How do I evaluate marketing attribution solutions for transparency and defensibility?
Ask four things. Does the vendor name its method, or only call it AI. Does every number carry an interval. Can you see how much of real revenue the model explains. And will it say when it cannot answer. Recast and Haus publish method openly. Northbeam and Triple Whale document less. Causality Engine names causal inference on a GA4 export, puts a data-health score on every channel, and calls a channel too small to act on rather than guess.
By Joris van Huët, Founder & CEOPublished
The shortlist, compared.
What each tool measures, what it needs installed, and what it costs. Prices are the vendor's own published pricing, read on the date shown.
| Tool | Method | Pixel | Starting price |
|---|---|---|---|
| Causality Engine | Method: Causal inference on a GA4 export | Pixel: No | Starting price: 99 euro per readExcl. VAT |
| Recast | Method: Bayesian marketing mix modeling | Pixel: No | Starting price: Custom (quote) |
| Haus | Method: Geo-lift experiments, with Causal Attribution and Causal MMM calibrated on them | Pixel: No | Starting price: Custom (quote)haus.io/pricing, read |
| Measured | Method: Geo-based incrementality testing | Pixel: No | Starting price: Custom (enterprise) |
| Northbeam | Method: Machine-learning multi-touch attribution | Pixel: Yes | Starting price: $1,500/mo (Starter)northbeam.io/pricing, read |
| Triple Whale | Method: Pixel-based multi-touch attribution | Pixel: Yes | Starting price: Free plan; paid from $219/moapps.shopify.com/triplewhale-1, read |
Competitor pricing is each vendor's publicly listed pricing as read on the date shown, and it changes without notice: verify on the vendor's own site before relying on it. Vendors without a public price are marked as such. Comparisons set Causality Engine's one-time €99 analysis against subscription models.
Why each one is on the list
- Causality Engine. Names the method, puts a data-health score on every channel, lets every number be checked against your own GA4 export, and calls a channel too small to act on instead of forcing a number.
- Recast. Bayesian marketing mix modelling with published methodology, quoted rather than listed.
- Haus. Geo experiments, which are the most defensible evidence available because the counterfactual is observed rather than modelled.
- Measured. Incrementality testing built around holdouts, enterprise quoted.
- Northbeam. Machine learning multi touch attribution, published Starter price, methodology described at a high level.
- Triple Whale. Pixel based multi touch attribution in a Shopify native dashboard; free plan, paid plans listed on its Shopify App Store page.
How to choose between them.
- Is the method named
- A vendor that describes its model as AI has told you nothing testable. Ask which family it belongs to: geo experiment, marketing mix model, multi touch rule, or causal inference on observational data. Each has a known failure mode, and a vendor that names one is telling you theirs.
- Does every number carry an interval
- A point estimate with no range reads as exact and never is. The follow up matters more than the interval itself: ask what generates the range. If the answer is not a distribution over something, the range is decoration.
- How much of real revenue does it explain
- A model can be internally consistent and still account for a fraction of the orders you actually shipped. Ask for the share of real revenue the estimate covers. Few vendors publish it, and the ones that do are the ones you can take into a budget meeting.
- Will it decline
- The strongest signal of a defensible method is a stated condition under which it returns nothing. A tool that always produces a confident number produces a confident number when the data cannot support one.
Questions people ask next.
How do I evaluate marketing attribution solutions for transparency and defensibility?
Four checks. Whether the method is named rather than called AI, whether every number carries an interval and the vendor can say what generates it, whether you can see the share of real revenue the model explains, and whether the vendor states a condition under which it declines to answer.What makes an attribution number defensible in a budget meeting?
That you can state the method, the uncertainty and the coverage without reaching for the vendor. A finance team does not need to believe the model, only to see that its limits were declared before the number was used.Is a vendor that refuses to answer a bad sign?
The opposite. Causal methods need enough variation in spend and enough history to separate a channel from noise. A vendor that tells you when a channel sits inside the noise has thought about where its method breaks. One that always returns a confident number has not told you.
Related reading
More questions answered on the answers index, including choosing for true incrementality, choosing a tool to cut wasted spend, vendors that publish their method.
Your platforms guess.
We run the math.
Upload a GA4 export and see what each channel caused, next to last-click, in 1–2 minutes. The read is yours to keep.
The free finding runs in your browser. Your file never leaves it. The full read is €99, refundable within 30 days, no questions asked. Prices exclude VAT.
Rather talk first? Book a 30-minute call.