.”

Current CAC review

Goal: understand paid efficiency now, not at month end.

Prompt: “Estimate current CAC for paid channels for the last 30 days by campaign. Return spend, conversions, CAC, and change vs prior 30 days. Flag campaigns above target CAC.”

Worst spending campaign detection

Goal: stop waste quickly.

Prompt: “List the 10 campaigns with highest spend and weakest conversion efficiency over the last 30 days. Classify each campaign as pause, fix, or scale and explain why.”

Funnel bottleneck diagnosis

Goal: find the biggest conversion drop-off.

Prompt: “Run a checkout funnel analysis for last 14 days and previous 14 days. Show drop-off by step and device, and identify the single highest-impact bottleneck.”

Best fit

Safe MCP is strongest when your team needs interactive, in-chat Google Analytics analysis.
If your main requirement is scheduled automation pipelines and cross-app syncing, pair Safe MCP with connector/automation tools.