Your Shopify order data is the lift evidence

Your Shopify order table is the only number in the stack that nobody else can revise. Every ad platform reports a claim about how much of it they caused, and the claims routinely sum to more than the table contains.

That arithmetic is the cheapest audit in marketing and almost nobody runs it.

The audit, in four lines

A ratio above one means the platforms are collectively claiming revenue that does not exist, which is expected rather than scandalous: each platform counts a conversion it touched, and touches overlap. The number is still worth knowing, because it sizes how much of your reported performance is double counting. We wrote it up as the claim ratio and as platform-reported ROAS against orders.

Why the order table is the anchor

It is transactional. It exists because money moved, not because a script fired, so it does not degrade with consent refusal, ad blocking or attribution window settings. It is the same number your accountant uses, which means a marketing report anchored to it is a report finance already trusts the denominator of.

Analytics sits in between: better than platform claims because it is yours, worse than the order table because it depends on collection. That is why the anchoring check, comparing export order counts against Shopify order counts, belongs in every read.

What the order table cannot do

It cannot tell you what caused anything. It records outcomes, not counterfactuals, and no amount of slicing turns an outcome table into a causal claim. That limitation is worth stating clearly, because there is a genre of Shopify app that implies otherwise by showing order data grouped by referral source and calling it attribution.

Grouping by source is a description of where visits came from. It is last-click attribution with a better interface, and its weaknesses are the ones set out in causal versus last-click attribution.

Combining the two

The useful pattern is: order table as the anchor and the denominator, a causal read for the allocation question, and platform claims as an input to neither. That gives you a number you can defend on both sides, because the total is auditable and the split is method-labelled.

On Causality Engine the causal read starts as a €99 one-time upload of a Google Analytics export, showing what each channel caused next to what last-click gave it, with a data-health score and a next step, refundable within 30 days, no questions asked.

The one thing to do this week

Run the claim ratio. It takes twenty minutes, needs no tooling, and it is the number that most reliably changes how a founder reads their ad dashboards afterwards.

Related answers

• The Shopify lift number your agency will accept
• The lift surprises hiding in your own Shopify data
• What a real Shopify channel lift answer requires
• Privacy-first attribution keeps customer data yours
• Own your alert rules, not the vendor defaults