stackgate.ai
Browse workflows

Marketing and growth teams

Ads performance with GA4, Meta Ads, and Search Console

Correlate paid social spend, site conversions, and organic search demand in one marketing performance brief.

Outcome: A weekly ads brief with Meta spend efficiency, GA4 conversion trends, and rising Search Console queries to inform budget shifts

Integrations in this workflow

Google Analytics 4
Meta Ads
Google Search Console

Example questions

Try asking your AI client

  • Build this week's ads performance brief on the private gateway. Pull Meta Ads account insights for the last 7 days, run a GA4 report for sessions and key events over the same period, and list rising Search Console queries for our primary domain. Highlight campaigns with rising CPA, pages where organic clicks grew but paid traffic fell, and recommend where to shift budget - do not change any campaigns.

  • Which Meta Ads campaigns spent more while GA4 conversions fell, and do those landing pages also lose Search Console clicks?

  • Find pages where Search Console clicks rose but Meta Ads CPA got worse. Write a budget-shift note - do not edit campaigns.

  • Compare last week versus this week across Meta Ads spend, GA4 key events, and Search Console impressions for the pricing page.

  • Draft a Friday Slack update that uses all three sources - Meta Ads, GA4, and Search Console - and do not post it.

The problem

Marketing teams review Meta Ads, Google Analytics, and Search Console in separate tabs. Spend decisions miss organic demand signals and on-site conversion context.

Gateway tool bundle

Use namespaced tools on the private gateway or business gateway:

Google Analytics 4 (read-first)

  • ga4__list-account-summaries-tool → ga4__get-property-tool to confirm the property
  • ga4__run-report-tool for sessions, users, and conversion events by date and default channel group
  • ga4__run-realtime-report-tool for a quick sanity check on live traffic

Meta Ads (read-first)

  • meta-ads__list-ad-accounts-tool → meta-ads__get-ad-account-tool for the active account
  • meta-ads__list-campaigns-tool → meta-ads__get-campaign-insights-tool for spend, impressions, and results
  • meta-ads__list-ad-sets-tool → meta-ads__get-ad-set-insights-tool when you need placement-level detail

Search Console (read-first)

  • search-console__list-sites-tool → search-console__get-site-tool to confirm the property
  • search-console__query-search-analytics-tool for clicks, impressions, CTR, and position by query and page

Start with list-services-tool or describe-service-tool to confirm all three integrations are connected.

Token scope recommendations

Private gateway (individual marketers)

  • Scope API tokens to mcp:private-gateway and add mcp:ga4, mcp:meta-ads, and mcp:search-console when you want explicit per-integration limits instead of full gateway access.

Business gateway (shared marketing workspace)

  • Use mcp:business-gateway so the brief respects the company integration allowlist.
  • Bind tokens to a company environment when staging ad accounts must not mix with production reporting.

Least-privilege example

{
  "abilities": [
    "mcp:private-gateway"
  ]
}

Instruct agents to call only list-*, get-*, run-report-tool, query-search-analytics-tool, and *-insights-tool endpoints. Keep Meta create-*, update-*, and delete-* tools out of automated playbooks.

Company integrations checklist

  1. Connect GA4 — OAuth with read access to the property you report on.
  2. Connect Meta Ads — OAuth with ads read permissions for the ad account under review.
  3. Connect Search Console — OAuth scoped to the site that matches your landing pages.
  4. Issue a gateway token — prefer scoped abilities over * full access.
  5. Verify with a read-only prompt - run one of the example questions and confirm audit logs attribute calls to the correct workspace.

Governance

  • Read-only by default: this playbook is for analysis. Do not enable Meta campaign edits or bulk budget changes without explicit human approval per action.
  • Attribution limits: GA4 and Meta use different attribution models; note the reporting window and model in any recommendations shared with stakeholders.
  • Search Console lag: query data is typically 2–3 days behind; align date ranges with GA4 and Meta when comparing week-over-week trends.
  • PII and client data: ad and analytics exports may include customer segments; paste redacted summaries into public Slack threads and review Audit logs before sharing externally.

Connect via private gateway

Activate the integrations above, create an API token scoped to mcp:private-gateway, then add this config to your MCP client.

Config file: ~/.cursor/mcp.json

{
    "mcpServers": {
        "stackgate-private-gateway": {
            "url": "https://stackgate.ai/mcp/private",
            "headers": {
                "Authorization": "Bearer YOUR_API_TOKEN"
            }
        }
    }
}