Keeping It Simple for Measurements

Make everything as simple as possible, but not simpler — Albert Einstein

One of the most insidious tendencies we managers and leaders have is to make things more complicated than they absolutely need to be. We tend to forget that giving simple, clear direction gives rise to intelligent behaviour. Complex rules and regulations, however, gives rise to simple and stupid behaviour.*

This tendency can really get out of control when we try to come up with measures that we can use to keep the pulse of our businesses and projects. Either we come up with many measures (trying to make sure we’re not missing something I suppose, which means we’re spending all our time analyzing measure and end up missing the part where we actually run the business). Or we come up with measures that don’t mean anything to anybody but a select few in an ivory tower (“$x net booked earning before interest and taxes this fiscal quarter”)

The best metrics are simple, few and understandable by everybody. Everybody in the company or project understands what it means and how they contribute to it. They give focusĀ  and energy to the organization’s execution, because people love seeing their progress measured.

This story illustrates my point:

Charles Schwab had a mill manager whose people weren’t producing their quota of work. “How is it,” Schwab asked him, “that a manager as capable as you can’t make this mill turn out what it should?”

“I don’t know,” the manager replied. “I’ve coaxed the men, I’ve pushed them, I’ve sworn and cussed, I’ve threatened them with damnation and being fired. But nothing works. They just won’t produce.”

This conversation took place at the end of the day, just before the night shift came on. Schwab asked the manager for a piece of chalk, then, turning to the nearest man, asked: “How many heats did your shift make today?”


Without another word, Schwab chalked a big figure six on the floor, and walked away.

When the night shift came in, they saw the “6” and asked what it meant.

“The big boss was in here today,” the day people said. “He asked us how many heats we made, and we told him six. He chalked it down on the floor.”

The next morning Schwab walked through the mill again. The night shift had rubbed out “6” and replaced it with a big “7.”

When the day shift reported for work the next morning, they saw a big “7” chalked on the floor. So the night shift thought they were better than the day shift did they? Well, they would show the night shift a thing or two. The crew pitched in with enthusiasm, and when they quit that night, they left behind them an enormous, swaggering “10.” Things were stepping up.

Shortly, this mill, which had been lagging way behind in production, was turning out more work than any other mill in the plant.**

Next time you need to come up with a way to motivate a team, project, division, or company, find the metric that everybody can understand and get behind. Being clear on the values and purpose of that team, big or small. will make it easier to find that number. Then let that be progress visible to everybody, and keep it up-to-date.

What gets measured gets done – just be careful what you measure.

* “Simple, clear purpose and principles give rise to complex and intelligent behaviour. Complex rules and regulations give rise to simple and stupid behaviour.” — Dee Hock
** From HBR’s “Why Keeping Score is the Best Way to Get Ahead*

3 thoughts on “Keeping It Simple for Measurements

  1. I’m basically in agreement with the KISS approach of presenting metrics, but coming out of the software production world, the complexities and “art” of the software life-cycle can be daunting to measure (think 6-Sigma), and once you are in that mind-set, it can be difficult to boil it down to simple outputs. Some managers may demand that the people reading the reports be at least conversant with the metric gathering and analysis techniques. That is not as simple as writing “6” on the floor, although I get your metaphorical example.
    To support your point though, here is a practical implementation that I have seen. A now retired merchant friend of mine would sit in the mall and count the number of people entering a popular store, and the number of people coming out with packages. He would then trend this sample with previous observations to predict how his business would do in the coming months. Simple and effective. -But I hope he didn’t forget about the “uncontrolled variables” like day-of-the-week or new staff in the store <8}
    Cheers and keep the information flow coming.
    Stu Cox


  2. Thanks Stu,

    Ah yes. The “It’s too complicated to measure”, or “You can’t measure x” where x is happiness, or effectiveness, or something intangible. I call B.S.

    You’re absolutely right that having the wrong measure can be just as bad or even worse than having no measure. I can think of a recent example from an engineering context where the measure was the number hardware changes to a particular design. More changes meant that either 1) the design was progressing nicely (good in the middle of the design phase) or 2) the design was churning (bad as the design deadline approaches).

    The problem was the measure was *net* changes (additions less subtractions). The meant lots of churn could be going on, but the actual number presented for analysis was zero. Totally useless and misleading.

    I would argue that having no measure is a failure of imagination and leadership. Especially in software where you can write code to measure just about anything, and you have archives of code to experiment with.

    The problem is not having a measure, it’s with picking the right one. Whether it’s cyclic complexity as a proxy of maintainability, or lines-of-code as a proxy for estimating effort and completeness, or function points as a proxy for programmer effectiveness and efficiency. There’s lots to choose from. The danger here is choosing too many and being overwhelmed with data.

    Top performers like the competition, and they strive for measurable feedback that shows progress. When I hear the excuse that we can’t measure something because it’s too complicated, or too subjective, or too something something I immediately think “What are they hiding?”

    Even if our first attempt at measuring something is wrong we can learn something from it. We need to do what we can with what we have where we are right now and stop making excuses. Measure something, figure out if it works or not, then change it if necessary. Just get on with it.

    Disclaimer: Stu and I used to work together, and it’s not him I’m hectoring. There was a team lead we both know who put off measuring progress for almost two years by “studying the problem” that I had in mind when I wrote this.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s