US DTC brands running paid media without an independent measurement layer are optimising to platform-reported ROAS numbers that systematically overcredit the platform doing the reporting

Meta Conversions API and Google Enhanced Conversions both address post-iOS14 signal loss but work differently. Meta CAPI sends conversion events from a server directly to Meta's CAPI endpoint, bypassing the buyer's browser entirely. It requires you to send event data — including customer identifiers like email or phone number for matching — from your server to Meta. Google Enhanced Conversions supplements the existing Google Ads conversion tag by sending hashed first-party data (email, phone, name, address) alongside the conversion event, which Google uses to improve conversion matching accuracy across devices and browsers. CAPI is a server-to-server event pipeline; Enhanced Conversions is a data enrichment layer on top of the existing browser-based conversion tag. Both are necessary for accurate conversion measurement on US campaigns at scale.

Klaviyo and GA4 use different attribution models that will always produce overlapping revenue counts for the same orders. Klaviyo uses a 5-day click, 24-hour open attribution window by default — any purchase within that window by a recipient who clicked or opened an email is attributed to the email. GA4 uses last-click UTM attribution — the same purchase is attributed to whichever channel's UTM parameter was most recently present in the session. The reconciliation methodology is to treat these as separate measurement systems for separate decisions: use Klaviyo revenue for evaluating email program performance (revenue per recipient, flow conversion rate, campaign lift), and use GA4 for channel acquisition decisions (which paid channels are driving new customer sessions). Do not add Klaviyo-attributed revenue and GA4 channel revenue together and compare to Shopify total — the overlap will make the sum exceed actual revenue.

Media mix modelling (MMM) is a statistical method that estimates the revenue contribution of each marketing channel by analysing the relationship between spend levels and revenue outcomes over time, without relying on individual purchase attribution. Last-click attribution assigns 100% of a purchase to the final touchpoint before conversion. MMM looks at aggregate patterns — when Meta spend goes up, does revenue go up? By how much? With what lag? — to estimate incrementality across channels. For US DTC brands, MMM through tools like Northbeam or Triple Whale's attribution model gives you a channel contribution estimate that is not dependent on tracking pixels or cookies, making it more reliable in a post-iOS14 environment where individual-level attribution is increasingly degraded. MMM is a complement to GA4 last-click attribution, not a replacement — GA4 tells you which channels are driving sessions; MMM tells you whether those channels are driving incremental revenue.

The GA4 e-commerce schema uses different event names and a different data layer structure from UA Enhanced E-commerce. In UA, the purchase event was typically pushed to the data layer as ecommerce.purchase with a products array. In GA4, the correct event is named "purchase" and the product data lives in an items array with specific parameter names: item_id, item_name, price, quantity, item_brand, item_category, and item_variant. The migration requires updating the data layer push in your Shopify theme files (or in your GTM custom HTML tag) to use the GA4 items array structure, and updating the GTM GA4 event tags to read from the new data layer variables. Running both structures simultaneously during a transition period is possible but requires careful deduplication to avoid double-counting events in GA4.

Triple Whale, Northbeam, and Rockerbox serve different needs at the $5M+ revenue level. Triple Whale is the most widely adopted among Shopify DTC brands in the $1M-$30M range; it offers a pixel-based attribution model, a summary dashboard (the "Whale Dashboard") consolidating key metrics, and a creative analytics view for Meta ad performance. Northbeam uses a more sophisticated algorithmic attribution model that tends to better handle brands with longer consideration cycles and multi-touch customer journeys; it requires 60-90 days of data to calibrate its model fully. Rockerbox is a cleaner channel-level attribution tool without creative analytics, making it a better fit for brands that want channel-level MMM without the creative performance layer. At $5M revenue with heavy Meta and Google spend, Triple Whale or Northbeam are the two most common choices — Triple Whale for brands that prioritise creative performance visibility, Northbeam for brands that prioritise attribution model accuracy.