Five stages in a row. Items flow left to right. Each stage has a capacity: how many items it can process per second. The queue shows items waiting before each stage.
Here is the thing. It does not matter how fast stages A, B, D, and E are. If stage C can only do 2 per second, the whole pipeline does 2 per second. That is the bottleneck. Everything else just builds up inventory.
Try it. Set one stage to 1 and the rest to 10. Then speed up every stage except the slow one. Watch the throughput stay flat. Then fix the bottleneck. That is the Theory of Constraints in 30 seconds.
Jargon you just learned
- Bottleneck
- The stage with the lowest capacity. It limits the entire system.
- Throughput
- Total output of the system. Items per second coming out the other end.
- Capacity
- How much a single stage can process per unit of time.
- Constraint
- The thing limiting system performance. Same as bottleneck, fancier word.
- Queue
- Items waiting in line before a slow stage. Where inventory and frustration pile up.
- Theory of Constraints
- Goldratt's idea: find the bottleneck, fix it, repeat. Ignore everything else.
What to notice
Queues pile up before the bottleneck, not after it. Stages after the bottleneck are starved. They sit idle no matter how fast they are. This is why "optimizing everything" is a waste. Find the constraint. Fix that one thing. Then find the next one. Repeat. Goldratt figured this out in the 1980s and most organizations still have not learned it.