Start with the mechanics, because they explain almost everything else. In a performance review, one manager writes an assessment of one person, signs it, and it goes into a personnel file that human resources, the next manager and possibly a lawyer can read. In a 360, eight to twelve people answer a rated questionnaire, the answers are aggregated so no single rater is identifiable, and the output goes to the person being rated.
One document has an author and a permanent audience. The other has no identifiable author and, done properly, an audience of one. Everything that follows comes from that difference.
It also explains why expectations run ahead of the evidence. A meta-analysis of 24 longitudinal studies by Smither, London and Reilly, published in Personnel Psychology in 2005, found the average improvement in ratings after multisource feedback was small: a corrected effect size of .15 for direct report ratings, .15 for supervisor ratings and .05 for peer ratings. Their own conclusion was that practitioners should not expect large, widespread performance improvement from feedback alone. So the question you are really deciding is not which process is better. It is what each one is capable of doing, and what happens to your data when you ask one to do the other's job.
Two documents, two readers, two purposes
A performance review is an administrative record. Its job is to document what a person delivered against expectations, justify a decision about pay or advancement, and create a defensible trail if the employment relationship ends badly. It is backward looking by design and it is supposed to be.
360 feedback is a development instrument. Its job is to close the gap between how a leader believes they are experienced and how they are actually experienced. It is descriptive rather than evaluative, it covers behavior rather than results, and it is most useful precisely where the manager has the least visibility.
Those two purposes are not adjacent. They pull in opposite directions the moment raters work out which one they are participating in.
What a performance review is built to do
The review answers a question with money attached. Did this person meet the commitments made at the start of the period, and what follows from that?
It works best when the commitments were written down, when the measures were agreed in advance, and when the manager has watched enough of the work to have an opinion worth recording. It works badly at senior levels, where a chief operating officer's results are shaped by a dozen other people and a market, and where the manager sees perhaps a fifth of the behavior that matters.
That gap at the top is the honest reason 360 feedback exists. Not because reviews are broken, but because a single supervisory vantage point is thin at exactly the level where behavior has the widest effect.
What 360 feedback is built to do
A 360 answers a different question. How is this person experienced by the people who work above, beside and below them, and where does that differ from their own view?
The useful output is rarely the score. It is the gap: a leader who rates themselves high on listening while seven direct reports rate it low has learned something a performance review could not have told them. Self-other agreement is the mechanism, and it is why the instrument is wasted on someone who will not be given a real debrief.
It is also worth being honest about what 360 data is not. Ratings from colleagues are perceptions, gathered at one point in time, shaped by who was asked and what happened in the preceding month. They are good evidence of reputation. They are weak evidence of capability, and they are not a measure of results.
Side by side
| Question | 360 feedback | Performance review | What breaks if you swap them |
|---|---|---|---|
| Who writes it | Peers, direct reports, manager, sometimes clients | One manager, on the record | Raters who know their words carry a pay consequence write differently |
| Who reads the output | The person rated, plus a coach or facilitator | The person, human resources, future managers | Candor collapses once raters learn the file is permanent |
| What it measures | Observed behavior and reputation | Results against agreed commitments | Behavior scores get treated as a proxy for delivery, which they are not |
| Time orientation | Forward, toward changed behavior | Backward, over a defined period | Development conversations turn into negotiations about the past |
| Anonymity | Aggregated so individual raters are not identifiable | Attributed and signed | Anonymous input used for decisions is difficult to defend or appeal |
| What it decides | Nothing directly; it informs a development plan | Pay, rating, promotion, sometimes exit | Colleagues start rating strategically rather than honestly |
What happens when 360 ratings set pay
This is the specific risk, and it is measurable rather than theoretical. A 1997 meta-analysis by Jawahar and Williams, published in Personnel Psychology, found that ratings collected for administrative decisions ran nearly one third of a standard deviation higher than ratings collected for research or development purposes. Same raters, same behavior, different stated purpose, systematically more generous numbers.
Apply that to a 360 and the instrument stops working. The gap between self and others narrows, not because the leader improved, but because raters became kind. A program whose entire value is the accuracy of that gap has just lost it.
The second effect is political. SHRM's assessment of the practice records practitioners warning about misleading input from people competing for the same promotion, and about coalitions forming among raters. Anonymous ratings with money attached create an obvious incentive, and some people take it.
The third effect is that you cannot defend the decision. An employee told they missed a promotion because of anonymous peer ratings has no way to examine the evidence, and you have no way to show the ratings were fair. That is a bad position in a grievance and a worse one in litigation.
None of this means feedback is harmless when it is kept developmental. A 2019 rapid evidence assessment from the Center for Evidence Based Management noted that Kluger and DeNisi's meta-analysis of 131 controlled studies with 12,652 participants found an average effect size of .41, while also finding effects highly variable, with feedback in some situations producing no improvement at all or making performance worse. Feedback is a strong intervention with a real downside, which is an argument for running it carefully rather than for running it as an appraisal.
