Negative Feedback
Negative feedback keeps systems stable and self-correcting.
Act, measure, adjust. If you don't measure the outcome, you're not learning.
Walk into a cold room. Your body temperature starts to drop. Thermoreceptors in your skin detect the change and send signals to your hypothalamus - a almond-sized region in your brain that acts as your body's thermostat. The hypothalamus triggers shivering (generating heat through muscle contractions) and vasoconstriction (reducing blood flow to your skin to conserve heat).
Your temperature rises. The thermoreceptors detect this. They signal the hypothalamus again: "Getting warmer." The hypothalamus dials back the shivering response.
This is negative feedback - the most common control mechanism in biology. When a system deviates from its setpoint, feedback pushes it back. Thermostats. Cruise control. Blood sugar regulation. Predator-prey populations. All negative feedback loops, all serving the same function: maintaining stability in an unstable world.
The mechanism is simple: sensor → control center → effector → sensor. The sensor detects change, the control center processes it, the effector acts to reverse the change, and the sensor confirms the system is returning to baseline.
Blood glucose illustrates this perfectly. After you eat, glucose floods into your bloodstream. Beta cells in your pancreas detect the rising concentration. They release insulin. Insulin tells muscle, fat, and liver cells to absorb glucose. Blood glucose falls. The beta cells detect this and reduce insulin secretion.
Clean. Precise. Self-correcting.
Business Application of Negative Feedback
Negative feedback keeps systems stable and self-correcting. In organizations, it means measuring outcomes and adjusting when results deviate from targets. Most companies do open-loop control - acting without measuring results. Netflix's content recommendation algorithm exemplifies closed-loop negative feedback: suggest → user watches (or doesn't) → log outcome → model updates → next suggestion adjusts. The loop must close for learning to occur.