It doesn't check your title, your tenure, or how many people report to you. It shows a color, and the color is the same for everyone who looks at it. That is the entire mechanism. Not enforcement. Not trust. A shared signal that nobody has to interpret twice.
Most organizations do not run on this mechanism. They run on the other one — the one where alignment means trusting the person, not reading the signal. If a senior person says green, the room hears green, even when three people privately clocked amber. Nobody built a light. They just agreed to watch a face.
This works, for a while, at a small enough size. You can hold ten people's judgment in your head. You know who tends to round up, who catches real risk, who goes quiet when they disagree. Trust-in-person alignment is just pattern recognition running on a small enough sample that it hasn't broken yet.
It does not survive scale. Not because people get worse. Because the sample gets too big to hold. At two hundred people, you cannot personally calibrate your trust in each one. So either you build something that doesn't require calibration — a real signal, legible the same way to everyone who looks at it — or you get theater. Nodding at whoever is most senior in the room. Calling it alignment because no one contradicted the color that was never actually named.
Here is the part that makes the structural version unpopular: it costs something the personal version doesn't ask for.
Building a real signal means writing down what red actually means before you need it. It means someone junior can point at the criteria and say this is red, and be correct, even when the most senior person in the room wants it to be green. That is the whole value of the mechanism — and it is also exactly what makes people avoid building it. Personal-trust alignment is comfortable because it is unfalsifiable. Nobody can be wrong about a feeling in the room. Everybody can be wrong about a stated color.
So teams keep the informal version on purpose, without ever framing it as a choice. They call it "reading the room." They call it "trusting the team." What they are actually doing is keeping the signal personal, because a personal signal never has to be defended, and a structural one does.
You can find the fracture point in almost every account of an alignment failure, if you listen for it. We thought we were aligned. Reread that sentence. It's not describing a false signal. It's describing the absence of one. What existed instead was a room full of people reading a tone, a face, a pause before someone spoke — and calling their aggregate reading of that tone "alignment," because there was no color to check it against.
A traffic light removes that ambiguity by design, not by trust. Cars don't slow down because they believe in the driver coming the other way. They slow down because the color holds regardless of who's behind the wheel.
That's not a limitation of the mechanism. That's the whole reason it works.