A bottleneck in manufacturing is any resource, process, or workstation that limits the overall throughput of the production system because its capacity is less than the demand placed upon it. Named after the narrow neck of a bottle that restricts liquid flow regardless of how wide the bottle’s body is, a manufacturing bottleneck constrains how much you can produce even if all other resources have excess capacity.

Bottlenecks can be machines, workstations, skilled labour, inspection processes, or even material supply. For example, if you have five production steps and four can process 100 units per hour but the third step can only handle 60 units per hour, that third step is your bottleneck. It doesn’t matter how fast the other steps run because the entire system is limited to 60 units per hour. Work piles up before the bottleneck (creating excess work-in-progress inventory) and resources after it sit idle waiting for work (creating underutilisation).

Identifying bottlenecks is crucial for improvement efforts because, according to the Theory of Constraints, any enhancement to a non-bottleneck process doesn’t improve overall system performance. Only by improving the bottleneck can you increase throughput. This might involve adding capacity at the constraint (buying another machine, adding shifts, or cross-training operators), reducing setup times to increase available capacity, improving quality to reduce rework at the bottleneck, or offloading work to alternative resources.

Bottlenecks can be fixed or floating. A fixed bottleneck remains the constraint regardless of product mix, whilst floating bottlenecks shift depending on what you’re producing. Modern production planning systems help identify bottlenecks through capacity analysis and simulation, showing where queues form and where utilisation exceeds 100%. Real-time monitoring through MES systems can alert managers when bottlenecks develop unexpectedly due to equipment issues or demand changes, allowing quick response to maintain flow and hit delivery commitments.