A reading guide 11 articles
Systems and Incentives
When the same problem keeps returning, examine the connections, rewards and information behind the visible event.
A recurring problem may involve more than the person or event that first draws attention. Delays can amplify reactions. A target can change how people report results. Unequal information can make a deal difficult even when both sides would benefit from cooperation. This collection offers several ways to investigate those possibilities without assuming every bad outcome has one hidden cause.
Start with systems thinking when you need a view of the connections. Use the incentives route when a reward or metric appears to be shaping behavior. Use the information route when participants know different things. The full collection includes feedback loops, the iceberg model, power dynamics and comparisons with design thinking. Keep observations separate from your proposed explanation: drawing a plausible loop is a starting point for investigation, not proof that the loop caused the outcome.
Choose a starting point
-
Map the connections
Look beyond an isolated event and identify the relationships that may sustain a recurring pattern.
Systems Thinking: What MIT's Beer Game Teaches -
Examine a target or reward
Follow what changes when a measure becomes something people are rewarded for improving.
Goodhart's Law: How Wells Fargo's Best Metric Went Wrong -
Locate an information gap
Ask who knows what, and how signals, screening or incentives might change the interaction.
Information Asymmetry: Akerlof's Lemons and How to Fix It
All articles
11 in this collectionCobra Effect: How a Rat Bounty Ended Up Breeding Rats
In 1902, Hanoi paid for every rat tail. People cut tails off live rats and let them breed. Here is how the cobra effect works, and how to test a reward.
Feedback Loops in Systems Thinking: The SVB Bank Run
Feedback loops in systems thinking, through the 2023 Silicon Valley Bank run: how reinforcing and balancing loops work, why delays matter, and how to read one.
Follow the Incentives: What the Sears Auto Scandal Shows
In 1992 Sears oversold car repairs in 90% of undercover tests after tying pay to sales. See how to trace incentives, spot bad targets, and avoid cynicism.
Goodhart's Law: How Wells Fargo's Best Metric Went Wrong
Wells Fargo hit its cross-selling goal with millions of accounts customers may never have authorized. Here is how Goodhart's law works, and how to avoid it.
Iceberg Model Systems Thinking: Deepwater Horizon
The iceberg model in systems thinking, applied to Deepwater Horizon: one misread test, the near misses and structures beneath it, and how to dig into your own.
Information Asymmetry: Akerlof's Lemons and How to Fix It
Information asymmetry explained through Akerlof's rejected lemons paper: adverse selection, moral hazard, signals, screening, and a 5-minute audit for deals.
Power Dynamics: What Robert Moses Shows About Influence
Robert Moses never won an election yet ran New York's bridges for 34 years. See how tolls, bill drafting, and bonds gave him power, then map one decision.
Power Law Explained: What 1,000 Networks Revealed
A famous 1999 paper said most networks follow a power law; a 2019 test of nearly 1,000 found it rare. See density vs tail, log-log traps, and a 7-step check.
Red Queen Effect: Why Standing Still Means Falling Behind
Van Valen's fossil data showed old species die out as often as young ones. See how the Red Queen effect works in biology, banks, and careers, and what to do.
Systems Thinking vs Design Thinking: The PlayPump Lesson
Systems thinking vs design thinking via the PlayPump, a pump kids turned by playing: where each method would have caught its failure, and how to combine them.
Systems Thinking: What MIT's Beer Game Teaches
Systems thinking through MIT's beer game, where flat demand still caused wild orders: stocks, flows, delays, why blame misleads, and a six-step investigation.
Before you read
How do systems thinking and incentives fit together?
Systems thinking examines relationships across a situation. Incentives are one part of those relationships: rewards, penalties and constraints can shape responses. They are useful to investigate alongside delays, resources and information.
Which article helps with a metric that is producing strange behavior?
Start with Goodhart's law, then read the cobra effect and the essay on following incentives. Compare the measure's intended purpose with the behavior that is actually being rewarded.
Does every recurring problem require a complex systems map?
No. Begin with a small set of observed events and plausible connections. A detailed map is worthwhile only if it helps you test an explanation or choose an action. More arrows do not automatically mean better understanding.