…to see if you see what they see

My sweetie and I have assembled a lot of IKEA furniture renovating a farm house kitchen and all. The first time we did it, she lay all the bits and pieces out on the floor. I don’t mean just in piles. Each nut, bolt, fastener, pin, shelf, leg and panel was laid out and aligned in a grid.

It made her happy, which made me happy. I didn’t think much of it at the time. I’m a simple man.

Years later I ran into Deming’s Red Bead experiment. For those who aren’t familiar, he rigged a pretend ‘manufacturing’ system with arbitrary results, and then applied the usual metrics and management techniques to the outcomes, as if that would make a difference. Watching the faces of the bemused participants is fun, I think, because they can see what’s happening, or it’s happened to them before, or they caused it to happen to others.

They’re being managed and measured against a system over which they have no control nor influence. They’re being praised or berated depending on random outcomes for processes that don’t matter and won’t change the outcome. Does this sound familiar?

Everyone standing around the table knows what the solution is: take the time to sort red beads out of the bins, then move the white ones where they need to go. Or maybe don’t put red beads in the bin to start. That way the ‘defective’ red beads don’t make it out to the ‘shipping dock’ and then on to the customer.

Nobody says anything, especially not the ‘leader’, who blissfully carries on collecting statistics, firing poor performers, praising good performers, and hilariously coaching his workers on their scooping technique.

It’s a bit of absurdist management theatre that highlights several important points:

Deming’s Core Ideas

Deming used the exercise to illustrate several of his core ideas: workers can’t produce results the system isn’t capable of producing; if you want fewer defects change the system; ranking and rewarding workers based on certain data is meaningless and demoralizing; management theatre has no effect on the outcome when the cause of variation is systemic.

All important , but the three most important to those who want to improve their systems are:

Targets Without Methods Are Meaningless

The lap time for a race-car is just the start of what makes a race car fast, understanding what goes into that time (target), and being disciplined and curious about how to effect real change over things you can control (engineering, driving) is how you get better results.

https://practicalmanagers.com/2026/05/23/the-kpi-is-not-the-work/

You should know your process deeply enough to help your team improve it rather than just monitoring the numbers. This is good leadership.

Know the Difference

There’s a difference between built-in, systemic variation and when something unusual happens. Do you count the lap time if the car was forced off the road by someone else’s crash? Probably not.

Over-reacting to special cases is not helpful and counter-productive, even more than under-reacting to systemic issues is not helpful. It seems obvious: freaking out about stuff you have no control over, or that doesn’t matter, is poor leadership. Knowing the different and acting accordingly is good leadership.

Slow is Smooth, Smooth is Fast

It doesn’t matter whether you’re running a race team, sales team, a construction team, design team, or building IKEA furniture. Sometimes you’re up against a deadline, and you know if you don’t stop and change how you’re doing things, it’s going to be even worse.

Sometimes you need to stop, reset, pick out the red beads, or sort the bolts, sharpen the chainsaw before you cut a hole in the wall for the new kitchen windows. It’ll go faster. This is also leadership.

Doing Better

Deming was critical of Drucker’s Management by Objectives (MBO). He felt MBO as often practiced was exactly the wrong approach, because it focused management on the numerical output instead of understanding and improving the system that produced that number. It gave managers the illusion of control without the substance.

https://practicalmanagers.com/2026/04/11/smart-goals-what-are-they-good-for/

Most modern performance management systems have that kind of numerical target logic baked in, or it gets added in, because “you can’t manage what you can’t measure”, even though that’s not always true, or we often measure the wrong thing.

https://practicalmanagers.com/2026/05/16/showdown-at-the-okr-corral/

But if you don’t manage using measures, how do you know if things are getting better? Deming argued that you need to understand your processes statistically, track trends and variation over time, and manage the system as a whole, not the scoreboard.

Which is hard work, but leadership.