(2026-02-17) Poor Deming Never Stood A Chance

Lorin Hochstein: Poor Deming never stood a chance. The two management giants of the mid-twentieth century were Peter Drucker and W. Edwards Deming... it was Drucker that proved to be more influential in America.

If you’ve ever been at an organization that uses OKRs, then you have worked in the shadow of Drucker’s legacy

While you can tell a story about how Deming influenced Toyota, and Toyota inspired the lean movement, I would still describe management in the U.S. as Deming in exile.

Here I want to talk about why I think it is that Drucker’s ideas were stickier than Deming’s in the U.S. It all comes down to the nature of organizations and people.

the bigger the organization is, the hairier and more complex it gets. Managers, on the other hand, have a very finite amount of bandwidth

How is a manger to make sense of this mess?

OKRs as a mess-reduction mechanism

Key results reduce the bandwidth required to make sense of what’s happening in the system

It’s no coincidence that when John Doerr wrote a book on his experience with OKRs at Intel and how he brought them to Google, he titled his book Measure What Matters.

Deming’s approach to the problem of management was radically different from Drucker’s

He uses the term deadly disease to describe managing through numerical targets:
"Eliminate management by objective. Eliminate management by numbers, numerical goals. Substitute leadership.""

He argued that if you wanted improvements, you had to make systemic changes. Furthermore, you had to understand the system if you wanted to come up with a system improvement that would actually work.

Classical control versus statistical control

Deming was not opposed to the idea of goals: indeed, he was a passionate believer that management should strive to improve quality and productivity, and both of those are goals. He was also not opposed to metrics: he was an advocate of applying Walter Shewhart’s statistical techniques for management. It’s the use of metrics that’s radically different in the two approaches.

The Drucker-ian approach is akin to a classical control system, like a thermostat.

Deming wrote in explicit terms about control, but he meant it in a different sense: we wrote about statistical process control, and statistical process control is about the variability of the output.

Deming argued that you had to understand whether your system was under statistical control in order to determine what intervention to do in order to make an improvement. For example, if your system was out of control, the next intervention would be to do a qualitative investigation into the outliers. On the other hand, if the system was under statistical control, then you’d have to figure out what systemic change to make to improve things.

Note how you build a classical control system, whereas you observe whether your system is under statistical process control. Statistical process control is about understanding the system.

One of the virtues of OKRs is that they are straightforward for managers to apply

Deming’s approach, on the other hand, requires a much greater commitment of management bandwidth

Deming requires a never-ending research program, with no upper bound on the kind of information that might be relevant.