A Bottleneck, also called a Constraint, is the process step with the lowest capacity in a system. It sets the pace of the whole flow: the system can never produce faster than its bottleneck, no matter how much capacity exists at any other step. Managing bottlenecks is one of the most leveraged activities in operations.
Look for the station with the longest queue in front of it, the highest utilization, and the most overtime. Downstream steps starve; upstream steps overproduce. Cycle time at the bottleneck matches or exceeds takt time, while other stations sit comfortably below it. If you widen the bottleneck and total output increases, the diagnosis was correct.
The Theory of Constraints prescribes five steps: 1. Identify the bottleneck. 2. Exploit it, squeeze every possible unit of capacity from it (no starving, no defective parts wasted on it, no lunch breaks that stop it). 3. Subordinate everything else to it, every other station's pace serves the bottleneck's pace. 4. Elevate it, add capacity if steps 2 and 3 aren't enough. 5. If the bottleneck moves, start again.
Improving a non-bottleneck step is worse than useless, it produces more inventory in front of the bottleneck without any increase in system output. This is one of the most counter-intuitive lessons in operations: efficiency gains that don't happen at the constraint don't gain anything at all.
The bottleneck is where the leverage is. Focus improvement effort there and system output rises. Ignore it and every other improvement is decoration. Every operations leader benefits from being able to identify the constraint in their process at any moment.
