Shopify attribution: why every dashboard tells a different story

Ask three dashboards where your Shopify sales came from and you will get three confident, incompatible answers, and the discomfort of that experience is now a permanent feature of running a DTC brand. Attribution is the assigning of credit for a sale to the marketing that supposedly caused it, and the honest version of the subject in 2026 starts with an admission, nobody can do it precisely anymore, including the platforms charging you while claiming otherwise.

The break happened years ago. Apple’s privacy changes cut the tracking thread that once followed a person from ad to purchase, and everything since has been estimation wearing the costume of measurement. A small brand that understands this spends better than a large one that believes its dashboards, which is the entire argument of this piece.

Why the numbers disagree

Each dashboard answers a different question and labels the answer with the same word. Meta reports conversions it believes its ads influenced, on a window it chose, modeled where data is missing, and the model is built by the company selling you the ads. The ad platform grading its own homework is not a scandal, it is the disclosed design, and the grade skews generous everywhere. Google Analytics answers with last-click logic most of the time, handing credit to the final touch, which quietly overpays branded search and underpays the ad that introduced the brand in the first place. Shopify counts orders, the one thing counted correctly in the whole stack, and stays silent on cause.

Add the three reports together and a brand doing 100 orders learns it apparently did 170. The missing 70 were counted twice, or modeled into existence, or credited to a click that happened after the decision was already made.

The expensive mistakes the confusion causes

The first mistake is scaling into the mirage, a channel reports beautiful efficiency, budget doubles, revenue barely moves, because the channel was harvesting demand created elsewhere and its dashboard took the credit. Branded search does this constantly. The second mistake runs the other direction, killing a channel whose true contribution the last-click view could not see, top-of-funnel work that introduced the brand and got paid nothing for it. Prospecting keeps dying this way at brands that then wonder why their retargeting pool dried up.

Both mistakes trace to one belief, that a dashboard number is a measurement rather than a claim. The number is a claim, made by a party with a position, and the sooner it gets read like one, the safer the budget.

What a small brand should actually do

The corporate answer to broken attribution is a data team and incrementality testing, and a small brand has neither, and the good news is the small brand needs less. Three practices cover most of the value.

Trust Shopify for the what, orders, revenue, the numbers that pay rent, and treat every platform report as a directional claim about the why. Add a one-question post-purchase survey, where did you first hear about us, because the customer’s memory is crude and belongs to nobody’s sales pitch, and over a few hundred orders the pattern it draws is more honest than any modeled dashboard. And when a channel’s claim really matters, test it the blunt way, turn it down for two weeks and watch total orders, not the channel’s own report. Total orders barely notice, the channel was taking credit for demand that existed anyway. Orders sag, the channel was real, scale it back up with more confidence than any dashboard ever sold you.

The deeper background on why credit assignment is genuinely hard, the attribution literature covers it, and the short version is that people buy after many touches and no math cleanly splits the credit.

Where the store fits in this

One number in the whole mess belongs entirely to you, and no privacy change can touch it. Conversion rate is measured on your own property, orders divided by sessions, no modeling, no window disputes, no platform grading itself. A brand that lifts its store’s conversion rate has improved every channel’s economics simultaneously, whatever the true channel mix turns out to be, which is the argument our unit economics guide runs in full. Attribution tells you where to spend the next dollar. The store decides what every dollar was worth, and the store is the part you control outright, the whole thesis of the conversion rate guide.

FAQ

Why do Meta and Shopify disagree?

Meta models conversions it believes it influenced, on windows it chose. Shopify counts orders that happened. A claim and a count, and they were never going to match.

What attribution should a small brand use?

Hold Shopify as the source of truth for totals, read platform reports as directional claims, and add a post-purchase survey for the honest rough picture.

Are post-purchase surveys reliable?

Crude and honest. Memory misses touches, but the answer belongs to no platform’s sales pitch, and across hundreds of orders the pattern beats modeled dashboards.

How do I know if a channel actually works?

Turn it down briefly and watch total orders, not the channel’s own report. Demand that survives the pause was never the channel’s to claim.

Did iOS updates really break tracking?

They cut the thread that followed a person from ad to purchase. Everything since is estimation, some of it good, none of it the precision the dashboards imply.

Want the honest read on which of your channels is real, and what your store’s own numbers say? Send me the store and I will walk the math with you. Free look, no obligation, usually a reply within the hour.