How to run a 360 that drives change, not just data

Most 360 exercises end the same way. A report is generated, a leader reads it once, a coach walks them through the scores, everyone agrees there are useful themes, and then nothing measurable changes. Six months later the next cycle runs and the same gaps appear, sometimes wider. The instrument worked. The process did not. The 360 is one of the most widely used leadership tools in the world, and one of the most reliably wasted, because organisations treat it as a measurement event when it is really the opening move in a longer process.

The report is not the point

The common mistake is to put almost all of the design effort into the survey and almost none into what happens after it. The questionnaire gets debated, the rater list gets agreed, and the debrief and follow-through are left to take care of themselves. They never do. A 360 produces awareness, and awareness on its own changes very little. The gap between knowing you interrupt people and actually stopping is the whole problem of behaviour change, and a bar chart does not close it.

Before you run a 360 at all, it is worth being clear-eyed about what the instrument can and cannot measure, because a 360 reads perception, not capability, and a great deal of misuse comes from confusing the two. We have set that out in full in what a 360 can and cannot tell you about a leader, and this guide assumes you have made peace with those limits. This piece is the other half. Granted that a 360 is the right tool, used as one honest input among several, how do you run it so that it moves behaviour rather than just filling a folder?

The answer sits in three phases, and the weight is not where most people put it. Before, during, and after a 360 each carry real work, but the after phase is where the change is either won or lost, and it is the phase that is almost always under-resourced.

Before: purpose, framing, and safety

Everything that goes wrong in a debrief can usually be traced back to a decision made before the survey ever launched. Three decisions matter most.

Decide what the 360 is for, and say it out loud

A 360 used for development and a 360 used for evaluation are different instruments wearing the same clothes. The moment raters or recipients suspect that the data will feed a promotion decision, a performance rating, or a restructuring shortlist, the honesty drains out of it. Recipients become defensive before they have read a word. Raters either inflate to protect a colleague or deflate to settle a score. You end up measuring politics.

So decide, deliberately, that this 360 is for development, and then protect that decision in practice. That means the raw report goes to the leader and their coach, not to their manager as a scorecard, and certainly not into a talent review. If your organisation genuinely needs an evaluative read for selection, that is legitimate, but it calls for a different design and ideally a different instrument. The wider question of matching the tool to the purpose, selection versus development, is one we treat in how to choose a leadership assessment. For a developmental 360, the single most important thing you can do before launch is to make the purpose unambiguous and behave consistently with it.

Frame it as a starting point, not a verdict

How a 360 is introduced shapes how it is received. If it arrives as an audit, leaders brace. If it arrives as a mirror, they get curious. The framing email and the briefing conversation should do real work here. Say plainly that the report describes how a particular group of people experienced the leader over a particular window, that it is one input and not a judgement, and that the most useful lines in it will be the surprises rather than the confirmations. Tell recipients in advance that almost everyone finds at least one comment that stings, and that the stinging comment is often the one worth sitting with longest.

Build the psychological safety the data depends on

A 360 is only as honest as the raters feel free to be, and that freedom is not created by a privacy statement in the survey tool. It is created by the culture around it and by a few concrete design choices. Aggregate rater groups so that no individual peer or direct report can be identified from a small pool. Be honest that perfect anonymity is impossible in a team of four, and design the question set so it does not force people to choose between candour and exposure. Most of all, make sure leaders who received hard feedback in the past were not punished for the messenger. If your last 360 cycle ended with someone quietly working out who said what, the next one will return nothing but noise.

During: the right questions and the right raters

The mechanics of the survey itself are where good intentions get either supported or quietly undermined.

Ask about observable behaviour, not personality

The questions that drive change are the ones a rater can answer from memory of something they actually watched happen. "Holds people to account on commitments" is answerable. "Is a strategic thinker" is an invitation to project. The more a question asks a rater to infer an internal trait, the more the answer reflects the rater's general impression and the less it gives the leader to work with. Favour items that describe a visible action in a recognisable situation. The test is simple: could a rater point to a specific moment that justifies their score? If not, the item is measuring reputation, and reputation is hard to act on.

Keep the questionnaire short enough that raters answer the last item as carefully as the first. Survey fatigue is real, and a long instrument produces a tail of straight-lined responses that quietly corrupt the averages. Make room for free-text comments, because the narrative is almost always more useful than the numbers, and a leader will remember one specific sentence long after they have forgotten a score of 3.4.

Choose raters for range, not for kindness

Who rates matters as much as what they are asked. The instinct, especially when the leader chooses their own raters, is to assemble a friendly panel. That produces a warm, useless report. A good rater set has range: the manager above, a genuine cross-section of peers including at least one the leader finds difficult to work with, direct reports across tenure rather than only the favourites, and where the role warrants it, a stakeholder outside the reporting line. Aim for enough raters in each group that the averages can bear some weight, but not so many that the exercise collapses under its own logistics. The leader can nominate, but someone neutral, usually HR, should have the final say, precisely to keep the difficult voices in.

