How Metrics Influence Behavior

Page Summary

An explainer of how metrics influence behavior, opening with Wells Fargo's cross-selling targets driving employees to open around 2 million unauthorized accounts. It covers why metrics shape behavior (signaling, accountability, and crowding-out effects), Goodhart's Law and how a measure that becomes a target stops being a good measure, the gaming taxonomy and unintended consequences like attention-narrowing and short-termism, the difference between measurement and targets, how to prevent gaming (balanced metrics, human judgment, monitoring, iteration), and what makes metrics drive good behavior.

Contributions

Every accepted correction to this page is recorded with the exact change, so readers can see how the page improved over time.

1. 11 July 2026 · corrected by Emir Baycan

   2 corrections applied: The CFO study is by Graham, Harvey & Rajgopal (2005), not 'Campbell'. | The 2016 $185M action cited ~2 million accounts; the 3.5 million figure came from a later 2017 review.

   Before

   Research by Graham, Harvey, and Campbell found that 78 percent of CFOs...; Domain-examples table: '3.5 million unverified accounts; $185 million fine'

   After

   Research by Graham, Harvey, and Rajgopal found that 78 percent of CFOs...; Domain-examples table: '2 million unverified accounts; $185 million fine'

   Why: The Graham/Harvey/Rajgopal attribution and the main-paragraph Wells Fargo figure (2 million) were already correctly fixed in body content. But a SECOND body-text leftover survived in the Goodhart's Law domain-examples summary table, which still said '3.5 million unverified accounts; $185 million fine', inconsistent with the already-fixed paragraph above it. Fixed the table cell to '2 million unverified accounts', PUT, and verified clean on re-fetch. FAQ decoded (4-step op-sequence) and checked - no mention of these facts.

   View the full record →

Contributors In This Page