Metrics · Goodhart's Law · Product Analytics · PTOS

Goodhart's Law in Product: Why Metrics Lie When Made a Goal

Explaining Goodhart's Law and its direct impact on product development: how optimizing for metrics can lead to product degradation.

Goodhart's Law in Product: How Not to Fall Victim to Metrics You Created Yourself

In product development, we are obsessed with metrics. They give us a sense of control and objectivity. But what if the very metrics we worship lead us astray? This is where Goodhart's Law comes into play.

What is Goodhart's Law?

British economist Charles Goodhart formulated it this way:

"When a measure becomes a target, it ceases to be a good measure."

In other words, as soon as you start optimizing a system for a specific indicator, people and systems begin to "hack" that indicator rather than improve the reality it was supposed to reflect. The number in the report grows, but the product or process degrades.

Goodhart's Law in Product Development

This phenomenon is ubiquitous. Teams begin to optimize a number, not real value for the user.

• Goal: "Increase time spent in the app"

  • How it's "hacked": They create a confusing interface, forcing the user to spend more time searching for the desired function. They hide the "unsubscribe" button. They introduce an infinite feed that causes addiction but brings no satisfaction.
  • Result: The metric grows, but user frustration also grows. Retention falls in the long run.

• Goal: "Increase the number of reports created"

  • How it's "hacked": They encourage the creation of empty or meaningless reports. They introduce KPIs for managers, forcing them to "proliferate" entities.
  • Result: The number on the dashboard looks good, but there's no real benefit from these reports. The system fills with junk.

• Goal: "Increase DAU (Daily Active Users)"

  • How it's "hacked": They begin to bombard users with push notifications and emails with clickbait headlines, forcing them to enter the app for at least a second.
  • Result: DAU grows, but these are "empty" visits. Engagement and depth of use fall. Users turn off notifications or delete the app.

How to Protect Against Goodhart's Law?

It's impossible to completely avoid this effect, but it can and must be controlled.

1. Focus on the Value Event, not proxies.

   • Measure not just activity (clicks, views, time), but an action that proves the user received real value. Not "created a document," but "shared a document with a colleague, and they commented on it."

2. Always use Guardrail metrics (protective metrics).

   • Any target metric (North Star Metric) must have a "guardrail." Ask yourself before starting: "How can we 'game' our main metric while simultaneously harming the product?" The answer to this question is your list of guardrail metrics.
   • Example:
     • North Star: "Number of messages sent."
     • Guardrail: "Percentage of messages marked as spam," "Number of blocked users."

3. Use qualitative and quantitative data together.

   • Numbers tell you "what," but not "why." Metric growth can have completely different reasons. Combine analytics data with interviews, user feedback, and support tickets to understand the full picture.

4. Regularly review your metrics.

   • Metrics are not set in stone. The market changes, the product changes, user behavior changes. What was a good metric a year ago may be leading you down a blind alley today.

Conclusion: Metrics are a powerful tool, but not an idol. Treat them as hypotheses, not absolute truth. Your true goal is not to "move a number," but to solve a real user problem. And Goodhart's Law is a painful but important reminder of this.