About Yili
Yili Group is China’s largest dairy company by scale and breadth of range, Asia’s number-one dairy company for nine consecutive years, and one of the world’s top five. Its portfolio spans liquid milk, milk powder, yogurt and ice cream — produced across dozens of plants and distributed nationwide.
The challenge
Yili wanted true integrated planning: sales, production, logistics and procurement working from one shared view of the supply chain, with the ability to adjust production, material and transport plans quickly as the market moved.
At the same time, it wanted a better network: optimized line-haul routes to cut transport time and cost, data-driven decisions on where to open new warehouses, and a distribution structure where every node worked together efficiently.
Underneath all of that was a non-negotiable requirement: usability. Planners needed an interface that made building and adjusting plans fast and accurate — and flexible enough to fit very different business units and scenarios.
The solution
One platform for integrated planning
Convect deployed its planning platform at Yili, bringing sales, production, logistics and procurement onto a single platform. Every function now works from the same real-time data and responds to one unified supply chain plan — instead of reconciling separate spreadsheets in endless meetings.
More importantly, the plan no longer depends on individual planners’ rules of thumb. It is computed by optimization against Yili’s actual constraints, so it meets customer demand at the lowest total cost — and can be recomputed quickly when conditions change.
Network optimization
Using the same model, Convect helped Yili optimize warehouse locations to balance cost against responsiveness, redesign line-haul routes to remove unnecessary distance and time, and rebalance the distribution network so supply in each region matches demand — reducing both excess stock and shortages.
An optimization engine built for scale
The platform’s embedded network optimization engine evaluates and compares tens of millions of plan options in minutes, which is what makes it practical to re-plan as often as the business needs to.
Fast to learn, fast to deploy
Planners compare scenarios side by side in an intuitive interface. New users go from zero to proficient in about a week, and new planning applications can go live within a month.
Results
- +5% total capacity — the equivalent of more than half a new plant, with no new capital investment.
- 17% lower line-haul transport cost — savings worth millions of dollars a year from optimized warehouses and routes.
- Master planning cut from 5 days to under half a day. What used to take several departments the better part of a week now takes a morning.
The planning cycle went from five days to half a day. That speed gives Yili far more room to respond to market changes — and takes a heavy, repetitive workload off the planning team.
What’s next: agents on top
With the ontology and optimizer in place, the same foundation now supports Convect’s planning agents — which monitor demand and supply signals continuously, replan around disruptions automatically, and bring planners recommendations with the reasoning attached.
