Throughput
Why it matters
A business can feel frantic and still have low throughput, because activity and output are different things. The process's constraint sets the ceiling. Improving a step that is not the constraint just moves the queue around.
A related idea is Little's Law, a result from queueing theory: the average amount of work in progress equals throughput multiplied by the average time each item takes. If throughput is fixed, more work in progress means longer waits. Running everyone at full capacity therefore works against flow, because a system with no slack has nowhere to absorb variation.
How to apply it
- Pick one honest measure of finished output for the process, in a fixed period.
- Track it separately from busyness and from the amount of work in progress.
- Find the step that limits the number before trying to speed anything else up.
- Limit how much work is started at once. Finishing is worth more than starting.
- Recheck the number after each change to confirm it moved.
What it is
Throughput answers one question: how many things came out the end? A support team measures tickets closed per week. A studio measures projects delivered per month. A shop measures orders dispatched per day. Work that is started but not finished does not count, however much effort it has absorbed.
In the Theory of Constraints, Goldratt also used the word for the rate at which a business turns work into money through sales. In everyday operations the first meaning is the one to use.
Common mistakes
- Counting started work. Tickets opened or projects begun are not output. Count only what is finished.
- Confusing busy with productive. Full diaries and long hours say nothing about how much came out the end.
- Measuring every step. Only the constraint sets throughput. Speeding up other steps moves the queue.
- Ignoring work in progress. More work in flight at once means each item waits longer. Limit how much starts.
- Counting unlike things together. A quick fix and a major project are not the same unit. Measure comparable items or group them by size.
- Pushing for 100 percent utilisation. A team with no slack cannot absorb surprises, so queues and delays grow.
