Goldratt's Theory of Constraints is an observation about production systems that a chain is only as strong as its weakest link. Throughput is set by the bottleneck. Capacity added anywhere else does not raise output; it only grows inventory, WIP, and the illusion of progress.
It was introduced by Eliyahu M. Goldratt in The Goal (1984). The practice is a repeating loop:
- Identify the constraint that limits the whole.
- Exploit it—get every useful unit of work through it before spending elsewhere.
- Subordinate every other step to the pace and needs of that constraint.
- Elevate the constraint only after it is fully exploited.
- Repeat—when the bottleneck moves, chase the next one.
- Staff every team equally while one shared platform queue blocks all releases, and headcount outside the queue does nothing for delivery.
- Optimize local cycle time on a non-constraint feature team, and you fill the pipeline faster without shipping more.
- Fund six initiatives when three already compete for the same scarce people, and the system slows under WIP rather than under ambition.
Local efficiency is not system efficiency. Until the constraint is named and protected, "more capacity" is just more work waiting.