Process bottleneck
A process bottleneck is the step in a workflow with the lowest throughput — the constraint that limits the entire process's output. Identifying the bottleneck is the prerequisite to any throughput-improvement work, because non-bottleneck improvements don't move the system-level throughput.
Bottlenecks reveal themselves through queues: the step with growing backlog upstream of it is usually the bottleneck. Bottlenecks shift: fixing one moves the bottleneck somewhere else (usually visible within days). The Theory of Constraints prescribes a repeating cycle: identify, exploit, subordinate, elevate, repeat. The most common bottleneck-management mistakes are working on the wrong constraint (perceived but not actual) and breaking subordination (pushing work into the bottleneck faster than it can drain). Modern observability — process mining especially — has made bottleneck identification dramatically more rigorous than it was in the spreadsheet era.
Discussed in our use-cases
ICP-targeted pages where process bottleneck is part of the framing.
Related terms
- Theory of constraints
Theory of constraints is Eli Goldratt's management methodology built on the observation that every system has one constraint (bottleneck) that limits its total throughput — and the only improvement that increases throughput is the one that elevates the constraint.
- Flow efficiency
Flow efficiency is the ratio of value-add time to total elapsed time for a case — typically expressed as a percentage.
- Cumulative flow diagram (CFD)
A cumulative flow diagram is a stacked area chart showing, for each state in a workflow, how many items are in that state over time.