There's a gap between knowing what systems thinking is and thinking systems. Reading about reinforcing feedback loops or watching a supply chain simulation is not the same as feeling a system respond to your decisions. The exercises on this page are designed to close that gap — some are played in groups, some are drawn by hand, one runs in a browser. All of them produce the same result: participants start seeing structure instead of events.
No software required for the first six. The seventh is the exception — and it's worth it.
Exercises 1-3 are best for groups of 5 or more. Exercises 4-6 work with 2-3 people or solo. Exercise 7 is a digital companion for after a workshop. Pick based on your setting, not the list order.
The Beer Distribution Game is the most studied system dynamics exercise in existence. It was invented at MIT in the 1950s and has been run with thousands of groups — executives, students, military officers. The results are always the same: the system collapses, and nobody can explain why.
Here's how it works. You have a four-layer supply chain: Factory → Distributor → Wholesaler → Retailer. Each layer can only order from the layer below, and orders take two weeks to arrive. Customer demand at the retailer is constant — 4 cases per week, every week. The goal: keep inventory low while filling every order.
Why it's powerful: Within 20 minutes, inventories swing wildly. The factory starts building mountains of stock while the retailer runs empty. Everyone blames the layer above them. Then you reveal the data — and show that the upstream amplification of a perfectly stable demand signal looks exactly like the bullwhip effect. The system did it. Nobody did it.
Setup: Each participant gets a single sheet of paper or a simple order form. No computers needed. You can run it with paper cards, or use the free Beer Game web version if you have screens available.
Debrief questions that land:
- At what point did you start ordering more than you needed? What were you reacting to?
- Was there a moment you felt the situation was out of control?
- Who was "responsible" for the problem? (Usually: everyone, which means the system is responsible.)
The Iceberg Model is a Donella Meadows visualization that shows the relationship between events, patterns, systemic structure, and mental models. Most people respond only to events — they see the stock market drop and panic, or the project misses deadline and blame the team. The iceberg asks: what's underneath?
The model has four levels:
- Events: What happened. The stock crashed. The customer left.
- Patterns: What keeps happening. Recurring revenue decline. Consistent lateness.
- Systemic Structure: The underlying architecture that produces the pattern. A sales comp plan that rewards new logos over expansion. A team that can't say no to scope creep.
- Mental Models: The assumptions driving the structure. "We need to grow at all costs." "Quality is a later problem."
How to run it: Draw a large iceberg outline on a whiteboard or flip chart. Write "EVENTS" above the waterline. Put the three underwater levels below. Pick a problem the group cares about — a product failure, a customer churn event, a market shift. Ask: What is this an example of? What pattern does it fit? What structure keeps producing this? What assumptions are we making that keep that structure in place?
Why it works: Most problem-solving happens at the event level. Moving up the iceberg to structure is where leverage lives. Moving to mental models is where deep change starts. This exercise makes that progression tangible.
A Connection Circle is a deceptively simple exercise: given a set of variables, participants draw arrows showing how each variable influences the others. The variables might be: "Population," "Births," "Deaths," "Food Supply." The question is always the same: what influences what, and in what direction?
A positive influence (+): when A increases, B increases (or when A decreases, B decreases).
A negative influence (−): when A increases, B decreases.
How to run it: Write 6-10 variables on sticky notes around the outside of a large sheet of paper. Participants draw arrows between variables, marking each with + or −. After drawing, identify any closed loops — variables that eventually connect back to themselves. Label each loop R (reinforcing) or B (balancing) based on whether it has an odd or even number of negative links.
Why it's powerful: Connection Circles are often where people first encounter feedback loops by name. The act of drawing the connections makes the loop structure visible — something that reads abstractly in text becomes obvious when you trace it with a pen. Reinforcing loops amplify change. Balancing loops resist it. Both show up in the diagram.
Start with a Connection Circle before moving to causal loop diagrams. The Circle uses the same mental operation — identifying influence direction — but without the notation overhead. Once participants are comfortable with + and −, adding polarity labels to CLD arrows is a small step.
This is one of the foundational practices in system dynamics education, developed by the Waters Foundation. The instruction is simple: given a real or hypothetical situation, sketch how a variable changes over time — without numbers, without axes labels. Just the shape of the curve.
Examples to practice with:
- Your energy level over the course of a week (if you drink coffee every morning vs. not)
- The population of a fish species after it's introduced to a new lake, with and without predators
- Your skill at a new instrument over a year of practice
- The price of a hot commodity over a 10-year market cycle
How to do it: Give participants blank paper. Ask them to draw the horizontal axis (time) and vertical axis (the variable). Then sketch the shape. After individual work, compare curves — discuss why shapes differ. This is where the systems thinking begins: different shapes imply different underlying structures.
