Solve the Right Problem: Why KPIs Drift

Most KPIs don’t fail because they’re wrong.
They fail because they replace the objective.

KPIs are defined as measurable indicators of progress toward a desired result.

The logic is simple:

  • Identify the objective
  • Choose a metric
  • Track it
  • Improve it

Clean. Logical. Defensible.
Until it isn’t.


Where It Breaks

The first mistake is subtle:

A KPI is not the objective.
It’s a proxy for the objective.

And proxies drift.


A Simple Example

“I want more customer accounts.”
The KPI becomes: new accounts.

That assumes:
more accounts = growth

Not necessarily.

Growth can come from:

  • higher revenue per customer
  • better retention
  • increased usage

You can increase accounts and not grow at all.
You can grow accounts and damage the business.


When It Inverts

When the metric becomes the mission, things flip:

Instead of using a metric to understand reality,
reality gets shaped to improve the metric.

That’s when the system breaks.


The Wells Fargo Case

The KPI: number of new accounts
Employees were pressured to hit targets tied to compensation.

What happened:

  • millions of unauthorized accounts opened
  • funds moved without consent
  • customers charged fees

From a KPI perspective:

  • accounts increased
  • activity increased

From a business perspective:

  • trust collapsed
  • fraud increased
  • long-term growth declined

The KPI didn’t just fail.
It replaced the objective.


The Real Problem

Most discussions stop here and ask:
“What would have been a better KPI?”

That sounds reasonable.
It’s the wrong question.

The problem is not:
Which KPI should we use?

The problem is:
Why are we using a KPI to represent something more complex?


Why KPIs Drift

KPIs are attractive because they:

  • compress complexity into a single number
  • make performance visible
  • create the appearance of alignment

That compression is the risk.

When a number becomes the goal,
the system optimizes for the number—
not the outcome.


A More Useful Frame

Instead of asking:
What should we measure?

Ask:

  • What outcome actually matters?
  • What assumptions make this metric valid?
  • How could this metric be gamed?
  • What behaviors does it incentivize?
  • What would success look like without it?

If you can’t answer those clearly,
the KPI is leading you.


When KPIs Work

KPIs work when:

  • the system is well understood
  • the metric is tightly coupled to the outcome
  • feedback is fast
  • gaming the metric still improves the outcome

Most real-world systems don’t meet these conditions.


Final Thought

The issue isn’t that KPIs are bad.
It’s that they are substitutes.

And substitutes tend to take over.

This is not about KPIs.
It’s about what happens when a measurable proxy replaces a complex reality.

Once you see that pattern, it shows up everywhere.

Solve the right problem.

Leave a comment