The feedback about you is filtered long before it reaches you

The most useful information is the least likely to be volunteered

Suppose something about how you work is causing a problem for the people around you. For that to reach you, somebody has to decide that telling you is worth the awkwardness, the risk of a defensive response, and the possibility that they have misread the situation. Each of those weighs against speaking, and none of them weighs against staying quiet.

The result is not that you receive no feedback. It is that you receive a filtered sample, and the filtering is not random. Favourable observations pass easily because they cost nothing to deliver, while unfavourable ones pass only when somebody has a strong enough reason to bear the cost. Whatever arrives has already been through that gate.

This is a separate problem from bad feedback loops in general

Elsewhere on this site the argument is that experience only teaches when the world answers back quickly, clearly and honestly. That concerns the structure of outcomes — delay, noise and ambiguity in what actually happens.

What is described here is different. The outcome may be perfectly clear to everyone else, and the failure is in transmission. A person can be the only one in a group who does not know something about their own performance, and nothing in the environment is ambiguous at all. That distinction matters because the remedies are unrelated: one requires better measurement, the other requires making it cheaper for somebody to speak.

The filter tightens with standing

Two things reliably thicken it. The first is any form of power, including informal influence: the more a person can affect somebody’s situation, the more that somebody weighs the risk of speaking. The second is your own history of response. A single defensive reaction is remembered, discussed and generalised, and it can end the flow of information from an entire group without any of them deciding to withhold anything permanently.

Neither requires the person to be intimidating. Ordinary seniority and one bad afternoon are sufficient, and the effect is invisible from the inside precisely because its symptom is the absence of something.

What gets through is distorted in a predictable direction

The feedback that does arrive tends to be positive, general, delayed and softened, and it often comes from whoever has least at stake. Specific criticism is the hardest category to deliver, so it converts into vagueness — a remark about the general approach rather than about the paragraph, the meeting or the decision.

That vagueness is what makes such feedback useless rather than merely scarce. A general impression cannot be acted on, and the natural response to it is to ask for examples, which puts the cost back on the person who was already reluctant. This is why so many attempts to gather honest opinions produce nothing usable.

What actually surfaces it

Narrowing the question does most of the work. Asking what somebody thinks of your work invites a verdict, which is socially expensive. Asking what the weakest section of a specific document is, or what they would have done differently in one particular meeting, requests an observation about an artefact rather than a judgement about a person.

Asking for one thing rather than everything also helps, because a single named improvement is not an indictment. And treating whatever arrives without argument is the whole of the technique for keeping the channel open: the response to a criticism determines the price of the next one, and the price is set by the recipient rather than by the speaker.

Reading the silence

The hardest part is interpreting absence, and the honest position is that silence is weak evidence in both directions. It is compatible with nothing being wrong and equally compatible with the filter working normally, and there is no way to distinguish them from where you stand.

What can be checked is whether anybody has told you anything uncomfortable recently. A long stretch of exclusively favourable input is not usually evidence of a long stretch of faultless work. It is more often evidence about the channel, and the channel is the thing you can actually do something about.

One further check is available and it is worth taking occasionally. Ask somebody who has already told you something unwelcome, since they have demonstrated that the cost is one they will bear, and their answer is less filtered than anybody else’s by construction. The people who are easiest to ask are, for the same reason, the least informative.