01 / 04 · Noise
Meta says 4.2x ROAS. Google says 3.8x. Shopify says something else entirely.
Which number do you tell your CFO?
Causality Engine.
You spend money on marketing. Some of it works.
We tell you which parts, and how sure we are.
Refundable if the first read does not move a budget decision.
Or book a 30-min call.
Scroll to remove the noise
Before you rip anything out
Nothing to replace. It is a layer, not a migration.
Nothing to migrate
Triple Whale, GA4, Shopify and Klaviyo keep running.
€99 to test it
Your own export, 1 to 2 minutes, refundable.
Answers the allocation question
Where the next euro goes, at any budget.
Lands where you work
Exports, plus API keys and MCP on Pro.
If your spend is too small for the model to find a signal, we say so and decline. Two of five published cases are exactly that.
- Shopify
- WooCommerce
- Meta Ads
- Google Ads
- Amazon Ads
- eBay Ads
- TikTok Ads
- Klaviyo
- Awin
- CJ Affiliate
- Claude Commerce
- and more
Four Dutch DTC brands and our own store. Real engagements, numbers they let us publish.
No anonymous logos, no inflated brand counts. Four brands ran a real causal read on real GA4 data and signed off on the figures and quotes on the case-studies page; two of them also tell you when we said the product wasn't the right fit yet. The fifth, Maison Tanger, is ours: we run it on the model's output alone, to live the day-to-day our users live rather than just measure it.
Me Gorgeous
megorgeous.nl
+30% revenue lift, 2025
Causal read flagged 8K EUR/mo of Google spend with no incremental contribution.
The Two Sisters
the-twosisters.nl
+60% revenue lift, 2024
Last-click said cut Pinterest. Causal read kept it live.
Twinkels
twinkels.nl
+60% revenue lift, 2024
Hidden Meta return-rate gap surfaced by causal lift, not gross ROAS.
Maison Tanger
maisontanger.com
Our own store
Founder-owned. We run it on the causal read alone, to live the day-to-day our users live.
OFFFTRACK
offftrack.com
Honest fit floor
Causal gap inside the confidence interval at current spend. Revisit when scale lifts.
The offer
Two ways in. Both start with your GA4 export.
No annual contract, no pixel, no onboarding call. A one-time causal read for the price of a dinner, or the same model running continuously.
One read
Start hereThe fastest way to see whether causal moves a decision.
- Runs alongside your current stack, nothing to migrate
- Per-channel incremental ROAS with confidence intervals
- Platform-reported vs causal comparison for every channel
- Budget reallocation recommendations
- Exportable report
- 14-day Pro trial included
- Full refund if it doesn't move a decision
Pro
Continuous causal attribution, ingestion, chatbot, API.
- Automated GA4 ingestion (no more manual upload)
- Insights that compound month over month
- Chatbot for ad-hoc questions of your attribution data
- Developer API so marketing agents can pull insights
- MCP server, so the model works inside agent and LLM workflows
- Same causal model, run continuously
- Cancel any time
Choose €99 for a one-shot answer to one question. Choose Pro once you want the model watching every month. Full comparison on the pricing page.
This is what €99 buys. The report, not a dashboard tour.
One view per channel: what the platforms claimed, what actually caused revenue, and how sure the model is. Exportable, CFO-ready, yours to keep.
116%
of revenue, claimed by the ad platforms together.
It only happened once.
- Paid search31.2 → 14.9% (95% interval 13.3 to 16.5%)−16.3 points over-claimed
- Paid shopping22.7 → 10.2% (95% interval 8.9 to 11.5%)−12.5 points over-claimed
- Cross-network16.1 → 7.3% (95% interval 6.2 to 8.4%)−8.8 points over-claimed
- Email12.4 → 4.8% (95% interval 4.0 to 5.6%)−7.6 points over-claimed
- Referral5.8 → 2.1% (95% interval 1.6 to 2.6%)−3.7 points over-claimed
- Organic search9.6 → 19.0% (95% interval 17.2 to 20.8%)+9.4 points under-credited
- Direct18.4 → 39.6% (95% interval 37.2 to 42.0%)+21.2 points under-credited
Sample numbers for illustration. Your report shows your own channels, from your GA4 export.
Process
Here's exactly what happens. All four steps, before lunch.
The Q3 budget you already spent is auditable today: the model runs on GA4 history, so there is no ramp-up period and no model-training wait.
- 01~2 min
Export
Download any historical period from GA4. No pixel, no SDK, no access grants, nothing touches your stack.
- 02~1 min
Upload
Drop the CSV in and pay €99 by card. No contract, no sales call, full refund if it doesn't move a decision.
- 031–2 min
The model runs
Proprietary causal inference separates incremental revenue from would-have-happened-anyway, per channel.
- 04same day
Decide
Per-channel causal answers with confidence intervals and reallocation recommendations. Move budget with evidence.
Five checks before you spend €99.
If one fails, don't.
You run GA4.
The export is the only input. No pixel, no install, no developer.
You run four or more channels at once.
Meta, Google, TikTok, email, affiliate. Each one grades its own homework.
Every platform says it drove the sale.
Meta says scale Meta. Google says scale Google. Nobody says cut.
Your platforms claim more revenue than you made.
Add up what they report and put it next to your orders. It is more than you sold.
A budget decision is waiting on the answer.
Not a dashboard to admire. A call to make this month, and a number to defend.
