Google Analytics

@google-analytics list all accounts I have access to and show me their property counts

Found 3 Google Analytics accounts accessible with your credentials: 'Acme Corp Marketing' (accounts/123456) with 5 properties, 'Personal Blog Analytics' (accounts/789012) with 1 property, and 'Client Project Hub' (accounts/345678) with 12 properties. Each account resource name can be used to drill into specific properties or audiences. This illustrative example shows typical account enumeration output—actual counts and names depend on your Google Analytics setup.

This example demonstrates the List Accounts tool, which is typically your first step when exploring GA4 data via MCP. Requires OAuth2 consent with analytics.readonly scope at minimum. If you manage many accounts, response size can grow—consider filtering by display name in follow-up prompts.

@google-analytics create an expanded data set for property properties/987654321 combining session source, device category, and total revenue metrics

Successfully created expanded data set 'Session Source Revenue Analysis' (properties/987654321/expandedDataSets/abc123) combining dimensions [sessionSource, deviceCategory] with metric [totalRevenue]. This dataset will now appear in your GA4 property's data exploration tools and can be queried via the Data API. Note this is an illustrative example—actual dataset ID and availability timing depend on Google's processing queue, typically ready within minutes.

This write-action example shows the Create Expanded Data Set tool, useful for pre-aggregating custom dimension/metric combinations that you query repeatedly. Requires analytics.edit scope. Be mindful of GA4's quota limits on expanded datasets per property (typically 50). Once created, datasets persist until manually deleted.

@google-analytics fetch all audiences for property properties/987654321 and summarize which ones are actively collecting data versus archived

Retrieved 8 audiences from property properties/987654321. Active audiences collecting data: 'High-Value Purchasers' (14,203 users, last updated 2 hours ago), 'Cart Abandoners' (8,891 users, last updated 1 hour ago), 'Newsletter Subscribers' (22,456 users, last updated 30 minutes ago). Archived audiences: 'Q3 Campaign Test' (archived 45 days ago), 'Legacy Mobile Users' (archived 120 days ago). Three audiences are in draft state and not yet collecting. This illustrative summary shows typical audience metadata—actual user counts and update times reflect your property's real-time traffic.

This synthesis example pairs the List Audiences tool with AI reasoning to categorize audience states. Requires analytics.readonly scope. Audience membership counts update hourly in GA4, so expect slight lag between this output and the GA4 UI. Useful for auditing audience health before launching remarketing campaigns.

When this MCP makes sense for small marketing teams

A 3-person marketing team at a SaaS startup runs weekly performance reviews in Switchy, pulling GA4 audience data and account-level metrics into a shared workspace. The Google Analytics MCP works here because the team already uses OAuth2 for their Google Workspace, the 4 tools cover the read-heavy workflow (list accounts, fetch audiences, check expanded data sets), and no one needs to context-switch to the GA4 UI mid-meeting. The trade-off: if your team needs to create custom reports or dimensions on the fly, this MCP only handles expanded data set creation—you'll still open the GA4 console for most configuration work. Use this MCP when your review cadence is predictable and your data structure is stable.

Why this MCP falls short for real-time dashboards

A 6-person customer success team wants to pull GA4 engagement metrics into Switchy to flag at-risk accounts during daily standups. The Google Analytics MCP isn't the right call here. The 4 tools focus on account enumeration and audience listing, not the real-time event queries or funnel analysis a CS team needs to score user health. You can list audiences and confirm which properties exist, but you can't query session duration, conversion events, or user paths—the metrics that actually predict churn. If your workflow depends on querying raw GA4 event data or building dynamic segments, you need the GA4 Reporting API or a dedicated BI tool. This MCP is for account management and audience inspection, not operational analytics.

When this MCP speeds up multi-account setup

A 10-person digital agency onboards 2-3 new clients per month, each with multiple GA4 properties. The Google Analytics MCP fits this workflow: the team uses Switchy to list all accessible accounts, confirm property structure, enumerate existing audiences, and create expanded data sets for standardized reporting—all without opening the GA4 UI for each client. OAuth2 means one auth flow per team member, and the 4 tools map cleanly to the onboarding checklist (verify access, audit audiences, provision custom datasets). The boundary: if a client's GA4 setup is messy (50+ properties, inconsistent naming), the list tools return everything and you'll still need manual triage. Use this MCP when your agency has a repeatable onboarding process and clients grant you admin-level GA4 access upfront.