A feedback loop is simple: a system's output loops back and changes its own input. The thermostat is the classic example. Room gets cold → heater turns on → room warms up → heater turns off. The output (room temperature) feeds back into the input (heater trigger). That's a balancing loop — it resists change and pushes toward a target.
The other kind is a reinforcing loop — also called a positive feedback loop, though "positive" doesn't mean good. It means the loop amplifies. Output feeds back to make more output. A microphone too close to a speaker: sound enters mic → amplified through speaker → mic picks up more sound → louder → screech. That's reinforcing feedback.
Both types are everywhere. Here are five you've already lived through today.
Balancing loop: resists change, seeks equilibrium. Examples: thermostats, population limits, hunger and eating.
Reinforcing loop: amplifies change, accelerates away from equilibrium. Examples: compound interest, viral spread, market bubbles.
You post something. It gets three likes in the first five minutes. The algorithm notices early engagement and shows it to more people. More people see it, more like it. The algorithm shows it to even more people. Your post goes from 3 likes to 300.
Now flip it. You post something else. Zero engagement in the first hour. The algorithm buries it. Nobody sees it, so nobody engages. You get zero likes. Two posts with roughly equal quality — wildly different outcomes because of how the reinforcing loop operates in the first window.
This is why social media feels simultaneously random and rigged. It is both. The initial signal is somewhat random (who happens to be online, what else is competing for attention). But once the loop kicks in, outcomes compound fast. The rich get richer. The ignored get buried.
Early engagement → algorithm amplifies reach → more engagement → algorithm amplifies more. A reinforcing loop. The input (engagement) feeds back to increase the output (reach), which feeds back to increase the input. Unbounded — until you hit audience limits or the algorithm caps organic reach to push paid promotion.
Home prices rise. Investors buy more properties as an investment. More buyers, limited supply — prices rise further. News covers the price gains. More buyers enter the market hoping to catch the wave. Prices accelerate. Every headline about rising prices brings in more buyers. The loop runs until it doesn't.
Then it reverses. Prices plateau. Some sellers list to lock in gains. More supply hits the market. Prices soften. Buyers wait, expecting further drops. Fewer buyers, more supply — prices fall. News covers the price drop. More sellers rush to exit before it gets worse. Prices accelerate downward.
The housing market isn't a price discovery mechanism. It's a sentiment loop. Expectations become self-fulfilling. Rising prices create the belief that prices will keep rising, which makes them rise. Falling prices create the belief that prices will keep falling, which makes them fall. The fundamentals (jobs, population, construction costs) matter — but they operate slowly. The feedback loop operates fast.
This one has the longest time lag of any feedback loop most people will encounter in their lifetime, which is exactly why it's so dangerous.
Arctic sea ice is white. It reflects sunlight back into space. As temperatures rise, sea ice melts. Less white ice means less reflected sunlight. The ocean absorbs more heat instead. Ocean warms further. More ice melts. Less reflection. More warming. This is the ice-albedo feedback — a reinforcing loop built into the climate system.
Permafrost is another one. Permafrost stores massive amounts of carbon locked in frozen ground. As temperatures rise, permafrost thaws. Thawing releases CO₂ and methane into the atmosphere. Greenhouse gases rise. Temperatures rise. More permafrost thaws. More emissions. More warming.
Once a reinforcing climate loop passes a certain threshold, the loop self-sustains even if humans stop adding emissions entirely. The system has enough momentum to continue amplifying on its own. This is the definition of a tipping point — not a gradual slide but a shift into a different stable state. Systems thinking is the framework for understanding why this is different from linear cause-and-effect.
You avoid a task because it feels overwhelming. Avoiding it gives you a few minutes of relief. But the task is still there, and now you're behind on it. It feels even more overwhelming. You avoid it more. Now you're further behind. The deadline pressure spikes. The overwhelm becomes genuine panic. You avoid it even harder. You do anything but the task.
This is a reinforcing loop, and it has a name in psychology: avoidance reinforcement. The behavior (avoidance) is reinforced by the short-term reward (relief), while the long-term consequence (accumulating pressure) makes the original behavior more likely, not less.
The balancing loop that should counteract this — task completion providing relief — is delayed. You have to actually do the task before you get the reward. The procrastination loop provides relief immediately, which is why it dominates.
Both loops offer relief. Avoidance gives it now. Completion gives it later. Brains discount future rewards — a well-documented quirk called hyperbolic discounting. The procrastination loop wins because its payoff is faster, not because it's stronger. Understanding this changes the intervention: the goal isn't to feel more motivated, it's to make the completion reward more immediate (starting, not finishing, is what breaks the loop).
You go for a run. You feel good afterward — endorphins, accomplishment, better sleep that night. The next day, you feel slightly more energetic. Getting out the door is a little easier. You run again. Fitness improves incrementally. Running starts feeling less hard. You run a bit further. Energy baseline rises. You start looking forward to it.
This is a reinforcing loop operating over weeks. Each run makes the next run slightly more likely. The barrier to entry drops. The reward gets bigger. The habit compounds. This is exactly why starting is so much harder than continuing — you haven't yet built the loop. The first few runs feel purely costly, no feedback benefit yet, because the loop takes time to engage.
The inverse loop is equally real. Miss a week. Fitness dips a little. Getting out the door is harder. Miss another week. Guilt accumulates (which, counterintuitively, makes exercise less likely, not more). Fitness drops further. The thought of starting again feels even bigger. The loop runs backward just as efficiently as it runs forward.
In any reinforcing loop, the easiest place to intervene is at the start — when the loop is weakest and most dependent on external force. For exercise, that means lowering the entry cost (shoes by the door, shorter distance, same time every day) until the loop is strong enough to be self-sustaining. Same principle applies to any habit you're trying to build or break.
What These Five Have in Common
Every loop on this list has the same structure: an output that loops back and changes the conditions that produced it. The mechanism differs — algorithms, prices, ice, avoidance, endorphins — but the underlying dynamics are identical.
This is what systems thinking teaches you: behavior emerges from structure. The algorithm isn't trying to make you addicted. The housing market isn't conspiring to exclude buyers. Your brain isn't broken. These are just feedback structures doing what feedback structures do — amplifying or stabilizing, depending on the loop type.
Once you can identify the loop, you can find the leverage point. Change the structure, change the behavior. That's more useful than blaming the actors.
Get Early Access to More Scenarios
We're building structured scenario progressions, guided curriculum tracks, and collaborative scenarios. Join the waitlist and be first to know.
Want more examples? The bullwhip effect shows feedback loops destroying supply chains, and our systems thinking intro explains the framework from scratch. Or join the early access list for structured scenario progressions that build systems intuition step by step.