Running both without contaminating either
Most organizations need both, and the two can coexist if you separate them deliberately.
Separate them in time. Put several months between the 360 cycle and the review cycle, so raters are not answering one questionnaire with the other in mind.
Separate them in ownership. The 360 report belongs to the individual and their coach. The manager should hear the themes the individual chooses to share and the resulting development commitments, not the item-level scores.
Separate them in language. Say plainly, in writing, that 360 results are not used to set ratings, pay or promotion, and then hold that line the first time a senior leader asks for someone's scores. The program's credibility is decided in that one moment.
What can legitimately cross over is the commitment. If a leader receives feedback, chooses two behaviors to work on and tells their manager, then progress against those commitments is fair to discuss in a review. The plan is shared. The raw data is not. The same boundary applies when talent data feeds a 9 box talent review: bring the conclusion the individual has agreed to, not the anonymous item scores behind it.
Why the second cycle is quieter than the first
- Nothing visibly changed. Participants spent forty minutes rating colleagues, watched nobody behave differently, and concluded the exercise was ceremonial. They are not wrong, and they will rate faster and blander next time.
- The debrief was skipped or rushed. A report delivered without a trained conversation is a list of numbers about which the recipient is defensive. Almost all the value of a 360 is created in the hour after it lands, and that hour is the first thing cut when budgets tighten.
- No commitment was made in public. Insight with no named action and no follow-up date decays inside a quarter. The second cycle then measures the same behaviors and finds them unchanged, which is read as proof that the tool does not work.
- Anonymity was broken once. A leader worked out, or claimed to work out, who wrote a comment. One incident of that travels across the organization faster than any communication plan, and subsequent cycles get nothing but safe answers.
- The questionnaire measured the wrong things. Generic competency items produce generic feedback that no leader recognizes as being about them. If the items do not reflect what this organization actually asks of its leaders, the output is not actionable and the program deserves to fade.
- It became an annual event. Once a year is too infrequent to build a habit and too heavy to run well. The programs that survive attach feedback to a live development question rather than a calendar slot.
Start with one page and two hours
Before you commission anything, write a single page that answers two questions: which decisions in this organization will 360 data touch, and which it will never touch. Then get the head of human resources and one line executive to sign it. That is the whole first step, and it costs about two hours of your time plus one meeting.
It is a cheap step and it is the one that decides whether the program works, because every later failure traces back to an ambiguity on that page. If you cannot get agreement that 360 output stays out of pay decisions, you have learned something important before spending money, and the right move is to fix the appraisal process instead.
After that, the sequence is short: pick behaviors your leaders are actually accountable for, run a small pilot with a proper debrief for every participant, and check at sixty days whether anyone did anything differently. If choosing the right instrument is the harder part, our leadership assessment work covers that design decision, including where a personality instrument fits alongside feedback, which we compare in Hogan vs DISC.
Frequently Asked Questions (FAQs)
Can 360 feedback replace performance reviews?
No. A 360 measures how a person is experienced by colleagues, not what they delivered against commitments, and it carries no attributable author. You still need an evaluative record for pay and advancement decisions. The two answer different questions and removing either leaves a gap.
Should 360 feedback be used for promotion decisions?
Not as a rating input. Once raters understand their answers affect someone's advancement, the evidence shows ratings become systematically more lenient and the honest signal disappears. If you want colleague input into a promotion, gather it through attributed, structured reference conversations rather than an anonymous questionnaire.
Is 360 feedback really anonymous?
It is anonymous in aggregate, not absolute. With a small rater group, or a distinctive written comment, people can often guess the source. Use a minimum of three to five raters per category before reporting a group, and edit verbatim comments for identifying detail, or promise less than you can deliver.
How many raters should a 360 have?
Eight to twelve is the usual working range for a senior leader, including the manager, four to six direct reports and three to five peers. Fewer than that makes the categories unreportable. Many more raises the burden on the organization without improving the signal much.
How often should a 360 be run?
Every eighteen to twenty four months for an individual leader is typical, which is long enough for behavior change to be visible. Running it annually usually measures noise and creates survey fatigue. Attach it to a development question or a role transition rather than a fixed calendar slot.
What if 360 results contradict the performance review?
That contradiction is information, not an error. A leader can hit every number while being difficult to work for, and the two instruments are correctly reporting different things. Treat the divergence as the agenda for the next conversation rather than as a reason to discredit one of the sources.
Who should deliver the 360 debrief?
Someone trained in feedback who is not the person's manager. The manager has a stake in the content and the recipient knows it, which changes how honestly the conversation can go. An internal coach, a skilled human resources partner or an external coach all work, provided they have actually been trained on the instrument.