Estimate current AND maximum throughput for each workflow step — the gap reveals hidden capacity that upstream constraints are masking
For each step in a multi-stage workflow, estimate both current throughput (what it processes now) and maximum throughput (what it could process if fully loaded) to reveal hidden capacity that upstream constraints are masking.
Why This Is a Rule
Current throughput alone can't identify constraints because upstream bottlenecks starve downstream steps. The dual measurement — current throughput (what is processed now) and maximum throughput (what could be processed if fully loaded) — reveals the hidden capacity distribution. A step with current=50 and max=200 has 75% spare capacity (not a constraint). A step with current=50 and max=55 has 9% spare capacity (nearly a constraint). This data produces Map the constraint cascade before intervening: rank all steps by max capacity to predict where the next bottleneck will appear's cascade map and enables Local metric improved 50%+ but throughput gained less than 25%? That's a constraint cascade — immediately map downstream capacity limits's local-vs-system diagnostic.
When This Fires
- When building Map the constraint cascade before intervening: rank all steps by max capacity to predict where the next bottleneck will appear's cascade map before a bottleneck intervention
- When current-throughput-only measurement makes all steps look equally busy
- When you need to predict where the constraint will shift after fixing the current one
Common Failure Mode
Current-only measurement: "All steps process about 50 items per week, so I can't tell which is the bottleneck." The current throughput is equalized by the upstream constraint. Only by estimating max capacity do you discover the true constraint.
The Protocol
(1) For each step, estimate: Current throughput (count from actual data) and Maximum throughput (hypothetical if unlimited input). (2) Calculate headroom: (max - current) / max × 100%. (3) Steps with headroom under 10% → current or imminent constraints. Steps over 50% → well below capacity. (4) Rank by maximum capacity (lowest to highest) to map the cascade (Map the constraint cascade before intervening: rank all steps by max capacity to predict where the next bottleneck will appear). (5) The step with lowest maximum capacity is the system's ultimate constraint.