Causality Engine

Disclosure: I founded Causality Engine in 2023 and build it. It is in this stack because I use it, and you should know it is mine before you read further.

What It Does

Causality Engine is causal attribution for ecommerce, run on a Google Analytics 4 export. Upload the export and it shows what each channel caused, next to what last-click reports, in one to two minutes. Its causal inference splits Direct traffic back to the channels that sent those buyers, and separates what each channel caused from what it only touched. It is built for ecommerce owners and in-house marketing leads who split budget across channels and run GA4 on the store.

Why It Matters for Marketing Operations

Every ad platform reports its own contribution, and the claims overlap: add up what Meta Ads, Google Ads and TikTok Ads each say they drove and the total is more than the store sold. Last-click in GA4 errs the other way: it hands the sale to whatever came last, and Direct collects the buyers that other channels sent. A budget decision needs the part each channel caused, not the part it touched, and how sure that number is.

How I Deploy It

In an engagement I run it on the client's GA4 export before any budget moves. The read shows which channels caused sales and which only appeared in the path, next to last-click, with how certain each estimate is. That sets the agenda for the channel conversation, instead of each platform's dashboard arguing its own case. Where the read cannot settle a decision on its own, a controlled test of that channel comes next.

When to Use It vs. Alternatives

Use it when you run an ecommerce store on GA4, spend across several channels, and want a causal read in minutes rather than an implementation project. Today it reads the GA4 export; Shopify, the ad platforms and Klaviyo are listed as coming soon. For an always-on attribution dashboard built on pixel tracking, Triple Whale and Northbeam are the established options. For enterprise measurement programmes, look at Rockerbox, Measured or MMM models. To prove a single channel beyond doubt, run incrementality testing.

The Operator's Take

An attribution tool is only as useful as its honesty about what it does not know. Causality Engine reports what each channel caused and how sure it is. Pricing is public: EUR 99 for one read, EUR 299 a month for Pro, with a 30-day refund. More at causalityengine.ai ↗.