How to Use GA4 Exports With Causality Engine: Setup Guide

Causality Engine runs on one file: the Attribution paths export from GA4, downloaded as a CSV. No pixel, no integration, no OAuth grant. Upload it and the read comes back in 1 to 2 minutes: what each channel caused, next to what last-click gave it, with Direct split back to the channels that sent those buyers, and a data-health score and a next step for every channel. It is €99 once, excl. VAT, with a full refund within 30 days, no questions asked.

This guide covers what to export, what to check before you upload, and how to read what comes back. It is the same method the how it works page describes.

Step 1: export the Attribution paths report from GA4

1. In GA4, open Advertising, then Attribution paths.
2. Set the date range. Longer reads better: 12 months or more.
3. Pick your purchase key event.
4. Click Share this report, then Download file, then CSV.

It takes about two minutes, and a short video shows the clicks.

Two checks before you close GA4. First, the export needs purchase revenue: pick the purchase key event, not a sign-up or an add-to-cart, or the read has no revenue to split. Second, note how much of the credit sits in Direct and Unassigned. That share is traffic GA4 could not tie to a source. The read splits Direct back to the channels that sent those buyers and marks Unassigned as a tagging fix. The Black Friday number nobody checks explains what to do about it.

Step 2: upload it

Sign up at causalityengine.ai/start and upload the file. No tag in your theme, no DNS change, nothing touches your storefront.

What comes back

On one screen, with a link you can send to your team:

• What each channel caused. Each channel's share of the credit once Direct is split back to the channels that sent those buyers. Direct is not a channel you can buy.
• Last-click, side by side. Last-click rebuilt from your own GA4 export, with Direct left out, next to the engine. Within 25%, the two agree. Under 1% on both is too small to act on.
• Data health on every channel. A score from 0 to 100 that grades the read, not the channel. Under 90, move budget in steps.
• A next step per channel. Raise, cut or hold, or fix tagging first, and how to test it: change spend a little in one region, then compare revenue.
• Your paths to purchase. The common converting paths from GA4, with revenue, key events and days to key event.
• Receipts. Every number checks against your export.

What the read does not do

It installs nothing and tracks no one. It reads the aggregated paths in your export, and the method is causal inference: not a last-click rule, and not an LLM guessing. Pro adds automated GA4 ingestion and lifetime look-back for brands that want the read every week.

Before you pay

The read is €99 once, excl. VAT, for one export. If it is not worth it, ask within 30 days and you get the full €99 back, no questions asked. You keep the read. Pricing has the details.

What to do this week

• If you own the budget: export 12 months of Attribution paths today and run the read. Start with the channels whose next step is raise or cut, and test each one in a single region first.
• If you have to defend the number: open the receipts before the meeting, so every figure you quote checks against the export.

Product facts as stated on causalityengine.ai on 25 September 2026: pricing and refund terms are on the pricing page and change there first.