Ecommerce analytics is where the numbers never match.

Shopify counts server-side completed orders. GA4 counts client-side purchase events, which are blocked by ad blockers (~30%), cookie consent rejections (~40% in EU), cache plugins stripping the dataLayer, duplicate native+GTM tags double-counting, and Shop Pay/Apple Pay bypassing the checkout funnel entirely. 15-40% revenue gaps are routine.

Partially. CAPI sends events server-side, bypassing ad blockers and ITP. But proper setup costs $1,500-5,000 plus $50-200/month for sGTM or Stape. Deduplication between the pixel and CAPI frequently breaks, double-counting or dropping events. Event Match Quality often stalls below 7/10, eroding attribution anyway.

Klaviyo uses a 5-day last-interaction window plus view-through credit. Shopify uses last-click URL referrer. Klaviyo claims revenue from any order where a subscriber opened an email in the last 5 days, even if the purchase was driven by Meta or direct. Combined with Klaviyo flows running alongside paid channels, the same order often gets credited three times. Summed channel ROAS over 100% of actual revenue is the tell.

Baymard's 70% global benchmark gets misapplied constantly. Your real rate depends on traffic quality: research traffic abandons at 85%+, direct/branded traffic at 55-65%. A better question is which step is broken — shipping cost shock (20-30% of loss), forced account creation (15-20%), coupon-field hunt (10-20%), payment friction (10-15%). GA4's Funnel Exploration caps at 10 steps and silently drops users who skip any one, so most teams can't diagnose this reliably.

High-Performance Order Storage (default since WooCommerce 8.2) moves orders from wp_posts/wp_postmeta to dedicated wc_orders tables. Any tracking plugin still reading from the old post-based APIs silently loses all order data. Google for WooCommerce, MonsterInsights eCommerce, and several older GA4 integrations either updated in 2024 or broke entirely. Most store owners discover it when revenue graphs flatline unexpectedly.

Yes. Revenue attribution, session counts, channel grouping, cart/checkout funnels, and engagement metrics all work without cookies — they just require a first-party script that doesn't store persistent identifiers. Cross-session identification (which customer returned?) does need some state, but tab-scoped sessionStorage covers the same session. What you lose without cookies: cross-device identity stitching and multi-day retargeting audiences. What you gain: no consent banner, no GDPR liability, no ITP/ATT signal loss, no consent-mode-v2 complexity. See it in action.