Causality Engine vs Supermetrics: Report the Spend, Read the Cause

The short answer: Supermetrics tells you what each platform reported. Causality Engine tells you what each channel caused. A budget decision needs the second, so Causality Engine is the pick for it. One read of your GA4 export costs €99, and the first finding is free on the start page (prices from the pricing page, checked on 27 October 2026). Keep Supermetrics for the reporting. Its Starter package costs from $44 a month billed yearly, or $55 monthly. It moves 3 data sources into one destination such as Google Sheets or Looker Studio, and into Claude, ChatGPT and Copilot (Supermetrics pricing).

Disclosure: I founded Causality Engine and own it, and this page recommends it. Every claim about it links to Causality Engine's own pages.

The morning sheet is usually where doubt starts. Meta's column, Google's column, TikTok's column, refreshed overnight and perfectly accurate to each platform. Then a total that the bank account does not recognise. Supermetrics did its job. So did you. The columns were never meant to be added together.

Before the next renewal:

1. Price the reporting you already have. Supermetrics Growth billed monthly is, for example, 12 x $222 = $2,664 a year. Billed yearly it is 12 x $177 = $2,124, so the yearly plan saves $540, from the pricing page. That buys data movement. It does not buy a verdict.
2. Total the claims in your own sheet. Sum each platform's conversion value for last month and put the store's revenue beside it. Shopify's help page names the mechanism: when a customer clicks both an email and a Google ad, each platform can count the same sale.
3. Add one column, not one dashboard. A Causality Engine read gives each channel a caused share beside its last-click share, and a health score from 0 to 100 (how it works). Paste those into the same sheet, next to the platform claims.
4. Agree the test before anyone acts. One channel, one region, revenue compared after 14 or 21 days: the windows the sample read on the pricing page prints beside each next step.

Pipe and verdict, compared

Where the jobs divide

A faithful copy of each seller's grade

Supermetrics moves numbers as the platforms report them, which is exactly what a reporting pipeline should do. The catch is upstream. Meta's conversions are Meta's count, Google's are Google's, and both can include the same order. A faster pipe carries the overlap faster.

An agent acts on whichever number it is given

Supermetrics' campaign guide lets Claude or ChatGPT create campaigns and change budgets once you grant write access. There are four approval settings. They run from "No human approval", where changes apply immediately, up to approval of any change to a live campaign. That is a real capability, and it raises the stakes of the input. Tell an agent to scale the channel with the highest platform ROAS, and it will scale whichever platform graded itself most generously.

Seats and destinations, or one answer

Supermetrics charges per package: Starter covers 1 destination, 3 sources, 1 user and 3 accounts per source. Extra users and destinations are priced on top (pricing page). A Causality Engine read covers every channel in the export, with no seat fees (pricing page). The shapes differ because the jobs do. One is a service you use daily. The other is an answer you need a few times a year.

An agent that moves money, with a person in front of it

Here the two tools meet in one workflow. Say the read's next step for paid social is "raise spend a little in one region", the wording of the sample read on Causality Engine's pricing page. In Claude, connected through Supermetrics, ask the assistant to draft that budget change for the one region's campaign. Set the approval level to "Budget + going live" on the Supermetrics Hub. A person then approves before any budget changes on a live campaign, and the change lands in the campaign change history (campaign guide). The human check is the approval itself: confirm the region, the amount and the end date against the read before you press it.

Prices, 27 October 2026

Supermetrics, in US dollars on its pricing page: Starter from $44 a month billed yearly, $55 monthly; Growth from $177 billed yearly, $222 monthly, with 6 sources, 2 users and 12,000 monthly AI credits; Enterprise on request, including data warehousing and the Data API. AI credits for Supermetrics Studio are the one metered element.

Causality Engine, excluding VAT, on its pricing page: €99 per read, refundable within 30 days. Pro is €299 a month on monthly billing, and adds automated GA4 ingestion, a lifetime look-back, a chatbot and budget alerts. Enterprise is a quoted 90-day pilot. That page also puts Supermetrics in its "build it yourself" column, as the data-collection layer. That is the owner's view, and a fair description of the job.

Every figure Supermetrics delivers from the ad platforms it connects was calculated by the platform that sold the clicks. A quicker, cleaner pipe still hands you the seller's own grade. Wire an agent to that pipe and it spends money on the seller's grade, unless something independent sits beside it.

Causality Engine suits

• A store that already reports through Supermetrics and now needs to know which channel caused sales before the next budget meeting.
• Teams about to let an AI agent move budget, who want an independent number in the sheet before they do.
• A quarterly check that costs a read, not another seat.

Supermetrics suits

• Daily and weekly reporting across many platforms into Sheets, Looker Studio, Power BI or a warehouse.
• Agencies building client dashboards, or teams asking Claude about live campaign data.
• Launching or editing campaigns from an AI chat with approval rules and a change log. Causality Engine does none of that.

Background: the Causality Engine glossary entry, the Supermetrics glossary entry, and how to use Supermetrics for the reporting side.

Add the column before the tool

The free version needs only the sheet you already have: platform claims summed, the store's revenue beside them, the difference in red. Then export the GA4 Attribution paths report and drop it on the start page. The preview never uploads the file; it reads it in your browser, per the privacy policy.

For the full column, run the read in Causality Engine, which I founded, from the start page: €99, back in 1 to 2 minutes, every channel's caused share with a next step to test in one region, as how it works sets out. Where spend is too small for the data to separate a channel, the read says so rather than filling the cell.

FAQ

Does Supermetrics do attribution?

Its pricing page lists connectors, destinations, AI chats, Supermetrics Studio, transformations, storage and warehousing, and no attribution model (pricing page). It reports each platform's own attribution as that platform calculates it.

Can Supermetrics send data to Causality Engine?

Neither documents a connection. The read takes the Attribution paths CSV exported from GA4, and Pro ingests GA4 automatically (pricing page).

Is BigQuery included in Supermetrics' published plans?

No. The pricing FAQ says Supermetrics for BigQuery is available in the Enterprise tier, priced through a demo (pricing page).