Huya's Data Self‑Service Product: Challenges, Design, and Practice

The presentation begins with an overview of the "data service + self‑service" product concept, which aims to automate code generation through data standardization, lower the barrier for data consumption, and enable ordinary users to obtain data more easily.

It identifies three core issues in existing data services: timeliness (slow response to data requests), flexibility (high analysis threshold and inflexible dimensions), and consistency (inconsistent metric definitions across teams).

To address these, the authors propose three pillars of a good data service: standardized metric definitions, low‑threshold self‑service data production, and diversified self‑service capabilities that cover reporting, data extraction, and data integration.

The MVP of the self‑service product consists of three main functions: custom metric definition, metric subscription (automatic task scheduling), and report generation using existing BI tools. The design also tackles challenges such as reducing code‑heavy data pipelines for non‑technical users and handling data source integration.

Product evolution is guided by four perspectives: cost (optimizing compute resources), efficiency (supporting complex dimensions), data quality monitoring (centralized quality control), and security (unified data permission management).

Results show that the product covers 48% of self‑service data requests with an average latency of 16 minutes, serves 76 users (90% product operators), creates over 2,400 metrics, automates 818 datasets, and supports 117 report consumptions, receiving positive feedback from business users.

The article concludes with a Q&A section addressing metric consistency, applicability to traditional industries, and handling of out‑of‑scope dimensions, emphasizing the product’s scalability and future roadmap.