Multi-channel attribution after iOS 14, GA4, and consent mode

Three years after iOS 14, two years into GA4, and a year of widespread consent-mode adoption, most teams still trust their attribution dashboards more than they should. The clean, deterministic view of "this ad drove this sale" is gone for good. The teams that measure well in 2026 have stopped chasing perfect attribution and started making confident decisions with imperfect data.

Accept that no single dashboard is truth

Platform-reported conversions, GA4, and your backend will never agree — and that's expected. Each measures a different slice with different rules. The mistake is picking one as gospel. Instead, triangulate: use platform data for in-channel optimisation, GA4 for cross-channel trends, and your own revenue system as the source of truth for what actually happened.

Invest in the measurement stack that survives

• Server-side tracking. Move conversions server-side so signal quality doesn't collapse with browser and cookie restrictions.
• Consent mode, done properly. Modelled conversions recover much of the signal lost when users decline tracking — but only if the tags are configured correctly.
• A holdout or incrementality test. The only way to know a channel's true contribution is to occasionally turn it off and measure the difference.

Make decisions on trends, not decimals

Stop optimising to the second decimal place of ROAS. Look at directional movement over two to four weeks, cross-checked against blended CAC — total spend divided by total new customers. Blended metrics are harder to game and far closer to the truth than any single platform's self-reported number.