The encyclopedia · People & Management · Operational decision · 2019–2025
Micun Binfan's three-system organization scaled from 100 to 1,800 stores
Micun Binfan (米村拌饭) grew from 100 to 1,800+ stores in six years using 三制合一: partnership for capital, apprenticeship for talent, racing for motivation.
Micun Binfan (米村拌饭)
The solution
Micun Binfan (米村拌饭), a rice-bowl chain born in Yanji, is seen as the organizational benchmark of Chinese chain dining because of a structure its backers say is globally unique: 三制合一 — partnership, apprenticeship and horse-race systems working as one.
Partnership solves money: a new store of about ¥1 million gets roughly 85% from an external partner while HQ provides design, supply chain and managed operations, so expansion needs almost no financing. Apprenticeship solves people: every manager must raise apprentices who later become managers, spreading skills like a family. Horse-racing solves motivation: weekly and monthly rankings of revenue, reviews and repurchase decide bonuses and new-store rights.
The combination took the chain from about 100 stores in 2019 to 1,800+ by 2025; stores run 90–150 square meters, mostly in malls, average about ¥27 per ticket, and the closure rate is under 1%. Its 1,800+ stores are managed by 200+ long-term partners.
Why it worked
- External partners supply about 85% of store capital, so HQ stays asset-light.
- Master-apprentice chains replicate managers faster than HR pipelines.
- Public rankings create internal competition without complex KPIs.
- Emotional bonds keep cohesion during explosive growth.
What can be applied
Where capital, training and KPI systems are scarce, substitute social machines: co-investment, master-apprentice replication and public competition — each replaces an institution the market lacks.
Aftermath
By 2024 the system's side effects surfaced: partnership governance grew complex, teacher-school styles diverged, and racing pushed stores toward short-term gaming. In 2025 Micun paused expansion, scrapped store-elimination and high-pressure assessments, and shifted to deep operation, showing the same machines must be rebalanced when growth stops.
Sources
spotted an error? The archive wants to know.