Can I run GA4 Explorations through the API?

No — Explorations are a UI-only feature in GA4. The Data API exposes the same underlying dimensions and metrics, but there's no programmatic equivalent of an Exploration report. Ask AI Data Connector lets you compose the same cross-cutting queries conversationally in Claude (and ChatGPT, Gemini, Perplexity) and joins the results back to Shopify revenue.

Does GA4 have an MCP server?

Not a turnkey one. Google offers the GA4 Data API, but it isn't conversational and doesn't surface the Explorations or attribution-comparison views you actually use. Ask AI Data Connector wires GA4 to Claude through a single connection — no MCP server to host or maintain.

Why is GA4's attribution different from my last-click numbers?

GA4's standard reports default to data-driven attribution, while most ad platforms still report last-click. The dashboard lets you compare models but the Data API makes the comparison clunky to replicate. Ask AI exposes GA4 acquisition data with revenue attached so you can compare patterns side by side without rebuilding the comparison yourself.

Can I combine GA4 sessions with my Shopify orders?

Yes — Ask AI keeps GA4 channel, device and country breakdowns alongside your Shopify orders, so you can ask questions like 'which acquisition channel converts best for repeat customers?' or 'is organic search trending up but converting down?' without exports or manual joins.

Do I need BigQuery to do this?

No. BigQuery export gives you raw event-level data but it's a heavy setup and still doesn't make the data conversational. Ask AI works directly off GA4's Data API — connect with Google OAuth, no SQL, no BigQuery project required.

Get GA4 Channel Attribution & Explorations-Style Reports into Claude

If you migrated from Universal Analytics to GA4, you've probably had this exact frustration: the report you actually need lives inside the Explorations builder, the dashboard makes you click through five menus to assemble it, and the moment you try to automate any of it through the API, the report doesn't exist there. GA4 has the data. What it doesn't have is a way to ask the cross-cutting questions Explorations was built for outside the dashboard — let alone in plain English.

This post is about closing that gap by wiring GA4 into Claude.

The problem with the GA4 Data API on its own

GA4 is more powerful than UA, but the surface area split between the dashboard and the Data API has caught a lot of teams out:

Explorations are UI-only. Path exploration, funnel exploration, segment overlap, attribution comparison — the reports that justify GA4's complexity all live in the dashboard. The Data API has the same underlying dimensions and metrics, but no equivalent "report" object. You can request the data; you can't request the report.
Attribution comparison is awkward to replicate. GA4 lets you switch between attribution models in the UI. Reproducing the same comparison through the Data API means running multiple queries with different model parameters and stitching results back together yourself.
The API isn't conversational. Even when the data is reachable, the Data API expects dimensions, metrics, filters and dateRanges. It doesn't translate "which channel converts best for repeat customers" into a query — you do.
BigQuery export helps, then doesn't. Streaming raw events to BigQuery unlocks event-level analysis, but it's a heavy setup, costs money, and still leaves you writing SQL for every question.

So even teams that connect the Data API directly find it isn't enough on its own. You get numbers, but not the cross-cutting views GA4 was supposed to be for.

What Ask AI pulls in that the raw API doesn't

Ask AI Data Connector syncs GA4 and exposes it to Claude (and ChatGPT, Gemini, Perplexity) through a single connection:

Acquisition data with revenue attached

Sessions, users, conversions and revenue broken down by channel, source, medium and campaign
Joined to your Shopify orders so attribution comparisons line up with real revenue, not just GA4's modelled revenue

Behavioural breakdowns

Engagement, bounce, session duration by device, country, landing page and audience
Ecommerce funnel data — add-to-cart, begin-checkout, purchase — so you can isolate where users actually drop off

Trend data over arbitrary windows

Day-by-day, week-by-week, or rolling 30/90/365-day windows on any of the above, without re-building the date picker in every report

The point isn't to reproduce GA4's UI. It's to let you ask the questions Explorations was supposed to answer — but conversationally, with Shopify revenue joined in.

Why "GA4 in Claude" is different from "GA4 in a dashboard"

Here's the move that's hard inside GA4 itself: cross-referencing.

A few questions Claude can answer once GA4 + Shopify are connected:

"Which acquisition channel converts best for repeat customers this quarter compared to last?"
"Organic search sessions are up 20% — is revenue up the same, or is conversion rate dropping?"
"Compare the top five landing pages by sessions vs. by revenue. Which ones are vanity traffic?"
"Which device drives the highest AOV in the UK vs. the US?"
"Which campaigns brought in customers who later subscribed to my Klaviyo welcome flow?"

These are Explorations-style cross-cuts. No single GA4 report screen answers them cleanly, and none of them survive translation to a single Data API call.

A note on attribution

GA4's attribution model selection — data-driven, paid and organic, position-based — is a dashboard feature. The Data API exposes channel-level data but doesn't let you flip between models in one query. Worth being straight about this: Ask AI doesn't reproduce GA4's model picker. What it does instead is give you the underlying channel- and source-level data alongside actual Shopify revenue, so the model comparison becomes a conversation rather than a setting. If your real question is "is this channel earning its credit?", that conversation is the answer, not the model name.

How to connect GA4

Setup takes a couple of minutes:

Sign up or log in to your Ask AI Data Connector dashboard
Go to Data Sources and find Google Analytics
Sign in with Google and pick the GA4 property you want to connect (you can connect more than one)
Click Sync now — we'll pull your channel, acquisition, behaviour and ecommerce data
Go to API Keys, generate a key, and connect it to Claude

That's it. GA4 is now queryable in plain English, joined to the rest of your Shopify stack.

The bottom line

GA4's data is rich. The gap is access: Explorations are stuck in the dashboard, the Data API isn't conversational, and BigQuery is overkill for most questions. Ask AI fills that gap by exposing GA4 to Claude with Shopify revenue joined in — so the cross-cutting questions Explorations was built for become a chat rather than a click-trail.