How to get 360 feedback, with or without the official process

A 360 is just feedback from every side of your work: your manager, your peers, the people who report to you, and the partners on other teams who see a version of you that none of the others do. Companies run formal versions once or twice a year. You don't need one to get the benefits, and if you wait for one, most of what you learn will arrive months after you could have used it.

This playbook covers how to run the informal version yourself: choosing the reviewers, asking questions people can actually answer, receiving the answers so a second round happens, and turning one round into a rhythm. It also covers how to work the official process when your company has one, and where 360s go wrong. If the problem is getting even one person to tell you the truth, start with the honest feedback playbook; this one is about assembling the full picture from many people at once.

The advice comes from people who have run this at every altitude, from Kim Scott, who wrote Radical Candor, to executives at Amazon, Netflix, Zapier, and Bridgewater. Every quote plays at the moment it was said.

Why you can't wait for the official cycle

The annual review is the worst place to hear about a problem for the first time. It's attached to pay and promotion, so everyone involved is careful, and it batches a year of observations into one document. Ethan Evans's blunt version: if the first time you hear about a weakness is in the review that sets your raise, you've already paid for it.

Even the best-run cycles deliver their news late. Elizabeth Stone gets about 300 pieces of feedback in Netflix's annual 360, and some of them describe things that happened six months earlier. "I think, oh, I wish you had told me this at the time."

The weaker versions have a different problem: the tool teaches people to soften. Liz Christo, a general partner at Stage 2 Capital, watched formal 360s come back "kind of nice" while the real conversation happened somewhere else, and the anonymity rarely survived contact: "we all know who it's from" (34:28 in her Delivering Value episode).

So the version you run yourself is not a substitute for the official one. It's the thing the official one is a periodic approximation of.

Pick a challenge network, not a fan club

The instinct when choosing who to ask is to pick people who like you. Adam Grant's research points the other way: left alone, we quietly drop our critics and surround ourselves with cheerleaders, which is exactly how blind spots survive. His fix is what he calls a challenge network: "the group of people that you trust to push you to get better. They tell you the stuff you don't want to hear, but need to hear" (12:11 in the WorkLife episode).

For a working 360, seven to ten people covers every side without becoming a survey program:

Give people permission before you ask

Nobody owes you candor, and most people have learned that giving it is risky. The person who tells you the truth pays a social cost if you flinch, so their rational default is "everything's fine." That means the first move isn't a question, it's permission.

Permission sounds like naming the risk out loud: "I'm working on how I run meetings, and I'd genuinely rather hear it from you than keep guessing. Trust me, I'm open to it." Say why you're asking, say what you'll do with it, and mean the last part, because the second round depends entirely on what you did with the first. Our honest feedback playbook covers why the people around you softened in the first place; this is the antidote applied deliberately.

Ask a question someone can actually answer

The question most people ask is the one guaranteed to fail. Kim Scott has watched it fail for years:

Her replacement is a go-to question you write once and keep: hers is "What could I do or stop doing that would make it easier to work with me?" The wording is doing real work there. It assumes something exists, it's scoped to working with you rather than judging you, and it can't be answered with "fine." But she's emphatic that you shouldn't borrow her sentence, because a script in someone else's voice reads as a script: "there's no such thing as emotional novocaine." Write one that sounds like you.

Wade Foster runs Zapier's formal 360s twice a year, and his personal technique for getting useful answers is to stop asking about himself entirely:

His version anchors to a shared artifact instead: "How could we have done twice as well on this project?" The answers are about the work, but you own the work, so the feedback is about you without anyone having to say so. It's the easiest question on this page for a nervous colleague to answer honestly.

Jules Walter, a product leader at YouTube, adds the version for when the power gap runs the other way, when people see giving you feedback as a risk they didn't sign up for:

Two of his techniques travel well. Ask about one named dimension you're already working on, "to what extent did I show executive presence in that meeting?", so the other person is confirming or adjusting rather than initiating. Or criticize yourself first and let them calibrate: "I thought I lost the room in the second half, does that match what you saw?" Agreeing with you costs them almost nothing, and disagreeing with you is, technically, a compliment.

Take it so the second round happens

Everything before this point buys you one honest answer. Whether you ever get a second one depends on the thirty seconds after you hear the first.

Kim Scott's mechanical trick for those thirty seconds: "the simplest way to embrace the discomfort is to close your mouth and count to six." Six seconds of silence feels endless, and it's usually where the real sentence arrives, because the speaker fills it with the thing they softened the first time. Ethan Evans supplies the only acceptable opening reply: "The correct answer to all feedback is thank you. Now, you can say more than thank you, but you don't argue." Then ask for a concrete example, and if they can't produce one, make the ask that turns one conversation into a standing arrangement: "the next time you see it, could you point it out in the moment?"

Ray Dalio got a written D-minus from an employee for a rambling meeting, and his response, asking every attendee to grade him and sharing the results with the company, is extreme. The idea underneath it isn't:

You can't control the grade. You fully control how you take it, and the people watching you take it are deciding whether to ever grade you honestly again. Mellody Hobson, co-CEO of Ariel Investments, held back tears through brutal feedback from Bill Bradley for exactly this reason: "if I cry, he won't ever give me feedback again" (7:56 in her WorkLife episode). Her coach's repetition rule is also worth carrying into any 360 read-out: one person's comment is a data point, but "the third time they call you a horse's ass, buy a saddle."

