The Best-Laid Improvements of Mice and Men Often Go Awry

Image created by DALL·E 2 (OpenAI)

In the 1980s, you could say that the notorious Charles Koch was obsessed with Edwards Deming. Charles Koch joined his father's business, Rock Island Oil & Refining Company (now Koch Industries), in 1961. Koch became president of the company in 1967. In a legal battle with his brothers, Frederick and Bill sold their stakes for $1.1 billion, and Charles and David gained majority ownership. In his 2007 book, "The Science of Success: How Market-Based Management Built the World's Largest Private Company," Charles Koch cites Dr. Deming as an important influence. Charles Koch started applying Deming's methods to Koch Industries in the early 1980s. Koch said that Deming was more than a business consultant; he was a Guru.

Koch Industries applied Deming's concept of continuous improvement, and the results were remarkable. Phil Dubose, an oil gauger in Louisiana, was one of the most successful students of Deming's continuous improvement. He managed a fleet of barges that collected crude oil from terminals in the Gulf of Mexico and then shipped it to refineries in Texas. A few of his barges even traveled north on the Mississippi River to Koch's refinery in Minnesota. As Koch and Dubose saw it, measuring and improving the economics of marine operations for a fleet of barges that collected crude oil was not an exact science back then. Small tanks, constantly changing conditions, turnover, and training were all issues making the process an uncertain art. As a perfect fit, Koch and Dubose chose Dr. Deming's philosophy of using math and statistics to improve operations.

Dubose provided each vessel with a simple run chart. The run chart broke down barge costs. Groceries, fuel, maintenance, ship damage, and supplies are into separate categories. Over time run charts could be used to track these costs using analytical statistics, Statistical Process Control (SPC). Each barge captain would have a set of these visual charts. Captains could see where they were spending money and saving it. This new process allowed them to make most of their decisions based on the charts (the data). Then Dubose began tracking profits and losses per vessel. The captains had all the information they needed to boost profits and the freedom to act on it. Dubose had total visibility into his fleet; he knew which boats were making money and which were losing money.

In her book Thinking in Systems, Donella Meadows tells a similar story. An entire suburb of single-family houses was built at the same time near Amsterdam. Some houses were built with electric meters in the basements. Other homes had an electric meter in the front hall. Electric meters like these have a glass bubble with a small horizontal metal wheel inside them. A dial adds up kilowatt-hours as the household uses more electricity. The wheel turns faster as the family consumes more electricity. Early in the 1970s, the Dutch began paying close attention to their energy consumption due to the oil embargo and energy crisis. Compared to other subdivisions, this subdivision used one-third less electricity. At first, there was no explanation for this. After further investigation, It turned out that the placement of the electric meter was the reason. The households with the meter in the hallway used less electricity than the ones in the basement. One of the strongest features of Deming's theory of variation is using charts to visualize the data to help organize and continuously improve.

Around 1988 a young lawyer named Ken Ballen went to work for the US Senate investigating committee. One of the areas he was tasked with looking into was looking at oil theft. Koch Industries had a notorious reputation for what the industry called going long on oil sales. Going long was polite to say that they received more oil than they paid. One of the primary whistle-blowers was Bill Koch's younger brother, who felt he had been screwed out of the family fortune. When Charles Koch was bought in for questioning, Ballen pressed him on their measurement process. Koch tried to explain Koch Industries' policies of continuous improvement. In Christopher Leanord's book, Koshland, the dialog went like this:

Just what was a continuous improvement, exactly? Ballen asked. "How much time do you have?" Koch replied. "How much time do you have?" Ballen replied. "Continuous improvement philosophy is a philosophy developed by Edward Deming, a statistician," Koch said.

Koch then went on for a long time to explain Dr. Deming's philosophy and how it was used to manage the fleet of barges that collected crude oil. To Ballen, it also seemed like an elaborate rouse to divert his attention from the fact that Koch Industries was involved in oil theft.

This story is unique in that both were true. They were improving, but they were also probably stealing. Koch Industries was in the business of distributing oil, and Koch was right that oil transportation was not an exact science. Thus, it was challenging to determine how much an oil company collected versus how much they were paid. In many ways, this process is similar to how hedge funds buy and sell stocks. However, there is also a fine line between aggressive market behavior and outright criminal behavior. When does a continuous improvement philosophy become continuous criminal activity? It reminds me of the fictional character Gordon Gekko in Wall Street.

One of the more interesting parts of this story is when Ballen interviewed the captains of the vessels; they all saw Koch's improvement philosophy as something he was doing to pressure them to make more money. Again this was true. Koch's improvement policies were about making more money. However, it was clear that the captain's never understood the context of their continuous improvement process. The interviews confirmed Ballen's assertion that Koch Industry was stealing oil. It is possible that Koch made a mistake by not explaining the why behind his improvement philosophy to the vessel captains. Continuous improvement was just seen as another accounting process by the captains. When talking to Ballen, they might have had a better explanation if they understood the real purpose of all those run charts. This kind of mistake is common in DevOps. Leaders in organizations demand that everyone start doing DevOps, but they never explain why or in what context. Thus, the implementors perceive the new processes as just tool changes, not philosophical ones.