How LeKe Scaled 1,200 Gyms Using Data‑Driven Ops, Agile Testing & AI Recommendations

LeKe, a fast‑growing fitness chain with over 1,200 stores in 24 cities and more than 10 million members, attributes its success to a data‑centric operating model built on Volcano Engine’s data‑flywheel philosophy.

Three Core Capabilities

The interview identifies three key capabilities that turn LeKe’s vision into results:

Fine‑grained operation : leveraging a digital middle‑platform to segment users, analyze their habits, and deliver highly targeted services before they even step into a gym.

Agile testing : using Volcano Engine’s DataTester for continuous A/B experiments, rapidly validating product naming, content‑to‑product ratios, and new feature rollouts.

Personalized recommendation : integrating Volcano Engine’s intelligent recommendation engine with LeKe’s data to serve customized content, courses, and merchandise.

Fine‑grained Operation

LeKe’s “user‑centered” approach starts with detailed user profiling—hundreds of tags covering gender, fitness goals, and preferred activities. The middle‑platform aggregates these tags, enabling stores to pre‑emptively tailor services, improve renewal rates, and boost ancillary sales such as personal‑training sessions.

Agile Testing with DataTester

DataTester has become LeKe’s “digital infrastructure.” By running controlled experiments, LeKe quickly decided between naming options (e.g., “Marketplace” vs. “Mall”), optimal content‑to‑product ratios (30 % proved most effective), and the best promotional tactics during limited‑time campaigns. These experiments cut decision latency and reduced wasted resources.

AI‑driven Personalized Recommendation

Through an API integration, Volcano Engine cleanses LeKe’s data, performs feature engineering, and feeds custom or pre‑built models into a recommendation service. After iterative A/B testing, the recommendation engine was fully deployed, raising community UV‑CTR by 65.1 %, doubling average session duration, and increasing per‑user clicks by 130 %.

Data‑Flywheel Concept

The partnership embodies the “data‑flywheel”: data consumption fuels business decisions, which generate more data, continuously accelerating growth. LeKe’s digital middle‑platform connects headquarters, regions, cities, and stores, ensuring data flows smoothly across the organization and supporting the “hundred‑city, ten‑thousand‑store” expansion without losing management efficiency.

Overall, LeKe’s experience demonstrates how a well‑designed data ecosystem—combining fine‑grained user operation, rapid experiment‑driven iteration, and AI recommendation—can transform a traditional service industry into a scalable, data‑driven business.