theory-of-constraints

To find the bottleneck in a slow delivery pipeline, identify the true constraint by locating work-in-progress buildup, queues, downstream starvation, and idle time. This isolates the single constraint limiting system throughput and flow.

Local optimization fails to improve throughput because it ignores the system constraint, scattering effort instead of focusing flow. You must subordinate non-constraints and manage global flow rather than local efficiency to increase output.

The best way to increase throughput without adding resources is to exploit the existing constraint with no-cost actions. Identify the bottleneck, then apply exploit, subordination, and elevation steps to maximize flow before investing elsewhere.

To identify a false bottleneck in process optimization, analyze WIP and queue signals alongside throughput evidence. A true constraint shows downstream starvation and WIP buildup, whereas a false bottleneck lacks these global flow starvation patterns.

Yes, theory of constraints can analyze scaling problems across agent workflows. It locates the limiting constraint via WIP and idle signals, then applies targeted exploit and elevation actions to improve throughput across the entire process.