Field notes on 360-degree feedback. Short version first: it is a good tool for helping someone see themselves, and a bad tool for judging them. Most of the failures I have seen described come from using it for the second purpose. That is my opinion, but the mechanics below explain why.
What it is
A 360 collects structured ratings and comments about one person from several angles: their manager, their peers, their direct reports, sometimes internal customers, plus the person’s own self-rating. The point is the gap. The manager sees the output. Peers see collaboration. Reports see how the person behaves when the boss is not in the room. No single viewpoint has all of that.
The output is usually a report that compares self-rating against others’ average on each behaviour, with the raters grouped by relationship. The manager’s view is normally shown on its own; peers and reports are shown as groups so individual voices stay hidden.
Where it earns its place
Leadership development. A first-time manager believes she is “approachable”; her team rates her low on “gives me time”. That gap is useful, and nobody else in the company was going to tell her. The same holds for a senior person who has stopped getting honest feedback because everyone below them is careful.
It also works for culture work. If you are trying to move a firm toward collaboration, a behaviour-based 360 gives you something to track other than slogans.
Why it breaks when money is attached
Rater behaviour changes the moment the score affects someone’s increment. Peers who compete for the same promotion have a reason to mark each other down. Direct reports who fear retaliation will inflate. Friendly peers will trade favourable ratings. The data stops measuring behaviour and starts measuring the politics of the team.
There is a quieter problem too. Raters often differ in what they are even looking at, so disagreement between them is not necessarily error. A manager may see a person who protects the team from noise; a peer in another function sees someone who is slow to share information. Both are accurate. A decision rule that averages them destroys exactly the information that made the exercise valuable.
My rule: if the 360 feeds the rating, the 360 is no longer about feedback. Keep it separate, or accept that the scores will be soft.
Design decisions that matter
- Few questions, tied to behaviours the organisation has actually defined. Sixty items produce fatigue and straight-line answers.
- Rater anonymity with a minimum group size. A common rule is that a group (say, direct reports) is only reported separately if at least three people answered. With a team of two, merge the groups or drop it.
- The person chooses some raters, the manager approves the list. Self-selection alone gives you friends.
- Open comments, but screened for anything identifying or abusive before release.
- A debrief with a trained person. Handing someone a report with a few low scores and no conversation is how you create resentment.
A small illustration
Take a team lead, Arvind, rated on a five-point scale for “listens before deciding”. He scores himself 4.5. His manager gives 4. His four reports average 2.6. His peers average 3.8. The overall average, around 3.6, looks harmless. The useful information is the gap between 4.5 and 2.6 and the fact that the people closest to his decisions disagree most with him. Reduce that to a single number and you lose it.
The next step is not a rating penalty. It is a conversation: what do you think your team means? Pick one behaviour. Ask the team in three months whether it changed.
Honest limits
It will not fix a manager who does not want to change. It is also weaker in small companies where anonymity is a fiction; in a team of five everyone can guess who wrote which comment. And in a strongly hierarchical workplace, upward feedback is often too polite to be useful. In those settings, a few face-to-face conversations may do more than any platform.
One compromise some organisations use is to let the manager read the 360 themes while forming their own judgement, without importing numbers into the rating. It is not clean, and I would still lean toward keeping the exercise developmental. If you must use it in decisions, tell people so in advance and expect the answers to change.
Figures and names in the examples are illustrative.
Leave a Reply