The number was green. The reality was rotting

Why measuring transformation is harder than we admit — and how to do it honestly

Every project manager has a dashboard story. Mine goes like this: a system we’d launched showed adoption climbing week after week. Logins up, active users up, everything green. Leadership was pleased. Then someone actually talked to the branches, and the truth came out — people were logging in each morning because a manager had told them to, clicking around for a minute, and then going right back to the old spreadsheet. The metric was thriving. The transformation was dying.

I’ve come to believe that measuring change is one of the most underestimated skills in our profession. Not doing the change — knowing whether it’s actually happening. Because a badly designed KPI isn’t neutral. It doesn’t just fail to help; it actively misleads you, and it does so with the full authority of a number.

The law nobody teaches you

There’s a principle from economics that should be printed on every project charter. It’s called Goodhart’s Law, and the sharpest version goes:“When a measure becomes a target, it ceases to be a good measure.”

The mechanism is quietly devastating. You pick a number to represent a complex reality. You put it on a dashboard. Someone ties it to evaluation, funding, or a steering review. And then rational people start optimizing the number instead of the reality it was meant to reflect. One analyst who audited fourteen dashboards found that nine had at least one metric where the number improved while the thing it measured got worse — and the average time from “new target” to “first distortion” was just over four months.

We’ve all seen it. The support team measured on call-handling time that rushes customers off the phone. The sales team measured on call volume that makes dozens of pointless calls. The metric turns green. The underlying system rots. This is the trap I want us to learn to avoid — because in transformation, where we’re literally trying to change how people work, it’s everywhere.

The four ways a KPI betrays you

Over the years I’ve noticed that bad indicators tend to fail in the same four ways. Naming them helps me catch them early.

• The vanity metric: the number that only ever goes up and makes everyone feel good, but never drives a single decision. “Number of logins” is the classic. My test is simple: if a KPI can never be red, it isn’t a KPI — it’s decoration
• The metric disconnected from value: you measure what’s easy to count instead of what actually matters. “Training courses delivered” instead of “behaviors changed.” The first is trivial to report; the second is the entire point
• The proxy mistaken for the goal: you track support tickets as a proxy for adoption — reasonable — and then the team starts working to reduce tickets instead of improve adoption
• The metric with no threshold: a number without a baseline and a target says nothing. “We’re at 60%.” Good or terrible? You genuinely can’t tell without a reference point — and yet we present numbers like this in steering all the time

Guide with one hand, report with the other

If there’s a single idea worth taking away, it’s the distinction between two families of indicators — because using them correctly is most of the battle.

Leading indicators anticipate the result. Usage, behaviors, readiness. They move early and let you course-correct before the outcome is locked in — but they’re proxies, so they carry uncertainty. Lagging indicators confirm the result. Benefits, euros saved, FTE freed, quality. They’re reliable and objective, but by the time you read them, it’s too late to change them.

The mistake is choosing one camp. Revenue is a perfect lagging indicator, but knowing last month’s number can’t fix last month. The discipline is to hold both: guide daily with leading indicators, account for results with lagging ones. In practice this is exactly the logic of measuring adoption first — do people use it, and use it well? — and only later certifying the benefit realized. One hand on the wheel, the other on the ledger.

Building a KPI that survives contact with reality

So how do you design an indicator that doesn’t collapse the moment it becomes a target? A few habits have served me well.

• Start from the value, not the metric: ask the question before you pick the number. What business outcome are we really after? Only then derive what would prove it. This one reversal kills most vanity metrics at birth — you stop measuring what’s easy and start measuring what’s true
• Pair every leading indicator with the lagging one it predicts: if “% of cases handled end-to-end digitally” is your leading signal, tie it explicitly to the cost-and-time benefit it’s supposed to produce. When the two diverge, you’ve caught a Goodhart distortion in the act
• Anchor every number between a baseline and a target: progress isn’t “how high is the value” — it’s how far you’ve travelled from where you started toward where you want to be. Same number, completely different meaning
• Watch the denominator: a “95% completion rate” can mean 95 of 100 people, or 95 of the 100 who were still around at the halfway point. Half of measurement honesty lives in the denominator
• Never let a single metric carry a decision alone: the moment one number becomes the number, it starts to corrupt. Triangulate

The point I want to leave you with

Measurement feels like the objective, scientific part of our job — the part where we finally stop arguing and let the data speak. But the uncomfortable truth is that measurement is not passive observation. It’s an intervention. The act of choosing what to count changes how people behave.

That’s not a reason to measure less. It’s a reason to measure honestly — with humility about what a number can and cannot tell us, with the courage to show the seven metrics that are red and not just the three that are green, and with the wisdom to keep asking whether the reality behind the dashboard is genuinely improving.

Because the goal was never a green dashboard. The goal was a change that’s real. And the difference between the two is exactly where our judgment as project managers earns its keep.