Why it's powerful: BOTGs train the habit of asking "toward what?" instead of "what happened?" Most people describe events. Systems thinkers describe trajectories. A graph of "what happened over time" forces the pattern into view before anyone can reduce it to a single cause.
Stocks and flows are the atoms of system dynamics. A stock is anything that accumulates or depletes — money in an account, water in a reservoir, a population. A flow is the rate of change — the inflow that fills it, the outflow that drains it. Stocks change only through flows. That's it. That's the whole framework.
How to run it: Pick a system the group knows well — a project backlog, a savings account, a gym membership, a team headcount. Draw a rectangle for the stock. Draw an arrow pointing in for the inflow (additions). Draw an arrow pointing out for the outflow (losses). Label each flow. Then ask: What slows down the inflow? What speeds it up? What controls the outflow?
Key insight to surface: Stocks have inertia. Flows take time to change. If you want to grow a stock (revenue, trust, capability), you need to increase the inflow — and that takes time before the stock changes. This is why short-term thinking produces such systematically poor outcomes: you're trying to move a stock by tweaking a flow, but the flow responds slowly and the stock moves even more slowly.
Analogy that usually lands: A bathtub. The stock is the water. The inflow is the faucet. The outflow is the drain. If you want more water, you turn up the faucet. The water doesn't appear instantly. If you turn off the faucet, the water keeps draining until it's empty. Systems that work this way — savings, knowledge, fitness, relationships — all have this bathtub structure.
Group Model Building (GMB) is a structured facilitation protocol developed at the University of Wisconsin. It's the most complete exercise on this list — a full workshop format that uses all the elements above to build a shared causal loop diagram of a real organizational problem.
Core steps:
- Scripting: Each participant writes down the most important thing affecting the issue at hand — anonymously. The facilitator reads them aloud without attribution. This surfaces the full range of concerns without anchoring on whoever speaks first.
- Variable elicitation: From the scripting output, the group identifies the key variables. These become nodes in the diagram.
- Connection drawing: The group draws arrows between variables, marking + or −, one at a time. Debates about direction are where the mental models surface.
- Loop identification: Find the closed loops. Label them R or B. Discuss: which loops are currently dominant? Which are being overwhelmed?
- Action testing: Pick a proposed action. Trace through the diagram: what does this change? Where does the effect go? Does it reach the intended goal, or does a balancing loop resist the change?
Why it's powerful: GMB creates shared language. After a GMB session, a team has a diagram they can refer to, argue about, and test actions against. It converts "I think we should..." into "Let's trace it through the model." That shift — from opinion to inference — is the whole point of systems thinking.
GMB rarely produces a "perfect" diagram on the first pass. That's not the point. The point is that a team that has argued about the structure of their problem understands it fundamentally better than a team that hasn't. The diagram is a side effect. The shared mental model is the output.
The first six exercises are analog. This one is digital — and worth it. The Emergent Playground drops you into a pre-built dynamic system where every decision you make propagates through a network of linked variables. You don't draw the diagram — you interact with the system and watch what happens.
Start with the ecosystem scenario. You'll manage predator and prey populations, vegetation coverage, and a feedback loop between them. Within minutes, you'll feel the system respond to your choices — you'll see a reinforcing loop accelerate, watch a balancing loop stabilize, and encounter a delay that makes you react too late. Then try the supply chain scenario to relive the Beer Game dynamics in real time.
After a workshop, assigning 20 minutes in the playground is the best bridge between understanding the concept and feeling the system. The exercises above give you the structure. The playground gives you the experience.
These seven exercises form a progression from simple and physical to complex and digital. The Beer Game and Connection Circles are entry points — anyone can participate, no background required. BOTGs and Stock & Flow mapping develop the core habits of systems thinking. Group Model Building is the deep end. The Emergent Playground closes the loop by giving participants a system to inhabit rather than just describe.
Most of these exercises can be run in a single workshop session. The investment in facilitation is low and the return in shared mental models is high. Systems thinking is a practice, not a theory. These exercises are where the practice starts.
Get Early Access to More Scenarios
We're building guided curriculum tracks, structured scenario progressions, and collaborative workshop tools. Join the waitlist and be first to know.
New to systems thinking? Our intro article explains the framework from scratch. Ready to look at real tools? See our comparison of 6 systems thinking tools or learn why supply chains break via the bullwhip effect.