Bottleneck
A bottleneck is the process point that limits overall flow or throughput because its capacity is lower than the demand placed on it. In supply chain management, this can appear in procurement, manufacturing, warehousing, transportation, planning, or information flow. The bottleneck sets the system’s effective pace, since total output cannot exceed what the slowest or busiest stage can handle.
On the shop floor, a bottleneck is the station or handoff that dictates how fast material can move. It often appears at a machine center, changeover-heavy workcell, inspection gate, or kitting step where cycle time exceeds surrounding process speeds. Work-in-process builds upstream while downstream stations run short, so throughput stalls even though other areas have spare capacity. In warehousing, receiving, putaway, picking, packing, or dock scheduling can become the bottleneck when arrivals outpace labor, scanners, forklifts, or slotting capacity. For raw-material tracking, verification and transaction points such as barcode scanning, lot capture, QA release, or ERP/WMS posting are common constraints; when information flow lags physical flow, inventory visibility becomes stale and replenishment errors follow. The operational signature is persistent queue growth, rising cycle time, and constrained output despite apparent efficiency elsewhere.
How can a bottleneck be identified quantitatively?
Use the station with the highest sustained utilization and the longest queue growth. Standard indicators include throughput, cycle time, WIP, and capacity utilization; the bottleneck is the process step where queue length grows persistently and output constrains overall flow.
Is a bottleneck the same as a constraint?
No. A constraint is any limiting factor, while a bottleneck is the active constraint currently limiting throughput. When the bottleneck is improved or removed, the constraint shifts to the next slowest or most limiting stage.
Can a bottleneck move?
Yes. Once the limiting step is improved, bypassed, or subjected to better scheduling, the bottleneck shifts to the next slowest stage. Therefore bottlenecks must be monitored continuously rather than fixed once.