Why your ad platform and your analytics disagree

Every month someone discovers that Meta claims 212 conversions, GA4 shows 143, and Stripe settled 168 payments. The reaction is usually an audit hunting for the broken tag. Fair enough. But the disagreement is structural, and it survives a perfect implementation.

Attribution windows do the first damage. Meta counts a purchase up to seven days after a click and one day after a view by default. Google Ads uses windows up to 30 days. Your analytics tool counts the session it saw. Three defensible definitions, three numbers, one purchase.

Reporting dates split them further. Google Ads books a conversion against the day of the click, so a Tuesday click that converts on Friday changes last week's report retroactively. Analytics books it on Friday. Month-end totals drift apart with nobody wrong.

Then the platforms stopped seeing. After Apple shipped App Tracking Transparency in April 2021, the analytics firm Flurry measured US opt-in rates in the single digits during the first weeks. Meta and Google responded with modeled conversions, which are estimates, published to two decimal places. Ad blockers take a further 15 to 30 percent of web tracking depending on the audience, and a developer-heavy audience sits at the top of that range.

Cross-device journeys blur what remains. The phone clicks, the laptop buys, and every tool with a partial view invents its own answer about who deserves the credit.

The workable posture is boring, and we have written ours down in full on how we measure. Pick one source of truth per decision. Platform numbers work for comparing campaigns inside that platform, since the bias is at least consistent. Payment data answers whether the business made money. Reconcile the two monthly, expect a gap of 15 to 25 percent, and write the expected gap down. The month it doubles, you will notice a real breakage instead of relitigating definitions from scratch.