All resources

Measurement · 7 min

Ecommerce ROAS and platform ROAS are not the same metric

Understand the difference between store revenue divided by total ad spend and the conversions attributed inside each advertising platform.

In this guide

Two ROAS figures can both be calculated correctly and still disagree. The useful question is not which number is universally true, but what each number measures and whether its inputs are complete.

Start with the business view

KPIAds defines ecommerce ROAS as store revenue divided by advertising spend for the same project, currency context, and period. It is a blended business view: it does not assign every order to a specific platform.

This view helps an operator answer whether the store generated enough revenue relative to total paid media investment. It should always be read beside coverage, freshness, refunds, and the revenue definition in use.

  • Use the same date range and timezone for revenue and spend.
  • Do not combine incompatible currencies into one total.
  • Check whether revenue is gross, net, or affected by refunds before comparing periods.

Treat platform ROAS as attributed performance

Google Ads, Meta Ads, and TikTok Ads report conversions according to their own attribution settings, identity signals, conversion windows, and modeled data. KPIAds labels this provider-attributed ROAS rather than presenting it as the store's total business result.

Comparing provider-attributed metrics is useful for campaign work. Summing attributed revenue across platforms is not a valid substitute for store revenue because more than one provider may claim the same purchase.

Explain the gap instead of hiding it

A practical report shows both perspectives, names their sources, and makes discrepancies visible. Differences can come from attribution windows, consent, cross-device behavior, refunds, timezone boundaries, missing connector days, or delayed platform data.

KPIAds is designed to keep those meanings separate. The goal is a defensible operating view, not a promise of perfect attribution.