After: the debrief and the development that follow

This is the phase that decides everything, and it gets the least design attention. Drop a report on a leader's desk with no skilled conversation around it and you have, in the kindest reading, spent the budget for nothing, and in the worst, left a wound. A report handed over with a good conversation but no follow-through produces insight that evaporates within a fortnight. The change lives in what happens after the leader closes the file.

Run the debrief as a conversation, not a readout

The worst debriefs walk a leader page by page through their own scores, which is the one thing they can do perfectly well alone. The job of a skilled debrief is to help the leader move from defending the report to becoming curious about it. That means naming the distortions openly, recency, rater leniency, the politics of who rated whom, so the leader does not either dismiss the whole report over one unfair line or over-index on a single low score. It means resisting the urge to dwell on the highest and lowest numbers, and going instead to the most revealing line in any 360: the largest gap between how the leader sees themselves and how the room sees them. That gap is a question, not an answer, and the conversation about why it exists is where development actually begins.

A good debrief ends with the leader having chosen, in their own words, one or two things worth working on. Not a tidy list of every gap in the report. One or two. Behaviour change runs on focus, and a plan that tries to fix nine things fixes none.

Turn insight into a small number of observable commitments

Insight that is not converted into a specific, watchable change is just a more sophisticated form of self-knowledge. The move from debrief to development is the move from theme to behaviour. "Be more inclusive in meetings" is a theme. "Ask the two quietest people in the room for their view before I give mine" is a behaviour, and a behaviour can be practised, noticed, and reinforced. Help the leader translate each chosen theme into one or two concrete actions they can take this week, in real meetings, with real people.

Then build in the thing that actually drives change: a feedback loop that does not wait twelve months. The leader tells a few of their original raters what they are working on and asks those people to notice. This does two things at once. It enlists the people who supplied the data as allies in the change, and it makes the leader publicly accountable in a low-stakes way. The research on what makes development stick is consistent here. Marshall Goldsmith's long-running work on leadership behaviour change found that leaders who followed up with their colleagues and involved them in the process showed markedly greater improvement on subsequent feedback than those who received the same coaching but did not follow up. The instrument did not change. The follow-through did.

This is also where a 360 stops being able to do the job alone. A 360 tells a leader how they are currently experienced. It does not reach the patterns underneath the behaviour, the identity-level drivers that explain why a leader keeps interrupting even after they have resolved not to. That is the work of going beneath behaviour, which is where our method and the deeper instruments we use, including Ennea International's Five Lens Development Platform, do a different job. The 360 surfaces the what. The development underneath it has to reach the why, or the changed behaviour will not hold under pressure. CapabilityFX's orientation is inside-out, on the conviction that lasting change happens at the level of who a leader is, not only what they can be observed doing, and a 360 is best understood as the visible surface of that deeper work.

Two ways a 360 actually drove change

The difference between a 360 that moves behaviour and one that gathers dust is easiest to see in specific people. Both of these are representative composites.

The operations director who stopped solving everything. Consider an operations director in a manufacturing group whose 360 returned a clear, consistent theme across all his direct reports: he was trusted and respected, and he was also a bottleneck. Every exception came to him, and his team had quietly stopped trying to resolve things themselves because he always could. The numbers alone would have read as a strength: highly responsive, deeply involved. The narrative comments told the real story. In the debrief, the gap that mattered was between his self-rating as a developer of people, which was high, and his reports' rating of the same, which was low. He chose one behaviour: when a team member brought him a problem, he would ask what they thought first and resist giving the answer. He told four of his reports he was working on it and asked them to push back when he relapsed. Within a quarter the escalations had dropped, two of his managers were visibly making calls they would previously have deferred, and the next informal pulse named delegation as something that had genuinely shifted. The 360 did not change him. It pointed at one behaviour, and the follow-through did the rest.

The HR business partner who reframed the debrief. Consider an HR business partner in a retail group who had run 360s for years and was frustrated that they produced so little change. The instrument was fine. The process stopped at the readout. She redesigned the after phase. Debriefs became coaching conversations focused on the self-versus-other gap rather than score-by-score walkthroughs. Every participant left with one or two observable commitments rather than a plan listing every gap. And each leader was asked to tell a small group of their raters what they were working on and invite them to notice. The first cohort through the redesigned process showed something the old one never had: at the lighter follow-up six months later, raters reported visible change on the specific behaviours leaders had committed to, precisely where the leader had focused and asked to be watched. The data was no better than before. What changed was that the organisation finally treated the report as the start of the work rather than the end of it.

Where to start

If you run 360s already and they are not changing much, the fix is almost never a better questionnaire. It is rebalancing the effort toward the after. Put your energy into a debrief that goes to the gap rather than the score, a habit of distilling the report to one or two observable commitments, and a follow-up loop that pulls the original raters in as allies. Decide before launch that the exercise is developmental, frame it as a mirror rather than a verdict, and protect the safety the honest data depends on.

A 360 run this way earns its place. It becomes the visible surface of a development process rather than a yearly ritual that produces graphs and nothing else. If you would value a working session on where a 360 sits within a fuller capability picture, or on designing the development that converts its findings into changed behaviour, the assessments page is a useful starting point, and you can reach us directly via contact.

The leaders described here are representative composites drawn from patterns we observe in practice, not identifiable individuals.