Most coaching engagements open with two accounts of the same person. The coachee describes themselves, and the sponsor who commissioned the work describes why they commissioned it. Both are useful. Both are partial in ways that are fairly predictable.
The coachee reports intent. They can tell you what they were trying to do in the meeting that went badly, which is worth knowing and is not the same thing as what the room experienced. The sponsor reports the version that justified a budget, which tends to compress a complicated situation into one or two headline behaviors.
A 360 does not replace either account. It adds the thing neither can supply, which is several independent accounts from people in different relationships to the coachee, collected at a known point in time.
The diagnostic
The value is not an average. It is the disagreement.
Peers, direct reports and supervisors see different parts of the same person, and when their answers diverge, the divergence is the finding. A leader whose reports rate them well on clarity and whose peers rate them poorly on it is not inconsistent. They are behaving differently in two settings, and that is a far more tractable thing to coach than a general instruction to communicate better.
The self-versus-other comparison does similar work. Where a coachee rates themselves well above their raters, you have the agenda for the first session, and you have it in a form that does not depend on you being persuasive. Where they rate themselves below, which happens more often than people expect with capable leaders, the useful conversation is about why they cannot see something everyone around them can.
Because a 360 is dated, it also establishes a baseline. A repeat at six or twelve months measures the engagement against the record rather than against the coachee's memory of where they started, and memory moves in the direction of effort.
A plan with fewer items in it
The common failure in a coaching plan is scope. Twelve development areas is a list, not a plan, and a coachee who leaves the first session with twelve things to work on will work on none of them.
A 360 narrows it by evidence rather than by instinct. Two behaviors, chosen because they are what the people around the coachee actually raised, and ordered by how widely they were raised. That ordering is hard to argue with, and the coachee is less likely to spend the engagement negotiating the diagnosis.
It also shows where somebody is already strong, which is systematically under-used. Strengths tend to get a sentence in a readout and then disappear from the plan. Knowing where a coachee is rated highly tells you where to spend less time, which is how the plan stays short enough to act on.
Clarity, which is mostly about specificity
Coachees rarely dispute that they should improve. What they cannot act on is an adjective. Being told they come across as remote is not actionable; being told that in three of four rater groups people reported not knowing where a decision had landed is.
Reporting by rater group rather than as a single score is what preserves that. An average across peers and reports hides opposite signals and produces a middling number that describes nobody's experience. Split the groups and the picture is specific enough to do something with.
Why administration matters more than the instrument
A well-designed 360 administered by the coach produces the same problem a well-designed upward review administered by the firm produces. The raters work for, or alongside, the person being assessed, and they are weighing who will see the answer. If they believe the coach holds the responses, and that the coach talks to the coachee weekly, they answer accordingly.
Small pools make this sharper rather than softer. A coachee might have eight to twelve raters across three groups, and a group of three direct reports cannot be reported as a group without effectively identifying them. Which groups will report, and against what minimum, has to be settled before anyone is invited, because once the responses exist the threshold cannot be relaxed to produce a cut somebody wanted.
The arrangement that makes raters candid is simple to describe and has to be true. Responses go to an administrator outside both the coachee's organization and the coaching practice. The report reaches the coachee and the coach. The employer receives nothing. Raters are told all of this before they answer, because being told is the part that changes what they write.
SRA administers 360s for coaching practices on that basis, using the rater group design and suppression discipline built for firm-wide programs and scaled to one coachee at a time.
Three things a 360 is not
It is not an evaluation. Nothing in a developmental 360 should reach a performance record or a compensation decision, and if the organization paying for the coaching wants a report of its own, that is a different instrument bought on a different basis. Running one exercise and letting it serve both purposes is how raters learn not to trust the next one.
It is not a personality assessment. A 360 reports observed behavior from people who work with the coachee. It does not explain why, and the explanation is the coaching.
And it is not much use once. A single assessment is a snapshot with no direction of travel. The repeat is where a coach can show a sponsor that something moved, and it is the only part of the process that answers the question sponsors actually ask.
Practical notes
Eight to twelve raters is usually enough, spread so that no group is so small it cannot be reported. Invite more than your minimum, because response rates in small pools are volatile and losing two raters from a group of four changes what you can publish.
Field before the plan is written rather than after. A 360 used to confirm a plan the coach has already decided on is an expensive formality, and the raters can generally tell.