🧠 Knowledge Base

Goodhart’s Law: When Metrics Become Targets

Focus
  • Metrics & Incentives
Category
  • Philosophy
Lens
  • Relational
Latest blog posts
A tired office worker sits late at night, resting his head on one hand as the unseen screen in front of him casts a cold blue light across his face.

The Politics of Productivity: Balancing Motion and Momentum

A tired woman sits alone at her desk in a dimly lit office, face illuminated by a computer monitor’s glow with red light in the background.

The Loyalty Loop: How Systems Use Belonging to Blind Us

A male office professional sits at his desk surrounded by glowing data dashboards, resting his head on his hand in quiet frustration as he stares at the screen.

Progress as a Product: Rethinking Digital Transformation

A woman in her mid-30s works from her kitchen table, surrounded by a laptop, tablet, and paperwork. She looks tense and overloaded, caught between multiple tasks during a video call.

The Meeting Maze: Time, Belonging, Agency

Focus
  • Metrics & Incentives
Category
  • Philosophy
Lens
  • Relational
Latest posts
A tired office worker sits late at night, resting his head on one hand as the unseen screen in front of him casts a cold blue light across his face.

The Politics of Productivity: Balancing Motion and Momentum

A tired woman sits alone at her desk in a dimly lit office, face illuminated by a computer monitor’s glow with red light in the background.

The Loyalty Loop: How Systems Use Belonging to Blind Us

Goodhart’s Law shows how targets distort the very systems they aim to improve. When institutions chase numbers, they optimise appearances rather than outcomes, leaving users trapped in cycles of dysfunctional incentives.
Explanation
What it is

Goodhart’s Law is a principle stating that when a measure is turned into a target, it loses reliability as a true indicator of progress. In practice, metrics stop reflecting reality once people and institutions begin optimising for them. The law highlights how numbers can become decoupled from the purpose they were designed to serve.

When to use it
  • When evaluating performance metrics tied to incentives.
  • When you see teams or institutions celebrating numbers without corresponding improvements in outcomes.
  • When diagnosing why reforms, audits, or dashboards aren’t producing lasting change.
  • When reviewing whether KPIs, OKRs, or benchmarks are still aligned with mission goals.
Why it matters

Goodhart’s Law matters because it exposes the danger of mistaking measurement for meaning. By showing how targets distort behaviour, it helps leaders and teams avoid the trap of “looking good on paper” while systemic issues remain unresolved. Applying the principle ensures metrics drive genuine outcomes rather than surface-level theatre, improving alignment, quality, and trust.

Definitions

Goodhart’s Law

The principle that when a measure becomes a target, it ceases to be a reliable indicator.

Metric

A quantitative measure used to track performance or progress.

Target

A benchmark or goal tied to incentives or accountability frameworks.

Proxy Measure

An indirect indicator used in place of a harder-to-measure outcome.

Campbell's Law

Related principle stating that the more a quantitative indicator is used for decision-making, the more it corrupts the process it monitors.

Measurement Theatre

Optimising for visible metrics while leaving underlying dysfunction unresolved.

Canonical Sources
  • Charles Goodhart, Problems of Monetary Management: The U.K. Experience, 1975 Bank of England paper (archival references)
  • K. Alec Chrystal & Paul D. Mizen, Goodhart’s Law: Its Origins, Meaning and Implications for Monetary Policy, 1999 ResearchGate
  • Michael F. Stumborg et al., Recognizing and Mitigating the Manipulation of Measures in Analysis, CNA, 2022 — CNA PDF
  • Michael Fire & Carlos Guestrin, Over-Optimization of Academic Publishing Metrics: Observing Goodhart’s Law in Action, 2018 — arXiv
  • Adrien Majka & El-Mahdi El-Mhamdi, The Strong, Weak and Benign Goodhart’s Law, 2025 — arXiv
  • Oliver Braganza, Proxyeconomics: The Inevitable Corruption of Proxy-Based Competition, 2018 — arXiv
  • Hilde Weerts, Lambèr Royakkers, Mykola Pechenizkiy, Are There Exceptions to Goodhart’s Law?, 2022 — arXiv
Internal Sources
Notes & Caveats
  • Scope: Goodhart’s Law is often applied beyond its economic origins into education, healthcare, AI, and management.
  • Misread: The law is sometimes overstated as “all metrics are useless” — in reality, it warns of distortion under pressure, not futility.
  • Controversy: Related formulations (e.g. Campbell’s Law) blur conceptual boundaries; academic debate continues over distinctions.
  • Versioning: Recent work distinguishes “Strong,” “Weak,” and “Benign” Goodhart effects, especially relevant in machine learning contexts.
Objective

