Pull Systems authorize work based on downstream consumption rather than upstream forecasts, helping teams control inventory, flow, and overproduction.
Definition
A Pull System is a Lean flow-control method where upstream work is triggered by downstream consumption, customer demand, or a defined replenishment signal. Instead of pushing work forward because capacity is available, teams replenish only what has been used or requested.
Pull systems reduce overproduction, expose flow problems, and create clearer links between demand, inventory, and work authorization.
History
Pull systems are central to the Toyota Production System and Just-In-Time production. Kanban became the most recognized signaling method, but pull logic can be applied in manufacturing, service, software, maintenance, and office work.
When to Use
Use Pull Systems when demand can be translated into repeatable replenishment signals, inventory is excessive, upstream processes overproduce, queues are long, or teams need better WIP control. Pull works best after basic stability, quality, and replenishment rules are established.
Step-by-Step
- Define the product or work family and customer demand.
- Map current flow, queues, replenishment paths, and information signals.
- Set standard containers, lot sizes, reorder points, or WIP limits.
- Choose the pull method: kanban, supermarket, FIFO lane, two-bin, or CONWIP.
- Create clear visual signals and operating rules.
- Pilot with a limited scope and adjust quantities from actual usage.
- Define response rules for shortages, defects, and demand changes.
- Review stability, inventory, lead time, and service level.
Examples
- Assembly: A downstream cell returns a kanban card to trigger replenishment.
- Maintenance: Two-bin spare parts trigger reorder when the first bin is consumed.
- Software: A team limits active work and pulls the next item only when capacity opens.
Common Pitfalls
- Installing cards without changing push behavior.
- Setting quantities without demand and replenishment data.
- No rules for abnormal demand or stockouts.
- Poor quality that causes hidden shortages.
- Too many part numbers in one unmanaged system.
- No ownership for audits and signal accuracy.