All five true? Run the first read. If it doesn't move a budget decision, the €99 comes back.
And one of two people is reading this.
You own the brand and the budget.
You decide alone, and you want the answer before the next spend. Upload a GA4 export, get each channel's incremental ROAS with confidence intervals in 1 to 2 minutes, and act on it. €99, refunded if it does not move a budget decision.
The refund terms are in the pricing FAQ.
You run the channels. Someone else signs off.
You have to defend a number to the owner or to finance. Walk through the model on a real store's data in a 30-min call, then take the read in with its intervals, the method in plain language, and the four answers your CFO will ask for.
The four CFO answers are on the pricing page.
What €99 actually replaces
Year-1 cost, itemised.
The platforms publish a license number. The real bill includes onboarding, engineering hours, and seat fees you only notice in quarter two. This is the side-by-side.
- Tool license
Causality Engine
€99 per causal read, or €299/mo Pro
Platform alternative
€100-300/mo published list (often higher tier)
- Contract
Causality Engine
Credit card, cancel by closing the tab
Platform alternative
Annual terms common above the entry tier
- Onboarding / kickoff
Causality Engine
CSV upload, 1-2 minutes
Platform alternative
2-6 weeks of analyst time
- Pixel or SDK install
Causality Engine
None
Platform alternative
Required for most platforms (5-15 dev hours)
- Per-seat fees
Causality Engine
Unlimited team access included
Platform alternative
€30-100 per added seat per month
- Methodology document
Causality Engine
On request, no charge
Platform alternative
Often a paid engagement
- Quarterly model review
Causality Engine
Included
Platform alternative
Often a paid engagement
- First-decision guarantee
Causality Engine
Refund if the first read does not pay for itself in one reallocation
Platform alternative
No refund
- Time before first usable number
Causality Engine
1-2 minutes from upload
Platform alternative
30-90 days post-kickoff is industry-standard
Causality Engine
€99 (one-time) · €3,588 (Pro, 12 months)
Platform alternative
€7,000-18,000 typical, license + onboarding + engineering
Platform numbers reflect published pricing tiers and industry-standard onboarding hours, not a quote from a specific vendor. Your mileage varies by team size and integration scope.
Your platforms guess.We run the math.
Last-click is lying to you. Run the model on your own GA4 export, €99, once.
Full refund if you don't see the value.
Or book a 30-min call.
You have questions.
We have math.
What is marketing debt?
Marketing debt is the compounding cost of budget decisions made on wrong attribution. Each quarter you allocate spend on correlated numbers instead of causal evidence, the misallocation carries into the next plan and grows. Like technical debt, but on the marketing P&L. A causal read on your GA4 export quantifies it per channel.
The full breakdown of marketing debtWhat is attribution debt?
Attribution debt is the gap between what your ad platforms claim drove revenue and what actually caused it, carried quarter after quarter into the budget. It is how marketing debt accrues: allocate on claimed conversions long enough and the plan itself becomes the liability.
Both terms, definedI'm using GA4 and I don't trust the attribution numbers. What do I do?
Upload any historical period from your GA4 export. Our proprietary causal-inference model runs on it and within 1 to 2 minutes returns a per-channel view of which campaigns actually drove incremental revenue, not just which ones happened to be there at the click. €99, pay-per-use. No subscription, no setup call. Go Pro at €299/mo to unlock the cool stuff: automated GA4 ingestion (no more manual upload), insights that compound on each other over time, a chatbot for asking questions of your data, and a developer API so your marketing agents can pull insights in real time. Works for any ecommerce vertical, fashion and beauty brands were our launch users, but the model is vertical-agnostic. GA4 is enough today; the native Shopify integration is on the roadmap.
What €99 and Pro includeWait, what even is attribution?
Attribution is how you figure out which bits of your marketing actually led to a sale. The ads, yes, but also the emails, the Instagram posts, the people who found you on Google. Most tools just credit whatever someone clicked last, even if five other things got them there first. It's like giving all the credit to the waiter and ignoring the chef.
Causal attribution, explainedIs this only for ads?
No. Ads are one part of it. Your Instagram posts, the people who find you on Google, your emails, they all push someone toward buying, and they all go into the same model. Most people say ROAS when they mean ads. We look at the whole thing and tell you how the parts fit together.
Incremental marketing, not just paid ROASWhy should I care if my attribution is wrong?
Because you're making spending decisions based on lies. If your dashboard says Meta is your best channel but it's actually TikTok creating the demand, you'll cut the thing that's actually working. Wrong attribution means wasted budget. Every single month.
Why last-click is lying to youMy Shopify dashboard shows me sales. Isn't that enough?
Shopify tells you what sold. It doesn't tell you why it sold. Was it the Instagram ad they saw Tuesday? The Google search on Thursday? The email on Saturday? Shopify just sees the finish line. We show you the whole race.
The Shopify attribution guideI'm not technical. Can I actually use this?
If you can export a CSV from GA4, you can use Causality Engine. No code. No pixels. No 90-day onboarding. Upload the file, wait 1–2 minutes, see which channels actually cause sales. Native Shopify integration is on the roadmap but GA4 alone is enough today.
How it works, step by stepHow is this different from what Meta and Google already tell me?
Meta says it drove the sale. Google says it drove the sale. TikTok says it drove the sale. They all take credit for the same purchase. We're the only ones who don't have an ad budget to protect. So we tell you the truth.
How the attribution tools compare