Business simulation games for training

Every business simulation makes the same promise: your people run a company, a factory or a crisis for a few compressed hours and find out what their choices do. It is a good promise. What varies is whether you bought a model that genuinely responds to decisions, or a slideshow that branches twice and calls itself a simulation.

We design and build learning games for clients, which means we have taken a lot of simulations apart. Here is what the format is good at, the kinds you will be offered, the failure modes that never appear in a demo, and a paper simulation you can run this week without buying anything.

What makes it a simulation rather than a game with a story on top

The difference is the model. In a real simulation the facilitator does not decide what happens next; a set of rules does, and they react to what every team just did. That is why the same simulation run twice produces two genuinely different sessions. Three things get sold under one word:

• Branching scenario. Pre-written outcomes attached to a menu of choices. Cheap, repeatable, good for procedure and compliance. It cannot surprise the person who wrote it.
• Simulation. A model with variables that interact. Decisions feed back into the state of the world, so consequences are calculated rather than scripted. Good for anything where second-order effects are the lesson.
• Role-play. Humans as the model. Infinitely flexible, wholly dependent on the skill of the people playing the other side, and impossible to run identically twice.

None is superior. They fail differently, and only one can teach you something the designer did not already know.

The five kinds you will actually be offered

Work out which of these is in front of you before anyone shows you the interface.

The last column predicts disappointment best. Every simulation has a dominant strategy and every group finds it faster than the designer expects; the question is whether finding it teaches the thing you paid for.

What the model rewards is what people learn

A simulation is an argument about how your business works, written in arithmetic. If the model pays for market share, teams will buy market share. If it never charges for staff turnover, nobody will protect their people. Participants read the incentives within two rounds, which is both the strength of the format and its central risk.

A simulation teaches whatever its model rewards. Get the model wrong and you have run a very engaging course on how to succeed in a world that does not exist.

Four failure modes turn up in almost every run. All are manageable if you name them in advance:

1. Optimising the sim, not the business. By round three someone has spotted the lever the model overpays for and the team stops reasoning about customers. Surfaced in the debrief, this is the most useful thing that happened: people gamed a measurement system, exactly as they do at work.
2. Spreadsheet capture. The most numerate, most confident person takes the controls and the other four become an audience. If your objective was team dynamics rather than finance, you have bought an expensive solo game. Make a different person justify each round out loud.
3. The black box. If a team cannot see why their number moved, they invent a reason and believe it. Superstition, not learning. Ask a supplier to explain the core of the model in two sentences; if they cannot, the debrief has nothing to stand on.
4. The compressed horizon. Four quarters in three hours makes anything with a slow payoff, such as training or maintenance, look irrational inside the game. Say so before the debrief ends, or you have quietly taught short-termism and scored people for it.

When a business simulation makes training worse

The section that decides whether the money is well spent. Five situations where a simulation is the wrong instrument:

1. The objective is knowledge. If people need to know a policy or a product spec, write a good document. A simulation costs many times more and teaches it less reliably.
2. Nobody will run the debrief. A simulation generates data and feelings. Without someone to turn those into conclusions, you have run an expensive board game and people will describe it that way afterwards.
3. The model contradicts your actual economics. Off-the-shelf sims are built to be generic. If yours rewards volume when your business lives on retention, the transfer is negative. Check it against the three drivers your business genuinely turns on.
4. Seniority is mixed and scoring is public. Put a director and two graduates on one team with a live leaderboard and the graduates stop offering ideas by round two. Split by level, or drop the public board.
5. It is secretly an assessment. If anyone is watching the results with talent decisions in mind, say so beforehand. Undisclosed assessment destroys psychological safety for every programme you run afterwards, not just this one.

Something you can run on Monday: the four-quarter paper simulation

You do not need a licence to find out whether this format suits your group. Four to six teams of three to five, ninety minutes, one sheet per team and a flipchart. Each team runs a competing business in the same market.