Ensure metrics remain aligned with institutional purpose rather than collapsing into targets that distort behaviour. 

This guide helps teams and leaders detect when measures have become self-serving and restore their relational integrity.

Steps
  1. Map each metric to its North Star
    Tie every measure back to the institution’s stated mission. 
  2. Audit behaviours around the metric
    Within a set review cycle (e.g. quarterly), check if actions serve the number or the purpose. 
  3. Cross-check outcomes with stakeholders
    Validate results with users, customers, or frontline teams to see if improvement is real. 
  4. Retire or refresh distorted metrics
    Once evidence of gaming or drift emerges, phase out or replace the measure.
Tips
  • Use a basket of indicators rather than over-relying on one metric. 
  • Make the purpose of each measure explicit and visible. 
  • Pair quantitative measures with qualitative feedback for balance. 
  • Introduce sunset clauses: no metric should live forever without re-approval. 
  • Frame reviews as opportunities to learn, not witch hunts.

Pitfalls

Treating dashboards as reality

Always triangulate metrics with user or frontline experience.

Assuming objectivity of numbers

Remember every metric encodes subjective choices.

Adding more KPIs as a fix

Focus on fewer, purpose-anchored measures instead of KPI overload.

Ignoring unintended consequences

Run “pre-mortems” to explore how metrics might be gamed before rollout.

Acceptance criteria
  • Evidence that metrics track real outcomes, not just compliance with targets.
  • Dashboard or reporting artefacts updated to include purpose-link notes.
  • Stakeholders across teams confirm alignment between measures and mission.
  • Clear process exists for retiring or refreshing metrics that drift.
Scenario

A hospital management team is under pressure to reduce waiting times in its emergency department. Government oversight ties funding to meeting a four-hour patient handover target. Administrators must balance this metric with the real objective of providing safe, effective care.

Walkthrough

Decision Point

The leadership team decides whether to prioritise “hitting the four-hour target” or addressing underlying process inefficiencies.

Input/Output

Staff begin moving patients into short-stay wards earlier to clear the queue.

Input: waiting-time data.

Output: apparent compliance on dashboards.

Action

A new reporting sheet is logged into the hospital’s performance portal, showing compliance with the four-hour standard.

Error handling

Clinicians flag that some patients are being transferred prematurely, creating risks. Resolution: governance board reviews discrepancy between patient outcomes and reported compliance.

Closure

The board establishes a parallel metric tracking readmission rates and patient safety incidents, anchoring both back to the hospital’s mission statement.

Result
  • Before → After delta:
    • Before: Waiting times reported as compliant but at cost of patient safety.
    • After: Broader metrics adopted, balancing time targets with quality and trust.
  • Artefact snapshot:
    New *Performance Dashboard v2.0* stored in hospital intranet reporting hub.
Variations
  • If government regulations tighten, add parallel indicators that measure quality as well as speed.
  • If team size is smaller, prioritise qualitative spot-checks with patient interviews instead of complex dashboards.
  • If tooling differs (e.g., no integrated portal), use simple shared spreadsheets annotated with outcome notes to maintain transparency.