“If you're a buyer, stop accepting the platform's own grade on its own work.”
Jeff argues that Google search escaped scrutiny because its dashboards looked good. Last-click credit and incremental impact answer different questions, and brand search shows the gap. Buyers should test what changes without the spend, and search must prove it earns its budget.
Jeff welcomes Reddit's new incrementality tools but says the seller shouldn't be the only one grading its own test. He recommends checking any platform's lift report with an independent vendor, using the advertiser's own analytics data.
Jeff doubts that IAB Tech Lab's Trusted Server Framework answers Apple's Safari restrictions, since it still builds a trail on one person. He says the web has moved away from user-level tracking for years, and advises measuring what spend changed in aggregate.
Jeff says Apple's latest crackdown on fingerprinting and identity-stitching continues a decade-long push away from user-level tracking. He argues that advertisers still relying on stitched user paths are exposed and should move to aggregated, incremental measurement.
Jeff argues that moving identity infrastructure server-side is just the latest circumvention of browser and walled-garden limits on user-level tracking. Teams that already moved to aggregate measurement are unaffected, while those rebuilding the old model gain only false confidence.
Jeff argues that server-side frameworks only move the collection point and don't change what marketers are permitted to see. With walled gardens blocking access and most of the journey already untrackable, he says measurement should shift to outcomes and aggregated data.
Jeff argues that Apple's latest announcement finishes off user-level tracking, following years of cookie deprecation and iOS changes. With most of the user journey now untraceable, he says marketers should stop chasing individuals and adopt aggregated data.