Causality Engine vs Northbeam: One €99 Read or a Measurement Stack

The short answer: spend under $1.5m a year on ads and Northbeam's own pricing page puts you in the Starter band, from a $1,500 starting rate. At that size, run Causality Engine first. €99 for one read of the GA4 export you already have answers which channels caused your sales, in 1 to 2 minutes. The first finding is free at app.causalityengine.ai/start (Causality Engine's pricing page; both checked on 27 October 2026). Northbeam fits a team that wants attribution dashboards every day, view-through credit, a mix model and managed tests in one platform. That team also needs the weeks and the analyst to run it.

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

Northbeam's documentation grades its own methods plainly. Its page on mix modelling sorts the three by what they can prove. Multi-touch attribution is correlational. MMM is correlational and slower to update. Incrementality testing is "the causal ground truth" (Northbeam MMM docs). That is candour, and it deserves the credit. It also means two of the three methods it sells are correlational by its own account. It is the right frame for this page as well.

Do the sums first:

1. What a year costs. For example, twelve months at Starter's $1,500 starting rate is 12 x $1,500 = $18,000. Professional at $3,500 a month is $42,000. Both come from Northbeam's pricing page. A Causality Engine read every quarter comes to €396 for the year on the checkup page.
2. What you already hold. Northbeam's touchpoint data starts when its pixel goes live. Its implementation guide puts a Shopify setup at 2 to 4 weeks, other platforms at 4 to 8. Causality Engine reads GA4 history already recorded, 12 months or more (how it works).
3. Which channel deserves a test. A controlled test costs held-back spend. So it belongs on the channel whose caused share sits furthest from its last-click credit. Finding that gap is what a read does first.

What each one asks of you

Read from each vendor's pages on 27 October 2026.

Where they part company

Proof by experiment, or a model on history

Northbeam's incrementality announcement is dated August 2026. It describes a holdout group withheld from ads, or geography-built groups where a holdout is not possible, with results fed back to calibrate MTA and MMM (announcement). That is the rigorous route. Each test answers one question, at the price of the spend held back.

Causality Engine estimates the counterfactual from the converting paths already in GA4. Its assumptions are on the page: complete GA4 data, enough paths per channel, and no hidden factor moving in lockstep with one channel (how it works). It is a model, so it grades itself. Every channel gets a data-health score from 0 to 100, and a channel under 1% of credit on both methods is called too small to act on. The test settles one question by experiment. The read tells you which question is worth testing.

Weeks of onboarding, or a file you have

Northbeam's guide is candid about the work. UTMs are required. Adding them can reset ads into a learning phase, and the longer they are delayed, the longer implementation takes (implementation guide). Its pricing FAQ adds that your performance will look different from Meta Ads Manager by design. A new set of numbers is the product. Causality Engine works on the record you already keep, so the money spent last quarter is readable this afternoon.

What Northbeam does that a read does not

Clicks + Deterministic Views credits ads people saw and did not click. Apex sends Northbeam's data to the ad platforms' delivery algorithms. Both are on its home page. MMM+ reaches retail revenue that never touches a website. And the Northbeam MCP lets Claude or ChatGPT query your dashboards, read-only. The check that goes with it: ask whether the channel the assistant calls a winner also shows caused lift. The dashboards it reads are the correlational ones.

Prices, as published on 27 October 2026

Northbeam's pricing page lists four plans by ad spend. Growth is custom, for brands under $200K a month on ads, run with an agency partner. Starter is from $1,500, month to month, for Shopify brands under $1.5m a year. Professional is from $3,500 a month for spend up to $500k a month. Enterprise is custom, above that. Incrementality is optional on Professional and Enterprise, and MMM+ on Enterprise.

Causality Engine's pricing page lists €99 for one read, Pro at €299 a month billed monthly, and a custom 90-day Enterprise pilot, all excluding VAT. The read is refundable within 30 days under its terms.

The FAQ answer on "start at"

A starting rate reads like a price. Northbeam's FAQ says the monthly price is informed by your pageviews, and is an initial estimate until it has learned more about your business. Professional and Enterprise run on annual terms. So the default after a demo is a year committed against an estimate of your traffic. Ask for the pageview basis in writing.

Causality Engine's own comparison with Northbeam is the view of the company I own, so weigh it as a vendor page. It suggests the same order as this one: run the read before signing or renewing.

Choose Causality Engine if your job is

• Deciding next quarter's split from GA4 history before a platform decision has been made.
• Finding the one channel that deserves an incrementality test.
• Getting a causal answer at a spend level where a platform subscription is hard to defend against the budget it measures.

Choose Northbeam if your job is

• Watching attribution daily across channels, with view-through credit for video and reach campaigns.
• Running controlled holdout and geo tests and calibrating MTA and MMM with the results.
• Working with a media strategist, and sending first-party data back to the ad platforms through Apex.

For the method in more depth, the Causality Engine glossary entry and enterprise marketing attribution cover the models.

Start with the free calculator

One free check works with no tool at all. Enter your average order value and gross margin in Causality Engine's break-even ROAS calculator. Compare the result with your MER, total revenue over total ad spend, and not with any platform's ROAS, as the calculator itself advises. If MER sits below break-even, the budget question is already on the table, whichever tool you buy. Then drop the GA4 export on the start page for the free first finding.

If the numbers say a channel decision is due, Causality Engine, which I founded, runs the full read for €99 at app.causalityengine.ai/start. It returns each channel's caused share, its last-click credit, a data-health score and a next step; the method is on how it works. The case studies report that Me Gorgeous's first read found roughly 8K EUR a month of Google spend buying conversions that would have happened anyway. That is the company's own published result, not a forecast for your store. Pro, at €299 a month, runs the same model continuously.

FAQ

Can Causality Engine replace Northbeam?

For the question of where the next budget goes, often. For daily dashboards, view-through credit, MMM+ or managed tests, no. The two can run together: Causality Engine installs nothing and reads a GA4 export.

Does Northbeam need UTMs on every ad?

Yes. Its implementation guide calls them required and says delays extend implementation. Causality Engine reads whatever GA4 recorded. Traffic with no tracked source shows as unassigned, with the advice to tag every link and run a new read, per the sample on its pricing page.

Is $1,500 Northbeam's monthly price?

It is Starter's listed starting rate. The pricing FAQ says Starter is month-to-month, usage-based pricing for Shopify brands, informed by pageviews. Exact quotes come from a call.