WHERE TO START

As an improvement expert, I am sometimes asked, “How do you know where to start? How do you zero in on a problem area with a high likelihood that sustainable improvements can be made?”

The answer is two parts consisting of: 1) the obvious and familiar starting strategy of ‘follow the money’; and 2) the “trade secret” of improvement.

"Following the money" simply means auditing the ‘critical path’ – a sequence of tasks that must be executed in order for the business to get paid. This path begins with the sale and ends when good product and a correct invoice is delivered to a satisfied customer. I look for the things you might expect along this path, capacities & bottlenecks, first pass yields, proper tools for the job, and so on.

The “trade secret” of successful improvement is this: Look for indecision, uncertainty, and trial & error behavior. Now, let me stress that this is not the problem. It’s a symptom or warning flag, an “X” that marks exactly the spot on the business ‘critical path” where one should “dig for treasure”. Improvement is not simply achieved by squashing disagreement and reckless boldness.

I have tracked these clues through all kinds of processes: call centers, accounting operations, technology infrastructure and manufacturing. One of the clearest illustrations that 'trial and error behavior' marks a serious process problem is the story of one client’s metal machining operation.

THE STORY OF THE “EIGHT” STATION

The plant had everything from low end computer controlled cutting equipment (“CNC” machines) to high end finishing equipment designed for better than 100 millionths of an inch tolerances.

The natural attractions to the engineers are all of the complex CNC devices, and the high precision finishing tools, but in this facility there was a “big payoff” opportunity that had not been discovered.

An inexpensive series of “low tech” workstations prepared every single metal casting before they went into the CNC or finish-work processes. The centerpiece of this line was called, the “eight” station. Designed for low cost and maximum throughput, the heart of this workstation was a turntable on which eight un-worked metal castings were placed at forty-five degree intervals.

The table would make a 1/8th turn, and each of seven machining stations would make their specific cut. Station #8 was for the operator who loaded and unloaded parts. It seemed to be a bottleneck, but was there room for improvement?

It was soon obvious to me that this particular operation was a den of confusion. The operators called it, “the machine from hell”. “Only Johnny can set it up, and it takes hours.”, they said.

As I watched, I quickly saw why the machine was so difficult to set up. Eight fixtures for holding parts must be bolted onto the turntable. They could be placed too close or too far from the center of the table. They could be placed too far to the left or to the right. Two “degrees of freedom” times eight fixtures equals sixteen different adjustments that must be exactly right.
There was more. The operators had to unbolt and lever the machining stations into position.

Seven moveable machining stations meant fourteen more dimensions to the problem. There were a grand total of 30 independent but interrelated dimensional adjustments that had to be brought into “harmony” before the machine was set correctly.

I watched as “Johnny”, the best operator in the house, struggled for hours to get all the pieces of this machine into position. Tweaking the setup by trial and error, one adjustment always led to a dozen re-adjustments. A “trial run” meant wasting raw material by making sample cuts that were sent to the measurement lab for evaluation. It was usually fatigue and an urgent demand for production that caused him to surrender the cause, and actually attempt to produce a product.

THE SOLUTION

I took a little time to think about the problem. How should the setup of this machine be handled? When I thought I had the answer, I went back to Johnny. “Tell me”, I asked, “Is there anything on this machine you won’t move to make a correct setup?”

Johnny answered, “Well, the engineers told me that I should try not to move the stations on the left and right of the operator position. They didn’t know why. I tried the leave them alone, but I’ve had to move them anyhow.”

His answer confirmed my suspicions. The designer had intended that two of the workstations together with the exact center of the turntable form a perfect ninety-degree angle. Those three points would establish the absolute reference frame from which every component could easily and systematically be placed in its one and only one correct position.

Once this “carpenter’s square” had been broken, the machine could no longer be set up with certainty. There was no longer “one and only one place” to position any component. The solution was: 1) to reestablish the permanent location of the two operator-adjacent stations by precision measurement, and 2) to teach the operators that the setup process must be remembered as:

A) The built-in 90-degree reference angle shows where to exactly place
B) Fixtures, which show you the “one and only one” place for the
C) Remaining five stations.

There were some optional elements to this improved setup, but it was no longer trial and error. Instead, it was now a consistent process with a predictable duration and a certain outcome.

THE REST OF THE STORY

The rest of the story has to answer one big question.

What drove the operators to break the rule, and move the two forbidden machine stations?

Of course, the obvious point is that no one in the plant truly understood the significance of those two stations on either side of the operator. An implicit right angle is a way to establish a repeatable frame of reference for measurement.

The not as obvious, but most fundamental flaw was in the management’s decision rules established for: “How do you know when a setup is correct?” The proper answer is, “When you are certain that you have adhered to a correct setup process.” It is engineering’s responsibility to see that a correct process will produce a correct product.

In this plant, a flawed decision rule had been substituted. Setup was deemed to be “wrong” if a sample part failed a QC measurement check. This policy is flawed from the start because of the assumption that a single selected sample will always be ‘average’. Questionable measurement precision in this plant compounded the problem, and drove operators relentlessly to a point where they felt they had not alternative but to “break” the machine to gain setup approval.

FOOTNOTE for the Mechanical Engineer:

“Pinning” this machine was not the solution, and only add constraints that make it impossible to correct the problem. After the reference angle was broken, engineering had tried to sort out the mess by pinning some of the fixtures, making matters worse.