Meta's Generative Ads Recommendation Model (GEM) delivers 5% ad conversion increase on Instagram and 3% on Facebook Feed
Meta's ads recommendation system needed to handle billions of sparse user-ad interactions across diverse surfaces, process a heterogeneous mix of advertiser and user data, and train large-scale foundation models efficiently at scale.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · User-ad interaction data ingestion
Billions of daily user-ad interactions across Meta's platforms provide sparse training signals such as clicks and conversions.
GEM delivered a 5% increase in ad conversions on Instagram and a 3% increase on Facebook Feed in Q2, with Q3 architecture improvements doubling the performance benefit per unit of data and compute, and the new architecture achieving 4x efficiency over the previous generation of ranking models.
What failed first
Traditional architectures struggled to process long user behavior sequences efficiently, existing approaches risked losing critical engagement signals by compressing sequences into compact vectors, and legacy recommendation systems failed to balance cross-platform learning with surface-specific optimization.