One logistical rule from Matt Mochary's coaching curriculum: never take serious feedback over Slack or email, and budget twenty to thirty minutes when you receive it live. Hard words in text hit the threat response with no tone and no repair loop, and his observation is that within a day or two, confirmation bias starts doing permanent damage that a live conversation would have resolved in minutes (Mochary Method Class 2).

Close the loop, then make it a rhythm

Feedback you visibly acted on is the only advertisement that gets you more. Report back to each person: what you heard, what you changed, and what you'd like them to watch for next. Then ask the follow-up that proves you meant it, Scott's favorite: "Did I overcorrect?"

For cadence, Scott's floor for managers is startlingly high: solicit feedback "every week, from each of your direct reports," which in practice means budgeting the last five minutes of every 1:1 and asking your go-to question until answering it stops being scary. For everyone else, the same question aimed at a peer or your boss once every few weeks keeps the channel open at nearly zero cost.

And when something real needs to change, make the loop measurable. After Samantha Leal, a growth leader, learned secondhand that her team was afraid of her, she convened the whole team, apologized, asked for specifics, wrote every item on the board, and committed to a follow-up meeting a month out. Then she added the step that turned an apology into a system:

The scores took three or four months to move. That's the honest timescale for a reputation changing, and the monthly number is what kept both sides from giving up in month two. Team members thanked her privately; she calls the whole episode career-defining.

Pick the channel: conversations, a form, or a tool

Two decisions here: which channel carries which conversations, and who stays anonymous. The everyday conversations are named by nature; anonymity is a property of the deliberate written round, and there it follows a working convention. Feedback from people above you, your manager and your skip-level, is unmasked, because they're paid to tell you directly and there's no way to anonymize a bucket of one. Feedback from your level and below, peers, reports, and the people you influence without authority, goes anonymous, because that's the direction where candor carries real risk. The mechanics follow from the convention.

Conversations carry the unmasked tier and the continuous rhythm. Your go-to question in 1:1s, the skip-level ask, the project debrief with a peer who'll say it to your face: named, live, and the only channel where the follow-up question works. Everything earlier on this page is this channel.

A quick anonymous form carries the lateral-and-below tier for the deliberate once-or-twice-a-year round. A Google Form set to collect no email addresses is genuinely enough: your go-to question as an open text box, a "what should I keep doing?" box, and one 0-to-10 score you can track between rounds, the way Samantha Leal did. Keep it under five questions so people finish it. The three-person rule is what keeps the promise honest: if a bucket has fewer than three respondents, merge it into the next one or drop the segmentation, because "anonymous" feedback from a bucket of two is a guessing game you've invited people into.

Tools take over when you want the round to run without you administering it. If your company already pays for an engagement platform, it can usually run a personal 360; Zapier runs its formal cycle on Culture Amp twice a year. And Work Coach's 360 feedback covers the layer none of these channels can: it observes your actual meetings and surfaces what a candid colleague would point out, continuously. It isn't your colleagues' own words, but it's the evidence to read them against, it never goes stale, and it needs no anonymity plan.

If your company runs a formal 360

Work the official process instead of just submitting to it. Evans's advice from the inside: if you've changed teams or managers recently, your record is thin, so prompt your current and former managers and the peers who actually saw your work to submit input. A 360 populated by the right eight people is a different document from one populated by whoever the tool guessed.

At the receiving end, compress before you react. Elizabeth Stone's method for her 300 pieces:

She distills the themes and proactively tees them up with her manager, which turns the pile into two or three workstreams and signals exactly the second score Dalio was talking about. Netflix deliberately decouples the cycle from compensation so people write the truth; if yours is attached to comp, read it knowing everyone was careful.

One caution on the themes themselves: interrogate them before adopting them. The executive coach Donna Lichaw tells the story of a client whose 360s kept praising her "attention to detail," a label she hated and that kept earning her more detail work (46:38 on Lenny's Podcast). A repeated theme is real signal, but it describes what people have seen you do, not what you should do more of. Sometimes the right response to a theme is to change the work, not to lean in.

The asks on one page

Frequently asked questions

How many people should I ask?

Seven to ten: your manager and your skip-level, three or more peers including one critic, your reports if you have them, and one or two cross-functional partners. The floor is set by anonymity, at least three people in any bucket that answers anonymously, and the ceiling by follow-through: more reviewers than you can close the loop with is a survey, not a 360.

Should the feedback be anonymous?

Only in the written round, and only below the line. Your everyday conversations stay named, and feedback from your manager and skip-level comes to your face in any channel, since a bucket of one can't be anonymized anyway. In the written round, peers and reports answer anonymously, with at least three people per bucket so the promise is real. Just don't let anonymity become the whole system: the anonymous form finds the theme, and the named conversation is where the follow-up question turns the theme into something you can act on.

How often should I run this?

Two rhythms. Continuous and small: the go-to question in 1:1s, weekly with reports, every few weeks with everyone else. Deliberate and deeper: the full seven-to-ten-person round once or twice a year, ideally offset from the official cycle so you hear things while they're still fixable.

What if my manager never gives me anything?

Bring the structure to them. Christian Idiodi of SVPG suggests handing your leader the frame: "You are the closest person to me. You see things I will not see. Where do you think I can get better, and if you had to come up with a plan for it, what would it look like?" Then reward whatever comes back, "I got tremendous value from your feedback", because a manager's first attempt at candor is as fragile as anyone else's.

Do I share what I heard?

Share the themes and what you're doing about them, at minimum with the people who contributed. You don't need Dalio-grade transparency, but a one-paragraph "here's what I heard and what I'm changing" note is the strongest possible signal that talking to you was worth it.