D2C · FMCG & Packaged Foods

Post-iOS 14, Meta signal quality score of 3.1/10. Pixel match rate 31% against the 70%+ benchmark Meta recommends for reliable optimisation. ROAS declining 40% year-over-year despite no change in creative quality or product range, the algorithm was flying blind and allocating budget to audiences with no purchase signal. The Meta account manager had recommended a 40% budget increase as the solution, which would have compounded the problem. Internal team believed the business had simply plateaued; the real cause was signal degradation.

Signal recovery before any campaign or creative change. Built a server-side GTM container on a first-party subdomain to avoid browser tracking restrictions, connected it to Meta Conversions API with deduplication logic against the existing browser pixel. Passed Shopify order data (email, phone, value, line items) as enriched purchase events, server-to-server, bypass iOS 14 restrictions entirely. Simultaneously cleaned the Shopify pixel implementation: removed 3 redundant event tags that were causing duplicate fires. Once signal was restored, rebuilt audience strategy from scratch: CRM first-party data → Meta customer list → LTV-weighted lookalikes replacing interest-based audiences. Rewrote the campaign hierarchy around the newly reliable signal.

Pixel match rate: 31% → 74% within 2 weeks of CAPI launch. Signal quality: 3.1 → 8.7/10. ROAS: fully recovered and +22% above the pre-iOS 14 baseline, the algorithm, now receiving clean purchase signals, found audiences the interest-targeting campaigns had never reached. Cost per result dropped 28% within 30 days of signal recovery, without any creative change. LTV-weighted lookalike audiences from CRM outperforming interest-based targeting by 3.1× on return. The budget increase the account manager recommended was never needed, the problem was signal, not scale.