Goodhart's Law in Sales

“The moment a metric becomes a target, it stops being useful.”

This shows up everywhere in sales organizations.

You introduce a metric to measure performance…
And the team starts optimizing for the metric—not the outcome.

Pipeline coverage becomes inflated deals.
Activity targets become low-quality outreach.
Forecast pressure leads to sandbagging.

The metric didn’t fail.

The system around it did.

Here’s where it gets more subtle—and more important:

Most sales organizations try to solve this by defining a “standard model” of success.

A healthy pipeline looks like this.
A good rep makes this many calls.
A strong funnel converts at these rates.

But that’s not how real sales teams work.

I’ve seen reps succeed in completely different ways.

One rep runs high-volume outbound:

• Lots of activity

• Smaller deals

• Lower conversion rates

• Still hits (or exceeds) quota

Another rep is highly targeted:

• Fewer touches

• Higher conversion rates

• Larger deal sizes

• Same (or better) outcome

Both are right.

The problem isn’t metrics.

It’s forcing everyone into the same ones.

Instead, I’ve found it far more effective to make metrics personal and diagnostic.

Sit down with each rep and understand:

• Average deal size

• Close rate

• Conversion by stage

• Pipeline velocity

From there, you can work backward:

👉 What does your pipeline need to look like?
👉 What level of activity actually supports your number?

This does two things:

1. It aligns metrics to real outcomes

2. It allows the rep to be a partner in the design of their work plan

3. It opens the door to meaningful coaching

Because now the conversation isn’t: “Make more calls”

It’s: “If we improve your close rate, you may not need to make half as many calls.”

That’s a very different—and much more motivating—conversation.

Goodhart’s Law doesn’t mean metrics are useless.

It means badly designed systems make them useless.

If your team is hitting the metrics but missing the business…

You don’t have a performance problem.

You have a measurement problem.