The systematic process of identifying, quantifying and characterising the resource(s), operation(s) or constraint(s) within a workflow whose capacity or variability limits the overall system throughput or performance; it distinguishes between capacity, utilization, variability and policy constraints and identifies where interventions will change system output.

Principle

A system’s achievable throughput is determined by its most constraining stage(s); therefore interventions that increase capacity or reduce variability at the true bottleneck change throughput, while equivalent efforts on non‑bottlenecks generally yield little system‑level benefit.

Demonstration

Illustrative scenario → Situation: a five‑station line shows work‑in‑process accumulation upstream of station 3. Recognition: cycle‑time measurement shows station 3 average cycle exceeds takt and has high variance. Action: apply targeted countermeasures (tool maintenance, staffing adjustment or buffering) at station 3. Consequence: upstream WIP declines and system throughput increases until another constraint appears.

Misapplication

Optimising the busiest or most utilised resource without confirming it constrains throughput (e.g., improving a downstream machine that is not on the critical path); the semantic error is conflating local utilisation with system bottleneck.

Consequence

Correct bottleneck analysis focuses resources where they change output, improving throughput, reducing lead time and guiding investment; misidentification wastes resources, may increase upstream WIP and can create new constraints elsewhere.

Reversal

In systems with frequent routeing alternation, service‑level demand limits, or rapidly shifting stochastic loads, the bottleneck location can move or be diffuse; in such settings a static single‑point bottleneck model is inadequate and continuous or probabilistic analysis is required.

Boundary

Clearly within: a paced manufacturing line where a single station has lower effective capacity than required. Boundary case: networked production with parallel machines and dynamic routing where bottlenecks may be transient. Clearly outside: market demand limitations where throughput is capped by external orders rather than production constraints.

Semantic Tension

Local efficiency (improving utilisation) ↔ System throughput optimisation: actions that increase local utilisation may not increase and can even reduce overall throughput if they are applied away from the constraint.

Synthesis

Bottleneck Analysis reframes improvement from isolated optimisation to constraint management: the goal is to change the system’s limiting condition, not to maximise every resource’s utilisation.