Why your Shopify ROAS is dropping: a tracking-first diagnostic

A drop in reported ROAS is one of the most stressful signals a DTC operator gets. The instinct is to blame creative, freshen audiences, or pause campaigns. Before any of that, audit your tracking. The drop is often optical, not real, and the recovery is closing the measurement gap.

Six tracking causes of a fake ROAS drop

1. Meta Pixel match rate dropped. iOS update, expired tracking cookie, or a theme change broke the customer-parameter coverage. Match rate fell from 80% to 60% and Meta now counts 25% fewer conversions for the same actual revenue.
2. CAPI stopped firing. A Shopify channel app update or a custom CAPI implementation broke without anyone noticing. Conversions that used to be deduped now show only the Pixel side, and Pixel coverage is lower on iOS.
3. Attribution window changed. Meta's default attribution window or Google Ads' default tightened in a recent update. Conversions that used to be credited now fall outside the window.
4. Refund event misalignment. A spike in refunds gets subtracted from Shopify revenue immediately but ad-platform revenue still shows the original gross. Net ROAS looks worse even though gross is stable.
5. Currency or timezone mismatch. International stores see ROAS swings purely from FX rate changes if Triple Whale or your dashboard is computing in USD while Shopify is in store currency.
6. Event deduplication broke. The same Purchase event now counts twice (once Pixel, once CAPI) because event_id stopped matching across the two sources.

How to diagnose which one is yours

Open Meta Events Manager → Event Match Quality → look at the past 30 days. If you see a step-function drop, that's match rate. Then check Test Events to see if Pixel and CAPI are both firing for the same conversion with matching event_id. Then check Refunds and Returns in Shopify Admin for the same period.

Most of the time one of these six is the culprit, not your creative. Running a tracking audit closes the diagnostic loop in 10 minutes instead of a week of testing creative variations.

When it's actually your creative or audience

If your tracking audit comes back clean (match rate above 85%, CAPI firing, dedup working, no refund spike, attribution window unchanged), then the ROAS drop is real. At that point the work is on the creative or audience side. But run the tracking audit first; it's faster, cheaper, and resolves about 60% of reported-ROAS-drop incidents.