1. The rules, on one page. Each quarter a team submits three numbers: a price per unit between 40 and 100, capacity bought this quarter at 30 per unit, and one investment of 500 in quality, marketing or holding the cash.
2. How the market resolves. Total demand is 300 units a quarter. Allocate it cheapest price first, capped by each team's capacity, until it runs out. Revenue is units sold times price. Costs are capacity times 30, paid whether or not the units sell, plus the investment. Quality adds 40 units of demand to that team next quarter; marketing moves them one place up the allocation order for one quarter.
3. The shock. After quarter three's decisions are submitted and before they are resolved, announce that demand has fallen to 180 units. No warning. Teams that bought capacity for a boom now pay for all of it.
4. The board. Publish a leaderboard after every quarter showing that quarter's profit. It goes on the wall where everyone can see it.
5. The scoring rule. Stated once in the briefing, in one line: the game is won on total profit across all four quarters. Nobody usually notices the mismatch with the quarterly board until you show them at the end.
6. Run it. Ten minutes to brief, ten per quarter to decide, three to resolve and post the board, thirty to debrief. The debrief is longer than any round for a reason.

The debrief question: which number did you actually manage, the one that decided the game or the one on the wall, and which of those is on the wall at work? Almost every group manages the visible measure. That is not a trick: it is how measurement runs organisations, and it lands harder when a group did it to themselves in the last hour.

Two more if you have time: who stopped speaking, and in which quarter? What did the shock reveal about assumptions nobody had written down?

How to interrogate a simulation before you buy one

• Explain the model in two sentences. If the answer is proprietary, the debrief will be a guessing game.
• How many decisions does a team make per round, and who makes them? Fewer, weightier decisions beat a form with thirty fields every time.
• What happens to a team that does nothing? If passivity scores respectably, the model is not tracking anything real.
• What does it produce for the facilitator? Decisions, timings, who changed their mind and when. A final score is a result, not evidence.
• Can it be run twice with the same group? The second run, with the model now understood, is often where the actual learning is.
• Where does the percentage come from? You will be shown an improvement figure. Ask for the study. Corporate learning is thick with numbers that trace back to nothing, and a supplier who admits the figure is marketing is one worth talking to.

The best-known example is still the beer distribution game, developed at MIT's Sloan School in the 1960s out of Jay Forrester's work on system dynamics. Four players, a supply chain, no communication except orders. It reliably produces wild swings in stock and a great deal of blame, all of it caused by the structure rather than anybody's incompetence. It costs nothing but paper.

Where the learning actually lands

The simulation is the experience. It becomes learning in the conversation afterwards, which is the part cut first when the schedule slips. Budget a third of the time for it and protect it. How to debrief an activity has the questions in the order that works, and Kolb's cycle explains why reflection has to come before anyone generalises.

Evaluate it honestly too. Simulations score beautifully on reaction because they are enjoyable, and that is the level everyone reports. Decide in advance which behaviour should look different, and check at ninety days. The Kirkpatrick model gives you four levels; the discipline is refusing to report level one as though it were level three.

When a stock simulation is the wrong shape

Sometimes the model you need does not exist, because the thing you are training is specific to you: your escalation rules, your regulatory pressure, the merger you are three months into. That is when building beats buying, and it is what our gamified learning apps are: one client, one objective, played on participants' own phones with a facilitator driving the pace live.

HEARTWARE is one of ours: five rounds over ninety minutes on judgment in the AI era, where teams decide what to trust, what to verify and what to own. Operation Keystone is a spy-thriller induction for two hundred new joiners in sixteen teams, deliberately hybrid: ciphers inside the app, physical challenges run by facilitators off-screen, each solved puzzle yielding the passcode for the next level. In both, reveals land on every phone at once so no team gets ahead of the conversation. Neither would have worked as a generic simulation, and neither would work without the debrief.

Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play Gamified learning appsThe Weekly Challenge to see the method in ten minutes.