Snap’s SMART cuts LLM costs 90% and lifts ad conversions 27.6%
_reachsumit · x · 2026-07-28
Snap’s SMART paper proposes a hybrid retrieval strategy for dynamic product ads that splits work between BM25 and LLM-driven prospecting.
- Keyword BM25 is used for retargeting known interests.
- LLM-generated queries are used for prospecting new categories.
- A lightweight gate sends only about 10% of users to the LLM path when semantic coverage gaps exist.
- The approach reportedly preserves most of the semantic-prospecting benefit while cutting LLM costs by 90%.
- In a two-week online A/B test, Snap says SMART improved conversion rate by 27.6% over a strong embedding baseline.
The key idea is adaptive routing: use cheap lexical retrieval by default, and call the LLM only where it adds clear value.
More from Companies & People
- HeyNox teases Caspian, an AI assistant that connects email, calendar and apps — avlok · 2026-07-28
- Big companies are hiring again as AI wipeout fears meet reality — jbarbier · 2026-07-28
- Ex-Shopify insider says Tobi Lütke was right to fix a culture problem — morganb · 2026-07-28
- Google, Meta and OpenAI sign open AI coalition letter as Anthropic stays out — FinanceYF5 · 2026-07-28
- AI engineering guide adds 4,894 job posts and real interview patterns across 51 companies — Al_Grigor · 2026-07-28
- A Stanford talk highlights a startup that cut its AI bill from $1.2M to about $100k a month — HeyAmit_ · 2026-07-28