Ah I don't remember the exact words but one of my favourite images he painted was something along the lines of

"Managing an organisation by quarterly reports is like driving a car by looking at the centre line in the rear view mirror."

>"It is wrong to suppose that if you can’t measure it, you can’t manage it – a costly myth."

Ironically, this may have been a reaction an early software management book in the 80s by Tom DeMarco of Peopleware fame and similar books that were popular at the time. It's kinda annoying that the people that are maintaining a website in Deming's name don't reference the first edition book dates for the quote attributions. The original book seem to have been published in 1993 (when Deming was 93...), whereas the quote is attributed to a 2018 third edition on the website.

The reason I find it ironic is that Deming seems to have been critiquing the young (relatively) influential upstarts at the time like DeMarco for over-emphasizing the importance of management metrics, And as it turns out, the nuances in his approach turned out to be more compatible with the approach that came to dominate the post-millennial software landscape (led by Google/Facebook at the time)--ie rigid process-averse lower case agile. It turns the out of touch senile old guy was more prescient than the new suits.

Although DeMarco started deviating from his early utopian management approach when he published Peopleware, it wasn't until the late 20000s when he did a a complete 180 and wrote a mea culpa, disavowing his earlier management books and the "You can't control what you can't measure" quote attributed to him [1].

[0] https://www.cs.uni.edu/~wallingf/teaching/172/resources/dema...

> It turns out the out of touch senile old guy was more prescient than the new suits.

Well, it was Deming. He is still more prescient than the state of the art in management. Managers read his books, make an effort to internalize them, and immediately go and do the complete opposite thing.

I think it's correct about the backlash from Deming against the early DeMarco philosophy. As a young business builder back then when the overuse of metrics began to get popular, I always could see the benefit of basic metrics to get an idea about things that actually could be measured meaningfully, but already was aware that things which made the most difference were not the measurable ones.

Since then I doubled down on the unmeasurable (so I could make the most difference), even though I'm a measurement person in a measurement business. In the commercial analytical chemistry lab, with Deming's influence I was way ahead of the alternative labs who were not yet using control charts, nor were they attempting to produce NBS-traceable (now NIST-traceable) results which I considered essential since my results were utilized for some of the most high-stakes commodities. Most of the people in the established labs at the time had never even heard of an NBS-traceable approach. Continuous improvement is essential and over a period of 40 additional years you can make the kind of progress that can not be accomplished in a mere 30 years or so. Eventually you've got alternative labs so deep into the metrics that you can run circles around them without even using control charts yourself, just the internalized learnings over the years applied to the most meaningful elements which are largely unmeasurable if not completely misleading when you try.

A reductionist view might be to consider statistical process control to be helpful but statistical people control to be harmful.

>Managers read his books, make an effort to internalize them, and immediately go and do the complete opposite

In my case I made an effort to internalize then gradually was able to successfully guide the continuous improvement actions over a period of 10 years of less-bold moves as I was racheting up based on accumulating that most-valuable experience. Precisely to avoid immediately going and doing anything really dramatic in what would naturally be so less-informed a time as to end up having something like the opposite of the desired outcome.

That's pretty fascinating stuff, thanks for sharing.

Was there ever a situation were you were tempted to just go by the measurable metrics alone?

>tempted to just go by the measurable metrics alone?

For me that's where I start out, using the metrics that are already available, and often needing to measure a few more things that seem good at the time.

Hard data can tell you a lot, I was a mathematical child and I've been known to do an equation or two, even a statistic now and then. But that's not enough, there are quite a few mathematically gifted business operators who are capitalized with resources way beyond my reach, whom I have always needed to outperform in their capitalist market while being a lowly entrepreneur without any of other peoples' money to rely on. There would be no way to compete with them on their terms.

So the overall objective is to get the early metrics into the rear view mirror where it belongs, since it is intended to fade away to a lesser degree of prominence as real knowledge and strategy is built more deeply to go forward more effectively than measurement would predict no matter how accurate or well-regarded the hard data is.

The temptation is always there to rest on your laurels in a number of ways, but yielding to that is the halting of the continuous improvement process.

Even though absolute perfection may not be fully achievable, that's the direction you want to progress toward as much as you can. But no further progress can be made when you've halted for any reason.

You need to find your North Star for the particular situation and follow it unwaveringly so you can leave past performance in the dust with the data which was once defined by it.