{"type":"post","id":"avito-recommendation-architecture-analysis-5ad956","slug":"avito-recommendation-architecture-analysis-5ad956","title":"AvitoTech reports that it redesigned the collaborative recommendation path for Avito’s home page because a","description":"The material is not merely a story about speeding up model training; it shows that recommendation freshness can be an end-to-end systems constraint when inventory has a short commercial life.","retrieval_nugget":"The material is not merely a story about speeding up model training; it shows that recommendation freshness can be an end-to-end systems constraint when inventory has a short commercial life.","published_at":"2026-09-25","updated_at":"2026-09-25","record_date":"2026-09-25","date_kind":"published_at","topics":[],"source_urls":["https://clc.to/ivJY6w?erid=2VtzqvQcHGx"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:avito-recommendation-architecture-analysis-5ad956","status":"published","visuals":[],"editorial_provenance":{"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:7cb1ebf35d16e01bccdb5375bb75efb3257095a2517178702f1da011429da3aa","reviewed_at":"2026-09-25T16:13:54.298Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0},"analysis":{"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"The material is not merely a story about speeding up model training; it shows that recommendation freshness can be an end-to-end systems constraint when inventory has a short commercial life.","observed_facts":[{"text":"AvitoTech reports that its collaborative model previously refreshed every six hours, while its new near-real-time path processes clickstream in 20-minute windows.","source_urls":["https://clc.to/ivJY6w?erid=2VtzqvQcHGx"]},{"text":"AvitoTech reports 240 million active listings and says that recommendations on the home page drive roughly half of views and about 30% of seller contacts.","source_urls":["https://clc.to/ivJY6w?erid=2VtzqvQcHGx"]},{"text":"AvitoTech reports a 6% Recall@K gain and a 1% increase in buyer–seller contacts in an A/B test.","source_urls":["https://clc.to/ivJY6w?erid=2VtzqvQcHGx"]}],"mechanism":"The mechanism is selective state mutation rather than repeated full retraining: a baseline model supplies initial embeddings, fresh event windows identify affected users and items, and workers update and publish only the relevant state.","why_now":"Why now: in marketplaces, a listing’s highest engagement may occur in its first hours, so a six-hour lag can remove commercially valuable inventory from the recommendation opportunity.","implications":["The implication is that teams should measure model freshness against the half-life of the underlying inventory, not only against offline ranking metrics.","An incremental model can create hidden operational debt if it runs alongside the batch model; consolidation may matter as much as the initial quality lift."],"evidence_boundary":"The evidence boundary is important: the reported Recall@K, A/B result, memory reduction, and architecture details are AvitoTech’s own account, and the supplied source does not provide experiment design, confidence intervals, absolute contact volumes, or independent replication.","watch_conditions":["What to watch is whether Avito reports how the 20-minute system performs across its category-by-region partitions, during queue backlogs or failures, and after changes in listing mix or user behavior."],"related_records":[],"new_branch_reason":"The supplied related references provide identifiers and match labels but no prior New Runtime record content; this source therefore starts a distinct analytical record centered on Avito’s documented near-real-time recommendation architecture."},"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/avito-recommendation-architecture-analysis-5ad956/","markdown":"https://newruntime.com/posts/avito-recommendation-architecture-analysis-5ad956.md","json":"https://newruntime.com/posts/avito-recommendation-architecture-analysis-5ad956.json"},"next_reads":[{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}