The Engineer's Fallacy: Why Your Message Is Never the One That Arrives

The meeting ended the way good meetings are supposed to. Heads nodded, the action items were repeated back, everyone agreed on "the direction." Two weeks later, three people had built three different things, each convinced they were doing exactly what had been asked. Nobody had lied. Nobody had failed to listen. And the person who called the meeting said the sentence managers say more than almost any other — the sentence that is almost always false: "But I was perfectly clear."

It is worth sitting with how strange that sentence is. It treats clarity as something you can possess and then hand over intact, like a document. It assumes that if the words leaving your mouth were unambiguous, the understanding arriving in someone else's head must be unambiguous too. That assumption feels like plain common sense. It is also one of the most expensive mistakes in organizational life, and it comes from a model of communication most of us absorbed without ever noticing we had adopted it.

The model in our heads is an engineer's

In 1948, Claude Shannon gave the world a theory of communication so powerful it built the digital age — the framework later formalized with Warren Weaver and known ever since as the Shannon-Weaver Model. A sender encodes a message, pushes it through a channel, a receiver decodes it; the engineer's task is to make sure the bits that come out match the bits that went in, despite the noise. I traced its anatomy — source, transmitter, channel, noise, receiver — in an earlier article; here I want to push at the assumption it quietly smuggles into the way we manage people. It is one of the most beautiful ideas of the twentieth century, and that is precisely what makes it dangerous when borrowed.

It is also, quietly, the picture almost every manager carries in their head about talking to people. I have a thought. I encode it into words. The words travel. You decode them back into my thought. Communication succeeds when your version matches mine.

The trouble is that Shannon was modelling the fidelity of a signal, not the creation of meaning — and he said so plainly. His theory is exquisite about whether the right symbols arrive and entirely indifferent to what they mean once they do. We borrowed the part about transmission and silently assumed the meaning came along for the ride. It doesn't.

Meaning isn't sent — it's rebuilt

Here is what actually happens. The words arrive, and the listener builds a meaning out of them, using their own context, their own assumptions, their memory of the last six things you said, and the mood they were in this morning. You do not transmit understanding. You hand someone a pile of raw materials, and they construct understanding in a workshop you cannot see and do not control.

This is why "move fast on this" produces a working prototype by Friday from one person and a five-week plan from another. Both decoded the words correctly. They reconstructed different meanings, because they brought different materials to the building site. Clarity, it turns out, is not a property of your message at all. It is a property of the listener's reconstruction. You can write the cleanest sentence in the world and still be misunderstood by someone whose materials differ from yours — and you will never feel it happen, because on your side everything was perfectly clear.

The curse of the one who knows

The cruelty of this is that expertise makes it worse. The more deeply you understand something, the harder it becomes to imagine not understanding it. Psychologists call this the curse of knowledge, and a famous little experiment captures it perfectly. People asked to tap out the rhythm of a song everyone knows were confident their listeners would recognize it about half the time. The real rate was closer to one in forty. In the tapper's head the melody plays in full; the listener hears only knocks on a table.

Every expert leader is a tapper. The strategy plays in full in your mind — the context, the trade-offs, the reasons it has to be this way. What reaches the team is the tapping. And because you can hear your own melody so vividly, you systematically under-supply the very context the listener needs to rebuild it. The clearer the message feels to you, the more you have probably left out.

Why "concise" can be dangerous

This is where a hard-edged, efficiency-minded instinct quietly backfires. We are trained to compress: the one-line email, the single-slide summary, the bullet that says it all. Information theory loves compression — strip the redundancy, send only what is strictly needed. But Shannon also knew the other half of the story: a noisy channel requires redundancy to survive. Error correction is repetition, restatement, the same idea approached from a second angle.

Human channels are extraordinarily noisy. Stripping a message to its leanest form does not make it clearer; it makes it more fragile, removing exactly the redundancy that would let a listener catch and correct a misbuild. The terse, elegant directive feels like superb communication and is often the most efficiently decoded into the wrong thing. Sometimes the leanest message is the one most reliably misunderstood.

The amplifier nobody warns you about

Add authority, and the physics shift again. A leader cannot not communicate — a raised eyebrow in a review, a question skipped, an email left unanswered are all reconstructed into meaning whether you intended it or not. And the higher the rank, the larger the structure the organization builds out of the smaller signal. An offhand "I'm not sure about this market," muttered in a corridor, becomes a strategy by Monday. You are not speaking into a neutral channel. You are speaking into an amplifier that takes your faintest signal and hands the whole organization a fully built interpretation of it.

From broadcasting to error-correction

None of this is a counsel of despair, and it is not a plea to talk more. It is an argument for a different job entirely: stop optimizing the message and start managing the reconstruction.

In practice, that means a few unglamorous habits. Build for the listener's materials, not your own — supply the thing they are missing, not the thing that already feels complete to you. Add redundancy on purpose: an important idea earns a second framing, an example, a why. And above all, close the loop. The only way to learn what someone has rebuilt is to make them show you — not "does that make sense?", which only ever earns a polite yes, but "walk me through how you'd approach it," which reveals the reconstruction while there is still time to repair it.

The best communicators I have worked with are not the most eloquent. They are the ones who treat every important message as a protocol with acknowledgment rather than a broadcast — who assume, by default, that what they sent is not what arrived, and who stay genuinely curious about the difference. They have traded the comfort of "I was clear" for something far more useful: knowing what was actually built.

The next time you are certain you were understood, ask the quieter question — not whether your message was clear, but what the person in front of you actually built out of it, and whether you have ever once bothered to look.