System Thinking

Do you really know your constraint?

Mohit Saini · 22 Mar 2018 · 6 min read

Do you really know your constraint?

Ask a room full of managers what the constraint of their company is, and you will get as many answers as there are people in the room. Production will say it is the old press line. Sales will say it is the price. Purchasing will say it is a supplier who never delivers on time. HR will say it is attrition.

Each of them is talking about a real problem. But are any of them talking about the constraint?

I have often heard people say, “TOC is just common sense. We identify our constraints and remove them every day.” It sounds right. But in my experience, this is exactly what Theory of Constraints is not.

Obstacle, limitation or constraint?

In everyday language, a constraint is anything that holds us back. So any manager who clears a hurdle on the way to the monthly target feels he is “working on the constraint”.

TOC uses the word much more narrowly. Let us separate two things:

  • An obstacle is a limitation that can be removed as soon as it is spotted. A missing drawing, a machine waiting for an operator, an approval stuck in someone’s inbox. Obstacles should simply be removed. There is nothing strategic about it.
  • A constraint is a limitation that cannot be removed in the near future. The capacity of a key resource that takes a year to expand. The demand the market is willing to give you this quarter.

Every system has at least one such limitation. If it didn’t, the company would be making infinite profit. Think about it.

If a constraint cannot be removed quickly, the only way to improve the system now is to get the most out of it.

That is what Dr. Eliyahu M. Goldratt meant by exploiting the constraint.

Common sense is not common practice

“Fine,” you might say, “getting the most out of the scarce resource is common sense.” It is. But let us test how commonly it is practised.

Think of a housing society in peak summer, when the municipal water supply drops to a third of normal for a few weeks. Water is now the constraint, and it cannot be expanded quickly.

What do people do? Many fill every bucket and drum they own, just in case. The overhead tank empties by morning. Families on the top floors get nothing. Seeing this, everyone stores even more the next day. The scarcity gets worse because of how people react to it.

What should the society do first? Not dig a new borewell (that takes months). The first step is to agree on a goal (every flat gets enough water every day) and then set simple rules: fixed supply hours, a fair share per flat, no storing beyond a day’s need. The same water now serves everyone.

Notice three things in this example:

  1. The goal had to be agreed first. Without agreement on “every flat, every day”, there is no basis for any rule.
  2. The biggest waste of a scarce resource is not using it badly; it is someone holding it while someone else goes without.
  3. The rules only work if everyone follows them. One family that keeps hoarding breaks the system for all.

This is TOC in miniature.

The steps, in plain words

To practise TOC, one has to internalise four steps:

  1. Agree on the goal of the system.
  2. Identify the constraint: the limitation that cannot be removed quickly, whether capacity or market demand.
  3. Decide how to exploit it: how to get the most out of it as it is.
  4. Subordinate everything else to that decision.

Only after these does it make sense to talk about elevating the constraint, that is, spending money to expand it. Very often, exploiting and subordinating reveal so much hidden capacity that the expansion is not needed at all, or at least not as soon as we thought.

Constraints and non-constraints

There is a dangerous misreading of this idea: “If only the constraint matters, we can ignore the rest.”

The opposite is true. Most resources in any system are non-constraints, and they are connected to the constraint. If a non-constraint upstream starves the constraint, or a non-constraint downstream scraps what the constraint produced, the whole system loses.

What changes is the rule for non-constraints. For the constraint, more output is always better: an hour lost there is an hour lost for the whole company. For a non-constraint, more is better only up to a point. Beyond that point, more is worse: it creates excess stock, clogs the floor and hides the real priorities.

And where is that point? It cannot be found by looking at the non-constraint alone. It is set by what the constraint needs.

This is why measuring every department on its own efficiency does so much damage.

So, are you subordinating?

Here is a simple self-check.

If capacity is your constraint:

  • Is the plant making items that aren’t needed yet, while urgent orders wait?
  • Does the constraint ever wait for material, drawings or an operator?
  • Is it doing work that a non-constraint could do?

If the answer to any of these is yes, the constraint is being wasted, not exploited.

If the market is your constraint:

  • Are your current customers getting what they want, when and where they want it?
  • Are you reaching every customer who would buy from you?
  • Is your sales team spending its time selling, or chasing dispatches and payments?

If not, the most precious thing you have (demand) is leaking away.

What an hour on the constraint is worth

Put a number on it. Picture a plant whose heat-treatment furnace is the constraint. Every order passes through it, and it runs 20 hours a day, about 6,000 hours a year. The plant sells ₹300 million a year, and materials make up about half of that. So the furnace, in effect, generates about ₹150 million of throughput a year (sales minus materials): roughly ₹25,000 for every hour it runs.

Now walk the floor with that number in mind:

  • The furnace stops for 40 minutes at each shift change while the next crew settles in. Two changes a day is 80 minutes, worth over ₹30,000, every day. Over a year, close to ₹10 million of throughput. Staggering the breaks so the furnace keeps running costs a little overtime.
  • A batch is rejected at final inspection after it has been through the furnace. That isn’t a loss of a few hundred rupees of material. It is furnace time that can never be recovered. An inspection before the furnace suddenly becomes very valuable.
  • The furnace waits an hour for a trolley because the forklift driver is busy elsewhere. The forklift is a non-constraint; its idle time costs almost nothing. The furnace’s idle time costs ₹25,000.

None of these needs a new furnace. Each needs someone to see that an hour lost on the constraint is an hour lost for the whole plant, and an hour saved on a non-constraint is often a mirage.

The real work is subordination

Identifying the constraint is often the easy part. Deciding how to exploit it is a matter of logic. The hard part, the step that separates companies that talk TOC from those that practise it, is subordination: getting every department to follow the constraint’s needs, even when it hurts their own numbers.

So the next time someone tells you they practise TOC every day, ask them one question: What did your non-constraints stop doing this week so that the constraint could do more?

The answer will tell you whether they know their constraint, or just their obstacles.

Mohit Saini

Strategy, supply chain & sales transformation consultant. About · Book a discussion