THE IDEA
The useful lie
Every model, framework, diagram, plan, and description is a simplification. It leaves things out. It has to - that’s what makes it useful. A map that contained every detail of the territory would be the same size as the territory and just as difficult to navigate. The value of a map is precisely that it’s simpler than reality.
The danger starts when we forget the simplification. When we start treating the map as if it were the territory. When the financial model becomes “what will happen” rather than “one possible scenario.” When the org chart becomes the organisation rather than a crude drawing of reporting lines. When the strategy document becomes the strategy rather than a description of intentions that will meet reality and change.
Alfred Korzybski coined the phrase in 1931, and it remains one of the most useful correctives in all of systems thinking. Every tool in this field - feedback loops, causal loop diagrams, stock and flow models, archetypes - is a map. They’re good maps. They reveal things that are otherwise invisible. But they’re still maps, and the territory will always contain surprises that the map didn’t predict, because the map was designed to simplify and the territory wasn’t.
IN PRACTICE
When the model meets the mess
A financial model projects steady growth for the next five years. The inputs are reasonable. The logic is sound. The spreadsheet is beautiful. The board approves the strategy based on the model. Two years later, a market shift that the model had no variable for changes everything. The model wasn’t wrong in its own terms. It was a map that didn’t include the mountain range the company was about to walk into. Nobody checked because the map was so convincing.
A personality test sorts people into types. The types are useful - they give a shared language for differences in how people communicate and work. But when a manager starts saying “she’s an introvert, so she won’t want to lead the presentation” or “he’s a thinker, so he won’t be good with clients,” the map has replaced the person. The type is a simplification. The person is the territory - full of contradictions, context-dependent behaviour, and capabilities that no four-letter code can capture.
A city planner uses a traffic simulation model to predict the impact of a new road. The model says congestion will decrease by 20%. The road is built. Congestion decreases by 10% for six months, then returns to previous levels as the new capacity induces new demand. The model was correct about the mechanical effect of the road. It was wrong about the behavioural response, because it modelled traffic as physics rather than as people making adaptive decisions. The map was precise. The territory was adaptive.
WORKING WITH THIS
Holding maps lightly
The goal isn’t to stop using maps. Maps are indispensable. The goal is to hold them with the right grip - tight enough to be useful, loose enough to update when reality disagrees.
Before trusting any model, ask three questions. What does this map leave out? Every simplification has an edge where it stops being accurate. Knowing where that edge is tells you where the map is trustworthy and where you’re flying blind. What would this map not predict? Think about the kind of surprise that would fall outside the model’s scope. If you can imagine a plausible scenario the map can’t handle, that’s information about its limits. When was this map made? Reality changes. Maps don’t update themselves. An accurate map from five years ago may be dangerously outdated today.
The most mature relationship with maps is to use several at once. Each one simplifies differently. Each one reveals things the others hide. A financial model, a stakeholder map, a customer journey, and a conversation with frontline staff are all maps of the same territory. None is complete. Together, they’re better than any one alone.
THE INSIGHT
The line to remember
The map is useful because it’s simpler than reality. It’s dangerous for exactly the same reason.
RECOGNITION
When this is in play
You’re confusing the map for the territory when a model’s prediction is treated as a certainty. When a framework’s categories become more real than the things they describe. When someone says “the data says” about a situation where the data is a small, selective slice of what’s happening. When a plan is followed precisely in a situation the plan didn’t anticipate. When the response to a surprise is “that shouldn’t have happened” rather than “our model didn’t include that” - the territory doesn’t care what the map says.