Feedback Loops
mental-model
Source: Physics
Categories: systems-thinkingorganizational-behavior
From: Poor Charlie's Almanack
Transfers
Control systems engineering — where a system’s output is routed back as input, either amplifying (positive feedback) or dampening (negative feedback) the signal — mapped onto business, social, and biological dynamics. The metaphor makes invisible causal loops visible and shifts attention from events to the structures that produce them.
Key structural parallels:
- Circular causation replaces root-cause thinking — in a feedback loop, asking “what caused X?” has no terminal answer because X partly caused itself through prior iterations. The cognitive move is to stop searching for a single root cause and start tracing the circuit. Instead of fixing a cause, you modify the loop.
- Positive feedback amplifies — a microphone pointed at its speaker creates a screech; each cycle amplifies the previous signal. In business, a popular product attracts more users, which attracts more developers, which makes the product better, which attracts more users. Understanding that you are inside a positive feedback loop changes your strategy from “push harder” to “don’t break the loop.”
- Negative feedback stabilizes — a thermostat detects temperature deviation and corrects it. In organizations, budgets, performance reviews, and quality controls serve as negative feedback: they detect deviation from target and trigger corrective action. The metaphor clarifies that negative feedback is not criticism — it is the mechanism that keeps systems near their target state.
- Delays cause oscillation — in engineering, feedback with a time delay causes overshoot and oscillation. The same happens in economics (the cobweb model: farmers see high prices, plant too much, prices crash, they plant too little) and in organizational behavior (hiring in response to last quarter’s demand, not next quarter’s). The metaphor explains why well-intentioned corrections often make things worse: the delay between action and effect means you are always correcting for the past.
- Loop dominance shifts — complex systems contain multiple feedback loops. Which loop dominates determines the system’s behavior. A startup might have a positive growth loop (more users, more revenue, more investment) and a negative quality loop (more users, more bugs, worse experience, fewer users). Understanding which loop currently dominates — and what might shift dominance — is the central analytical task.
Limits
The feedback loop model is one of the most useful analytical tools available, which makes its failure modes especially important to understand.
- Not everything is a loop — the metaphor encourages seeing feedback everywhere, including where it does not exist. Some cause-and-effect chains are genuinely linear: a meteor strike causes an extinction event without a feedback mechanism. Forcing every phenomenon into a loop structure can obscure simple causation.
- Loop identification is subjective — drawing the boundary of a feedback loop requires choosing what to include and exclude. Different analysts draw different loops from the same data. The metaphor implies the loops are out there waiting to be discovered, when in fact they are models imposed on reality. The map-territory problem applies: the loop diagram is not the system.
- Positive feedback sounds positive — the engineering terminology is confusing. “Positive feedback” means amplifying, not “good.” A bank run is positive feedback (fear causes withdrawals, which cause more fear). An arms race is positive feedback. The terminology smuggles in a valence that the engineering concept does not have, leading to consistent misapplication in business contexts where “positive feedback loop” is used as if it always means virtuous cycle.
- Oversimplification of delays — real systems have multiple delays of varying lengths interacting simultaneously. The simple “delay causes oscillation” model works for thermostats but breaks down for economies, ecosystems, and organizations where dozens of delays interact in nonlinear ways. The metaphor can create false confidence that understanding one delay gives you control of the system.
- Agency disappears — feedback loop diagrams show variables influencing other variables. People disappear from the picture, replaced by arrows and stocks. This can be analytically useful (removing personality from structural analysis) but also dehumanizing: it treats human behavior as mechanistic response to system variables rather than choice. “The system produces this outcome” can become an excuse for individual inaction.
Expressions
- “Virtuous cycle” / “vicious cycle” — the colloquial terms for positive feedback loops with desirable and undesirable outcomes
- “Flywheel effect” — Jim Collins’s metaphor for a business positive feedback loop that builds momentum
- “Doom loop” — Collins’s term for a negative spiral (technically also positive feedback, but toward destruction)
- “Runaway effect” — positive feedback without a natural check
- “Self-correcting mechanism” — negative feedback in institutional contexts
- “Checks and balances” — the political version of engineered negative feedback
- “Death spiral” — a positive feedback loop heading toward collapse (insurance, airlines, organizations)
- “The rich get richer” — the colloquial description of wealth as a positive feedback loop
Origin Story
Feedback as an engineering concept dates to James Watt’s centrifugal governor (1788), which regulated steam engine speed through mechanical negative feedback. Norbert Wiener formalized the concept in Cybernetics (1948), extending it from machines to biological and social systems. Jay Forrester applied feedback loop analysis to business and urban systems at MIT in the 1960s, creating the field of system dynamics. Peter Senge popularized the framework for business audiences in The Fifth Discipline (1990), making “systems thinking” and feedback loop diagrams standard tools in management consulting. Munger adopted feedback loops as a core mental model, emphasizing that understanding which type of loop is operating — amplifying or stabilizing — is essential for investment analysis and life decisions.
References
- Wiener, N. Cybernetics (1948)
- Forrester, J. Industrial Dynamics (1961)
- Meadows, D. Thinking in Systems (2008)
- Senge, P. The Fifth Discipline (1990)
- Munger, C. Poor Charlie’s Almanack (2005)
Related Entries
Structural Neighbors
Entries from different domains that share structural shape. Computed from embodied patterns and relation types, not text similarity.
- Gambler's Fallacy (probability/mental-model)
- Ouroboros (mythology/archetype)
- Nemesis (mythology/metaphor)
- Resilience (resilience/mental-model)
- OODA Loop (military-command/mental-model)
- Sharpening the Saw (tool-use/metaphor)
- Hansei (manufacturing/mental-model)
- Antifragile (resilience/mental-model)
Structural Tags
Patterns: iterationflowforce
Relations: causerestore
Structure: cycleequilibrium Level: generic
Contributors: agent:metaphorex-miner