Bottleneck vs Non-Bottleneck ROI
Compare two improvement projects on real system throughput value, not local capacity gain, since only bottleneck work actually increases output.
Project A
Project B
How this works
Improving a non-bottleneck resource's local capacity delivers little or no real system throughput gain, because the system's output is still capped by the bottleneck elsewhere. A project's real system value is its local capacity gain valued at the system throughput rate, but only if the project actually targets the bottleneck.
Key terms
- Bottleneck: the resource whose capacity actually limits the system's total output.
- Real system value: local capacity gain converted to real throughput value, zero for a non-bottleneck project regardless of size.
Assumptions
- The bottleneck's identity is known and stable across the comparison
Limitations
- Does not model the bottleneck shifting to a different resource after either project completes
Sources
Theory of Constraints (Goldratt, The Goal), the same framing already used in Constraint Improvement Value and OEE Loss Prioritizer.
Worked example
Project B's local capacity gain (500 units/year) is more than double Project A's (200 units/year), and costs less, a local-gain-only comparison would favor B. But B doesn't target the system's actual bottleneck, so its real system value is $0. Project A, targeting the bottleneck, delivers $8,000/year despite its smaller local number and higher cost.
Common use cases
- Prioritizing between two improvement projects competing for the same budget
- Avoiding a local-efficiency trap that funds a non-bottleneck improvement
- Following up a constraint-value analysis with a head-to-head project comparison