Manufacturing — and even services — often work a lot like traffic bottlenecks on your local freeway. To improve production, you need to focus on and then break the bottleneck or constraint. Dr. Eliyahu Goldratt, the founder of the theory of constraints, identified five steps for managing processes:
Identify system constraints.
Find the bottleneck that slows down the process.
Exploit the constraints.
Make sure that you get the most out of constrained resources.
Subordinate everything to the constraints.
Rearrange other processes so that they work at the same pace as the constraint.
Break the constraints.
Find ways to increase the capacity of the constraint.
Go back to Step 1.
After the constraint is broken, go back to Step 1 to identify another constraint to work on.
Step 1: Identifying system constraints
Inevitably, in any process, something slows everything else down; there’s always a weakest link. In a production process, it’s usually the department with the slowest output.
A number of factors can cause constraints. They can include
Poorly trained people who lack the skills to do their jobs
Inefficient equipment that has insufficient capacity to meet production needs
Company policies (written, or unwritten) that prevent people or equipment from accomplishing more
According to the theory of constraints, every system has at least one constraint. But what if every step in the process operates below capacity? In that case, the constraint is customer demand.
Goldratt calls the constraint the drum, because it sets the tempo that the entire production process will need to follow.
Suppose that the fictional Axiamm Corp. manufactures Tex’s Hot Sauce. Production requires five production steps.
Axiamm starts by buying the freshest and spiciest raw materials. These ingredients need to be prepared in various ways; some need to be pureed and/or cooked. Next, the Blending department measures the ingredients into a giant blender so that each batch can be mixed together. Then every batch needs to be tasted by a quality control inspector. Finally, each batch is bottled and packed into cases.
Alice, the plant manager, is concerned about the Quality Control department, which always seems backlogged. On any given day, six or even seven batches of hot sauce are waiting for inspection. This backlog then stops up the Blending department, forcing it to pause production while Quality Control catches up.
Alice identifies the Quality Control department as a constraint, primarily because of the lack of taste experts. The company has always had difficulty hiring well-qualified tasters who can verify that each product tastes right.
Step 2: Exploiting the constraint
After you identify the process that slows down the system, carefully consider how you can optimize the use of this constraint.
In parkway traffic, go to the bottleneck and try to get the automobiles to drive through it as quickly as possible. When hiking with small children, offer the youngest one incentives to move more quickly (snacks ahead, bears behind).
In a production process, you may be able to save time by setting up a buffer before the constraint. This buffer is a stockpile of inputs, ready and waiting for the constrained step in the process so that your constraint never needs to stop working in order to wait for more inputs.
You also may be able to exploit the constraint by maximizing contribution margin per unit of constrained resource. For example, consider how Axiamm Corp. can use this technique.
When making hot sauce, the Quality Control department already has a buffer waiting to get tested — the backlog of batches that stretch all the way back to the Blending department. Suppose that Tex’s Hot Sauce comes in three different flavors, each of which provides a different contribution margin:
Tex’s Very Hot Sauce: Contribution margin of $1,000 per vat
Tex’s Super Hot Sauce: Contribution margin of $700 per vat
Tex’s Atomic Hot Sauce: Contribution margin of $500 per vat
To exploit this constraint, Alice should first have Quality Control test the Very Hot Sauce; this strategy will bring in the most contribution margin: $1,000 per vat.
After the tasters get through the Very Hot Sauce, they should move on to the Super Hot Sauce, which has lower contribution, just $700 per vat. Finally, if they have any time left, they should test the Atomic Hot Sauce, which provides only $500 of contribution margin.
By exploiting the constraint based on contribution margin, the tasters are sure to test the most-profitable product before they work on less-profitable products.
Step 3: Subordinating everything to the constraint
In Step 3, you rework your entire production schedule so that it moves at the same pace as the drum. Any processes going faster than the drum produce unnecessary units. Goldratt calls this pacing of processes the rope, such that goods move through production like a single rope — all at the drum’s pace.
In traffic, subordinating everything to the constraint means getting all the cars to drive at the same speed — the maximum speed that they can drive through the bottleneck. For Axiamm Corp., it means purchasing, preparing, and blending the raw materials to make as much Very Hot Sauce as can be sold because this product has the highest contribution margin.
Step 4: Breaking the constraint
Breaking the constraint means creatively thinking about alternative ways around the constraint. In traffic, you can get off the parkway and take side streets. In hiking, maybe the little one can keep up by taking a shortcut.
At Axiamm Corp., Alice and her colleagues are thinking about how to increase capacity in Quality Control. They’ve tried recruiting more workers, to no avail. They’ve also tried forcing the tasters to test more vats of hot sauce each day. That didn’t turn out well, either.
Then Alice had a great idea. The Blending department usually blends 50 gallons of hot sauce at a time and then sends them to Quality Control in 50-gallon vats. However, the blenders have capacity to handle 100-gallon loads. Therefore, why not blend 100 gallons at a time and replace the 50-gallon vats with 100-gallon vats? This increase can cut the required number of quality control tests in half.
Problem solved. Now the Quality Control department can handle a full production load.
Step 5: Returning to Step 1
After you break the constraint in Step 4, start over again. Take a look at the production process to identify new constraints, exploit them, subordinate everything to them, and then break them.