This is the final exam for Bottlenecks & the Theory of Constraints. It pulls together the whole course: throughput as the capacity of the single slowest step, the tell-tale signs that let you spot a constraint, why a faster non-bottleneck adds exactly nothing, Goldratt’s five focusing steps, the hard ceiling that Amdahl’s law puts on any partial speedup, and the restless way the constraint moves the instant you fix it. Several questions look easy until you notice an answer that averages the stages, sums them, or credits the shiny fast machine. Reason each one through.
How this exam works
Read carefully — this exam is final. Each question appears one at a time. Once you submit an answer it is locked for good: there’s no going back, no retry, and no restart. Your score is hidden until the end, where you’ll see a pass/fail verdict. The pass mark is 70%. A few questions ask you to select all correct answers.
What is the throughput of a chain of steps?
Select an answer to continue.
Course Recap
Big picture
Bottlenecks & the Theory of Constraints, in one picture
- Bottlenecks & the Theory of Constraints
- Throughput = the bottleneck
- A system runs at the speed of its slowest step — the minimum capacity, never the average or the sum. The wide body of the bottle is irrelevant; only the neck matters.
- Spotting the constraint
- Work piles up IN FRONT of it (growing queue/inventory); stages AFTER it sit starved and idle. A stage faster than demand isn't binding — the constraint can move OUTSIDE to the market.
- Theory of Constraints (5 steps)
- Identify, Exploit (squeeze free output), Subordinate (pace everything to it; idle non-bottlenecks are fine), Elevate (spend to add capacity — only after exploiting), Repeat. An hour lost at the bottleneck is lost for all; an hour saved elsewhere is a mirage.
- Local vs. global optimization
- Improving a non-bottleneck adds ZERO throughput and can be worse than useless — WIP pile-up, longer lead times, the "keep everyone busy" utilization trap.
- Amdahl's law
- speedup = 1 / ((1 - p) + p / s). The un-improvable fraction caps everything: ceiling = 1 / (1 - p). 25% serial means 4x max, even with infinite hardware. It's the bottleneck in arithmetic.
- The constraint always moves
- Fix one bottleneck and a new slowest step takes over — it's a position, not a place. Whack-a-mole is progress. Watch for policy constraints (inertia). Highways, approvers, slow queries, the one bathroom.
- Throughput = the bottleneck
Key takeaways
Every system is a chain, and its real output — throughput — is set by the single slowest step and by nothing else: not the average of the stages, not their sum, not the shiny fast machine. Spot the constraint by the pile-up in front of it and the idleness behind it; remember a stage faster than demand isn’t binding, and the constraint can sit outside in the market. Then govern it with Goldratt’s five focusing steps — Identify, Exploit, Subordinate, Elevate, Repeat — squeezing free output before you ever spend to add capacity, and accepting that idle non-bottlenecks are fine. Beware the deep instinct to polish the wrong stone: improving a non-bottleneck adds zero throughput and can be worse than useless, piling up work-in-progress and stretching lead times. Amdahl’s law puts that in numbers — the un-improvable fraction caps the whole gain at , so 25% serial means 4x at most no matter the hardware. And expect the constraint to move the instant you fix it: it’s a position, not a place, and chasing it from highway merges to slow queries to the one bathroom is the work. Stop optimizing the parts. Govern the constraint.