Case Study: Mobile marketing leader AppsFlyer achieves sub-second, company-wide analytics at billion-row scale with Firebolt

Appsflyer - Customer Case Study

Appsflyer, a leading mobile marketing analytics and attribution platform used by over 12,000 companies, faced serious internal analytics limits: 35 PB of raw data across tables with billions of rows, 1,000 Looker users, and an Amazon Athena-based stack that hit hard limits (≈5 billion rows per query, ~20 concurrent queries), slow interactive query times, and inflexible, costly Spark ETL processes that could take days or weeks to change and rerun. Faced with a new report requiring joins on multi‑billion‑row tables, AppsFlyer decided to replace Athena and evaluate alternatives.

AppsFlyer chose Firebolt’s cloud data warehouse after a POC and migrated schemas and queries; Firebolt enabled joins on 3 billion and 26 billion row tables and let engineers perform ELT in hours using 100% SQL. The result: query performance improved 10x+ from the start (typical large joins now run in 1–2 seconds, earlier smaller joins fell from 40–60s to under 5s), filtering is millisecond‑fast, and costs dropped up to 10x versus prior options. With Firebolt, AppsFlyer’s 1,000 Looker users can run interactive analytics across terabytes and billions of rows in seconds, supporting broader, faster, company‑wide decision‑making.

AppsFlyer

Alexandra Sudilovski

Senior BI Expert & Looker Guild Master