Wednesday, September 13, 2020

Occam's Razor

Browsing around the Internet, I came across a succinct description of Occam’s Razor on the site SkepticReport.com. Here’s part of what it said:
Skeptics frequently talk about Occam's Razor. They use it to choose between alternative explanations for something, especially where no one alternative has been either proven or disproven. . . . Many people will tell you it says, "Choose the simplest solution". But it doesn't say choose the simplest solution. Opponents of Occam complain that it will not necessarily help you choose the correct solution. But Occam's Razor does not pretend to choose the correct solution. So what is it and what is its point?

Occam's Razor actually says: "Pluralitas non est ponenda sine neccesitate", which is [literally] translated as "plurality should not be posited without necessity." The words are those of the medieval English philosopher and Franciscan monk William of Ockham (ca. 1285-1349). The archaic English needs to be interpreted for modern times. What it means is this: Do not invent unnecessary entities to explain something.
I’d like to suggest an updating of William of Ockham’s idea to apply to business and leaders. Let’s call it Carmine Coyote’s Cutthroat:

“Do not invent unnecessary measurements and statistics to manage anything.”

I’ll start with setting objectives by statistical means: something that is so common as to go virtually unchallenged. Let’s challenge it, shall we?

Take an imaginary company whose sales for one product last quarter were $50 million. Their objective for the next quarter is a 5% increase to $52.5 million. So far so good. Another quarter passes, and the powers that be decide to be “kind” and keep their target the same (a 5% increase). No one considers that the base figure has increased (to $52.5 million). Five percent of that is $2.625 million, not the $2.5 million of the previous quarter. Now that product’s sales people have to find an extra $0.125 million in sales over and above the last quarter’s increase. They haven’t stood still. Their target has increased. If this continues, the next quarter’s target will be $2.756 (5% of $55.125 million). After one year of “standing still,” the quarterly target will be $3.04 million, compared to $2.5 million at the start. From there, the laws of compounding will quickly raise each new quarter’s target to stratospheric levels.

Many years ago, I discovered that the questions in math tests for college graduates most likely to produce wrong answers were any that included percentages. That still holds good.
Too often, those in charge look only at the percentage figure. Their reasoning is typically that if you made a 5% increase last quarter, you should increase that percentage this time around—maybe to 6% or more. You can imagine the effect on actual target amounts. But here is the kicker: those in charge, looking just at percentages and ignoring the effect of compounding, see such targets as wholly realistic. What’s a measly increase from 5 to 6? (For a start, it’s 20%, but don’t let reality or mathematical accuracy bother you.) Many years ago, I discovered that the questions in math tests for college graduates most likely to produce wrong answers were any that included percentages. That still holds good.

As a result of this kind of statistical idiocy, people find themselves facing impossible goals, even as their organizations raise market expectations that they are wholly incapable of fulfilling, however hard they drive their people forward. What’s even worse is that people truly do try their hardest to meet the objectives set for them. As W. Edwards Deming said, people with sharp enough targets will probably meet them even if they have to destroy the company to do so. He could also have added that they will likely destroy their own health and well-being at the same time.

A good part of the overwork and pressure that infects business today comes from people either collecting data to satisfy the organizational mania for measurement, or facing objectives that have been produced by the statistically illiterate.
Demming’s basic message was that quality is a management responsibility, and poor quality is almost always the result of imposing systems and objectives which thwart people’s desire to do high quality work. He also said the role of management was to drive out fear, something best done by eliminating quotas, numerical goals, and merit ratings. A good part of the overwork and pressure that infects business today comes from people either collecting data to satisfy the organizational mania for measurement, or facing objectives that have been produced by the statistically illiterate.

The mania for measurement wouldn’t be quite so bad were it not for the sad fact that most people involved neither understand the point of the figures nor are able to deduce anything useful from them. Many of these solemnly-reported figures are like the “statistics” beloved of baseball commentators: “That’s only the twelfth time since 1907 that the third player in a visiting team from south of the Mason-Dixon line has struck the ball on the first pitch so that it landed exactly 20 feet to the right of a woman sitting on a red cushion. Amazing!”

No statistical objective will matter if it defies common sense or flies in the face of reality.
Remember Occam’s Razor. Don’t invent unnecessary “entities” to explain what is more simply understood by using existing concepts, such as chance. Then add Carmine Coyote’s Cutthroat and stop inventing unnecessary measurements and statistics to manage or lead. As Simon Caulkin wrote recently in the Observer (a British Sunday newspaper):
Despite management's obsession with hard numbers, many organisations are a fact-free zone, swirling with untested assumptions. Horrifying sums of money are committed on superstition or whim. . . . As Peter Drucker says: “Thinking is very hard work. And management fashions are a great substitute for thinking.”
No measurement can be a substitute for hard thinking and taking your time to work out what may truly be the underlying causes of the problems you need to deal with. No statistical objective will matter if it defies common sense or flies in the face of reality. And no amount of performance management, incentivization, or bullying leadership will save an organization where those in charge are, in Caulkin’s words: “. . . in love with claptrap and blinded by ideologies.”


Stumble Upon Toolbar

Comments:
I totally agree. It sometimes feels like once the big leap to management takes place, number crunching skills get filed away along with all the other technical skills they may have had to begin with. I firmly believe that the deadlines that organizations set today are the reason that killer apps die before they're ever made public. Thank you for this blog, I enjoyed it completely.

Dan
 
Thanks, Dan. I'm glad you found it useful. Keep reading, won't you?
 
"the most important figures that one needs for management are unknown or unknowable Deming quoting: Lloyd S. Nelson" for more see: manage what you can't messure. Deming management posts from the Curious Cat Management Improvement Blog.
 
Thanks, CuriousCat. That's a most helpful link.
 
Post a Comment



<< Home
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 2.5 License.

This page is powered by Blogger. Isn't yours?