Goodhart's Law: When a Measure Becomes a Target in 2026
- What Goodhart's law actually says
- The common mistake: assuming people will optimize for intent
- A framework: how to set targets that don't collapse
- How to practice this
Goodhart's Law is a phenomenon familiar to many in management and performance measurement. At its core lies a simple truth: once a measure becomes a target, it ceases to be a good measure. This insight, originating from economist Charles Goodhart and popularized by anthropologist Marilyn Strathern, has profound implications for how organizations track performance and strive for improvement. In a world where metrics often dictate behavior, understanding and mitigating the pitfalls of Goodhart's Law has never been more crucial.
What Goodhart's Law Actually Says
Goodhart's Law emerged from discussions around monetary policy but applies broadly across various fields, including business performance. The law highlights that metrics, originally chosen as indicators of underlying realities—like customer satisfaction measured through Net Promoter Score (NPS) or code quality gauged by test coverage—become ineffective once they are tied to incentives or performance evaluations.
Consider a scenario where a customer support team is evaluated based on the number of tickets closed daily. Initially, this might seem like a reasonable metric, but over time, the team learns to close tickets quickly, often at the expense of quality. Customers may receive hasty, copy-pasted responses, leading to increased frustration despite the team meeting its targets. This divergence occurs because the focus shifts from genuinely resolving customer issues to merely hitting the numbers—demonstrating a classic case of Goodhart's Law in action.
The Common Mistake: Assuming People Will Optimize for Intent
A prevalent misconception is that team members will naturally align their actions with the underlying goals associated with a metric. Leaders often believe that the metric serves merely as a check-in point, with intrinsic motivation guiding the team to optimize for the desired outcomes. However, in reality, the metric becomes the de facto contract that shapes behavior. Employees, seeking promotions or bonuses, will prioritize the metric above the genuine intent behind it.
This leads to another common mistake: reacting to Goodhart's Law by abandoning measurement altogether. Such an overcorrection eliminates valuable feedback loops essential for learning and improvement. The solution lies not in forgoing metrics but in designing them thoughtfully to resist manipulation.
A Framework: How to Set Targets That Don't Collapse
To maintain the integrity of metrics and avoid the pitfalls of Goodhart's Law, organizations can adopt several practices:
Measure Outcomes, Not Outputs
Focus on measuring outcomes rather than outputs. For instance, instead of tracking the number of tickets closed, measure whether customer issues are resolved on the first contact, verified through follow-up surveys. Outcomes are more challenging to game because they involve external validation, whereas outputs are easily manipulated.
Pair Every Metric with a Counter-Metric
Implement counter-metrics to provide a balance. If you measure the speed of service delivery, also measure quality indicators such as customer satisfaction or retention rates. Counter-metrics help reveal when teams are gaming the system, ensuring that performance evaluations remain holistic.
Rotate Metrics
Goodhart's Law has a half-life. A metric that works well in the short term often becomes susceptible to manipulation over time. Regularly rotating metrics and introducing new ones can help maintain their effectiveness. This practice keeps teams honest and prevents complacency.
Keep Targets Directional, Not Point-Precise
Instead of setting specific numerical targets, provide directional guidance. For example, a team could be encouraged to improve NPS as a north star rather than aiming for an exact increase of 10 points. This approach makes it harder to game the system while still encouraging meaningful progress.
Separate Roles in Metric Design and Evaluation
To prevent gaming, separate the responsibilities of those who set the metrics, those who report on them, and those who are evaluated based on them. When the same individuals occupy these roles, the potential for structural gaming increases, leading to skewed results.
A Practical Example
Consider a software development team that uses "lines of code written" as a primary metric for productivity. Initially, this seems like a straightforward way to measure output. However, over time, the team starts prioritizing quantity over quality. Developers might rush to produce more code, leading to poorly designed features and increased technical debt.
To combat this, the team could instead focus on the impact of their work, such as user engagement or system performance metrics. By pairing the lines of code with user satisfaction scores and system reliability metrics, the team can better align their work with the organization’s goals. Regularly reviewing these metrics and rotating them every quarter can further ensure that the team remains focused on creating value rather than merely hitting arbitrary numbers.
Conclusion
Goodhart's Law serves as a critical reminder that metrics can easily become detached from their intended purpose once they are incentivized. By understanding this phenomenon and implementing strategies to design resilient metrics, organizations can maintain a focus on meaningful outcomes while still leveraging the benefits of measurement.
To improve your metric design and sidestep the pitfalls of Goodhart's Law, consider starting small. Assess one metric each month and apply the insights from this article to optimize its effectiveness.
Want to enhance your skills in metric design without overwhelming your schedule? Take the Omie Skill Assessment today and receive tailored lessons that fit your role and objectives.