The Theory of Constraints

In one paragraph

Whatever a business produces, one step limits the whole thing. Speeding up any other step just piles work in front of that one.

Goldratt's claim is that at any moment a system's output is set by a single constraint. It follows that effort spent improving non-constraints produces no additional throughput at all — it produces work-in-progress.

The five focusing steps

1. Identify the constraint — usually where work queues up in front of something. 2. Exploit it — make sure it is never idle, never working on the wrong thing, never waiting on setup. This step is free and is usually skipped. 3. Subordinate everything else to it — the rest of the system runs at the constraint's pace, deliberately, even though that means people and machines are not fully busy.

4. Elevate it — now spend money: more capacity, another machine, another person. 5. Repeat — because once you fix one constraint, it moves somewhere else, and the old rules you made for the old constraint become the new problem.

Step three is the one nobody can stomach. It means deliberately allowing parts of the business to be under-used, which looks like waste and every efficiency measure will punish it.

Throughput, inventory, operating expense

Goldratt argued for measuring three things: throughput (the rate money is generated by sales), inventory (money tied up in things bought), and operating expense (money spent turning one into the other). Conventional cost accounting, he argued, encourages local efficiency that damages all three.

How to actually use it

• Walk your process and find where work is waiting. That queue is the constraint, and it is nearly always a person or a single machine rather than a department.
• Before spending anything, exploit it: stop the constraint doing work that could be done elsewhere, cover its breaks, and make sure it never waits for materials or decisions.

Where it breaks down

• Clearest in manufacturing and other flow processes. In a business where the constraint is demand, cash or judgement, identifying it is much harder.
• It sits badly with conventional management accounting, so applying it usually means arguing with the reporting system as well as the process.

If you are being examined on it

Identify the constraint from the case, then apply the five steps in order. Recommending investment before exploitation is the standard error.

More on operations