6 360 feedback mistakes managers make (and how to avoid them)

A manager runs their first 360, gets the results back, reads "communication could be improved" and "strong leader, keep it up," stares at the screen for a minute, and quietly concludes the whole exercise was a waste of time. They're half right. The 360 didn't fail because 360 feedback doesn't work. It failed because it walked into one or more of the same six traps almost every first 360 falls into — each of which turns a potentially useful round into expensive noise.

The good news is that all six are avoidable, and none of the fixes require a consultant or a platform. They require knowing what to ask, who to ask, and what to do with the answers. Here's the list, in roughly the order they do damage.

Mistake 1: Asking about traits instead of behaviors

The single most common failure, and the one that poisons everything downstream. The questions ask whether someone "is a good communicator," "is a strong leader," "is approachable." These are trait questions, and they produce trait answers — vague, unfalsifiable, useless. Nobody can tell you how to act on "you could be more approachable."

A behavioral question asks about something observable: "When you brought a half-formed idea to this person, how did they respond?" or "In the last month, did this person interrupt you in a meeting?" Behavioral questions force the respondent to recall a specific moment, and specific moments contain the information that traits flatten into mush. The entire value of a 360 lives or dies here — we've put together a full set of 360 questions that actually surface useful answers precisely because most question banks fail this first test.

The fix: rewrite every trait question as "when did this person last…" or "describe a time when…". If a question can't be tied to an observable behavior, cut it.

Mistake 2: Reading the averages instead of the gaps

A 360 produces numbers, and the instinct is to average them — "my team rates me 3.8 on delegation" — and feel vaguely fine. Averaging is where the signal goes to die. The point of a 360 was never the mean; it's the spread, and especially the gap between how you see yourself and how others see you.

If you rated your own delegation a 4.5 and your team averages 2.8, the interesting thing isn't the 2.8 — it's the 1.7-point gap, because that gap is a blind spot you didn't know you had. Equally, when half your team rates you high and half rates you low on the same behavior, the average (a meaningless "medium") hides the real story: you're treating two groups differently. Zenger Folkman, a firm that has built its practice on 360 data, has long argued that the highest-value information in a 360 is in the outliers and the self-other gap — exactly the parts averaging erases.

The fix: never look at a 360 score without your own answer to the same question beside it, and never collapse a split distribution into a mean.

Mistake 3: Running it on others before running it on yourself

Most managers treat a 360 as something they administer to other people. The most valuable 360 you'll ever run is the one pointed at yourself — and running it on the team first, before you've ever sat on the receiving end, is both a missed opportunity and a credibility problem. You're asking your team to be vulnerable to a process you've never tested on yourself.

More practically, without a self-assessment baseline you have nothing to compare the incoming feedback against. The same peer rating means something completely different depending on whether it confirms or contradicts your own view — and you can't know which without having written your view down first. This is the foundational argument of the whole self-360 approach: self-assessment has to come before peer feedback, or the peer feedback has nowhere to land.

The fix: answer the questions about yourself first, in writing, before a single peer response comes in. The gap between the two is the entire point.

Mistake 4: Asking everyone the same questions regardless of what they can see

A 360 collects views from different vantage points — your boss, your peers, your reports — and that's its strength. But if you ask all of them the identical question set, you waste the difference. Your direct reports can speak to how you delegate and give feedback; they usually can't speak to how you handle your own manager or peers in a leadership meeting. Your peers see your collaboration and reliability; they don't see how you run a 1:1. Ask a respondent about something they've never observed and you get a polite guess, which is worse than no answer because it looks like data.

Consider a concrete case: you want to know whether you're a clear communicator. Your reports can tell you whether your written direction leaves them sure what to do next. Your peers can tell you whether you over-talk in cross-team meetings. Your boss can tell you whether your updates are pitched at the right altitude. That's three genuinely different questions hiding inside one word — and asking all three groups the single generic "is this person a clear communicator?" throws away the resolution that made the 360 worth running in the first place.

The fix: calibrate the questions to what each relationship can actually observe. Fewer, well-aimed questions per respondent beat a long generic battery everyone answers half-blind.

Mistake 5: Promising anonymity you can't actually deliver

Anonymity is what makes honest 360 feedback possible — but it only works if it's real. The classic mistake is a small team where "anonymous" feedback is trivially de-anonymizable: when you have three peers and one writes about "the time we disagreed about the roadmap," everyone knows exactly who that was. Worse is bolting a promise of anonymity onto a process that obviously can't keep it; the moment a respondent suspects you'll know who said what, they revert to the safe non-answer, and you're back to "keep up the good work."

Anonymity also has a failure mode in the other direction: used as a shield for cheap shots no one will own. The structure has to make honesty safe without making cruelty consequence-free — which mostly comes down to asking behavioral questions (hard to weaponize) rather than inviting open-ended character verdicts.

The fix: be honest about what anonymity your setup can really provide, aggregate where you can, and lean on behavioral questions that are safe to answer truthfully and hard to abuse. When the team is genuinely too small to anonymize, say so plainly and ask for signed feedback instead — an honest named answer beats a fake-anonymous one nobody trusts.

Mistake 6: Collecting the feedback and doing nothing visible with it

The most expensive mistake comes at the end. You run the round, read the results, have a private moment of reflection — and your team never sees anything change. You've now taught everyone who participated that giving you honest feedback costs them effort and risk and produces nothing. The next round, they won't bother, and you'll have killed the channel you spent real social capital opening.

A 360 isn't finished when you read it. It's finished when the people who gave you feedback can see that it landed — when you name one or two things you heard and what you're going to do about them. You don't have to act on everything, and you shouldn't. But the loop has to close visibly, or you've spent the round buying future silence.

The fix: after every round, tell the people who participated what you heard and what one change you're making. Then, later, show them it happened.

The thread connecting all six

Read the six together and a single principle runs through them: a 360 is only as good as the gap it reveals and the action it produces. Trait questions hide the gap; averaging erases it; skipping your self-baseline means there's no gap to measure; mismatched questions and broken anonymity corrupt the data the gap is built from; and doing nothing visible means even a perfect gap changes nothing. Get those right and a 360 stops being an annual box-tick and becomes the most direct view you'll ever get of the difference between how you think you lead and how you actually do.

That's the round Mirorly is built to run. You answer a set of behavioral questions about yourself first, send the same calibrated questions to the people around you, and read your self-view side-by-side with theirs — gaps surfaced, not averaged away. The peer feedback for a small team template is built for exactly the small-team reality where the six mistakes bite hardest: behavioral questions, real aggregation, your baseline first, and results structured to show you the gap instead of burying it in a mean.

The one-line summary

Most 360s fail the same six ways — trait questions, averaged signal, no self-baseline, mismatched questions, hollow anonymity, and no visible follow-through — and fixing them turns the round from a box-tick into the clearest view you'll get of the gap between your self-image and reality.