Designing Metrics That Drive Operational Excellence

Recently, I wrote about the illusion of visibility in a piece called “Your Metrics Are Too Pretty: Continuous Improvement Fails When Metrics Don’t Drive Action”.

I once visited a plant where every board was full of metrics, bar charts, trend lines, color-coded graphs. It looked great.

But when I asked a supervisor what one of the KPIs meant, he said: “Honestly, I have no idea. We just update it every day.”

That’s not performance management. That’s dashboard decoration, and it’s one of the biggest barriers to operational excellence.

Why Most Operational Excellence Metrics Don’t Work

Most companies don’t suffer from a lack of data. They suffer from a lack of clarity. Metrics get designed for reporting, not behavior. They become numeric wallpaper, visible, but useless.

If your frontline team can’t explain:

• What the metric measures
• What good looks like
• What to do when it’s off

…then it’s not helping you improve. It’s standing in the way of continuous improvement.

Cowboy Control: A Better Way to Think About Metrics

There’s an old graphic from Taiichi Ohno that shows a group of cows in a pen, with one cow standing outside the fence. The idea is simple: when everything is working, most cows (or processes) stay within the boundaries. The only one you need to focus on is the one outside, the one that’s abnormal.

That’s lean visual management in action: define what normal looks like, make deviations visible, and respond quickly and decisively.

But most companies take the opposite approach. They look at every cow, every day, analyzing all of them equally, even the ones behaving perfectly. They create massive dashboards, packed with data, and review everything with the same intensity. The result? Information overload. Teams lose the signal in the noise.

I call it cowboy control. Not because it’s reckless, but because it looks busy without being effective. The trick isn’t to stare at every number. It’s to build a system where the abnormal jumps out, and your people are trained to respond. You define normal, then act when something deviates. 

Metrics That Confuse, And Why

I once worked with a client who measured productivity as:

(Direct Hours ÷ Available Hours) x (Earned Hours ÷ Direct Hours) ÷ (Available Hours ÷ Total Hours)

Nobody could explain what it meant. Not even the person in charge of the chart. But it was reviewed daily.

Other examples I’ve seen:

• KPI lines with no targets
• Yield rates plotted with no threshold for action
• Cost ratios nobody can tie to a decision

If it takes a calculator or a spreadsheet to understand your metric,  it’s not driving behavior.

A Better Way: What Good Operational Excellence Metrics Look Like

The best operational excellence metrics are always:

• Simple: Easily explained without a cheat sheet
• Actionable: Clear response when outside the norm
• Visible: Seen daily at the point of use
• Behavior-linked: Reinforce the right habits

Think of a car dashboard:

• Fuel low? Fill up.
• Temperature high? Slow down.
• Door open? Close it.

Same in operations. Metrics should say: here’s the problem, now act.

What I Recommend to Drive Operational Excellence

To design metrics that actually help your team:

• Involve the people who use them
• Tie every metric to a behavior or decision
• Use targets and thresholds (not just trend lines)
• Limit dashboards to what matters, and retire what doesn’t
• Audit comprehension: Ask “What would you do if this chart went red?”
• Assure the team can impact the metric

The Bottom Line: Operational Excellence Starts With Clarity

Data isn’t power. Clarity is.

Operational excellence and continuous improvement both depend on timely, visible, and actionable information. If your metrics don’t drive behavior, they’re not helping. They’re hiding.

Make your metrics simple. Make them visible. Make them matter.