iOS 18, Advantage+ and the Meta attribution reset
What changed in reporting this summer and how we re-baselined our cost-per-booked-call benchmarks.
Two things converged in summer 2024 that broke the previous year's Meta benchmarks for bridging accounts. iOS 18 brought another round of attribution tightening, and Meta's continued push of Advantage+ as the default campaign structure changed both the reporting and the optimisation surface. Any benchmark older than June 2024 was, by August, basically unreliable.
What iOS 18 actually changed
Less than the headlines suggested, but enough to matter for low-volume finance accounts. The aggregated event measurement window shifted, the attribution priority logic changed, and the practical effect was another 8-12% of conversion signal disappearing from in-platform reporting.
For a bridging account doing 40-80 booked calls a month, that was the difference between knowing what was working and guessing.
Advantage+ as a forcing function
Meta's nudge toward Advantage+ campaign structures became hard to ignore through Q2 and Q3. Manual campaign performance degraded relative to Advantage+ on the same audiences, partly because Meta was directing more bidder attention to Advantage+ and partly because the ML genuinely did better on cold prospecting.
We migrated most cold acquisition to Advantage+ through summer 2024 and kept manual structures for warm retargeting and brand. That split has held.
How we re-baselined
Threw out every cost-per-booked-call benchmark older than June and rebuilt them on August data with the new attribution and the new campaign structure. Set a 90-day rolling baseline rather than a year-on-year comparison, because year-on-year was comparing two different systems.
The new benchmarks were roughly 18% higher than the old ones, which we initially read as channel degradation. Six months of data confirmed it was mostly the measurement system catching up to reality.
Summer 2024 was the third or fourth meaningful attribution reset on Meta in as many years. The brokers who handled it well were the ones who treated their benchmarks as living documents, rebased them on every major platform change, and stopped trying to reconcile current data with pre-reset history. The benchmarks before iOS 18 simply did not describe the channel after iOS 18.
Capital Edge publishes one note a month on UK bridging finance, paid acquisition, and AI-led outbound. Written for brokers, by the team running the playbook.