From Self‑Built BI to Volcano Engine: Challenges, Selection, Operations, and Future Outlook

Introduction – Firefly Thinking, an online education platform with over 200,000 users, migrated its data analytics from a self‑developed BI system to Volcano Engine in 2021, aiming to address performance, scalability, and usability challenges.

1. Pain Points – Limitations of the Self‑Built System The original BI relied heavily on SQL, lacked interactive analysis, suffered from slow query response (P95 > 30 s), had no clear dataset concept, and generated massive, hard‑to‑manage event points, leading to high maintenance costs and poor user experience.

2. Solution Selection – Why Volcano Engine A comparative evaluation of Volcano Engine, Superset, and FanRuan highlighted Volcano Engine’s superior query performance, richer visualization, robust data preprocessing (especially Feishu integration), advanced scheduling, built‑in auxiliary analysis, and a unique dynamic traffic‑splitting AB‑test capability, making it the optimal choice despite a higher price.

3. Operational Strategies – Turning Tool Potential into Business Capability The team shifted from a supply‑driven to a demand‑driven approach, promoted self‑service analytics, improved content production efficiency, and leveraged visual modeling to empower non‑technical users, resulting in higher analyst productivity and broader business adoption.

4. Future Outlook – Data Analysis in the Large‑Model Era Looking ahead, the vision is to move from traditional OLAP dashboards to conversational, AI‑driven question‑answering systems that directly deliver business insights, requiring deeper business‑metadata integration and semantic layer abstraction.

Conclusion The migration to Volcano Engine resolved critical performance bottlenecks, enabled scalable analytics operations, and set the stage for next‑generation AI‑augmented data analysis.