Supplement DTC · rebuilt Meta measurement post-iOS 14
Meta's reported ROAS had drifted 2.8× away from the brand's actual revenue. Every monthly board meeting ended in an argument about which number was real. The CFO was pushing to cut paid spend in half.
- Deployed server-side CAPI alongside the pixel, with dedupe logic on Shopify orders
- Built a blended attribution view in the warehouse — Meta, Google, organic, referral — reconciled to actual Shopify revenue nightly
- Ran an 8-week geo-holdout test to measure Meta incrementality directly, not from platform data
The brand wasn’t going to lose money on Meta. It was going to lose the argument about Meta. After iOS 14.5, Meta-reported ROAS on this account was 4.6 — but Shopify revenue only mapped about 1.6× back to Meta spend when measured directly. That 2.8× gap showed up as a $400k/mo “phantom contribution” the CFO couldn’t reconcile.
First fix was plumbing. Pixel-only tracking was bleeding 30–40% of events to Safari privacy and ad blockers. We deployed server-side CAPI with proper deduplication against Shopify order IDs. Attribution coverage climbed from 61% to 94% inside ten days. Reported numbers moved — closer to reality, not further.
The harder work was measurement hygiene. We built a warehouse view that reconciled every paid-media dollar against actual revenue, not platform-reported revenue, with a 7-day and 28-day lookback. The gap between Meta and reality shrank from 2.8× to 0.5× — within the noise band of attribution drift.
Then we ran the test that settled the argument. An 8-week geo-holdout across six US metros measured Meta’s true incremental contribution. The answer: Meta was driving about 72% of what it claimed. Worse than the platform said, better than the CFO feared. With an honest number, the brand unlocked another $1.4M in annual budget that had been frozen in uncertainty.