Situation–Behavior–Impact: the most effective structure for feedback that creates change without triggering defensiveness. It works by separating observable fact from interpretation.
▸ Try the interactive toolSituation–Behavior–Impact is the most effective structure for giving feedback that creates change without triggering defensiveness. It works by separating observable facts — what you actually saw happen — from interpretation — what you think it means — which is what makes feedback specific, fair, and hard to dismiss.
The structure has three parts: the Situation (the specific when and where — not 'always' or 'you tend to'), the Behavior (the observable action, stated factually without judgement — what a camera would have recorded), and the Impact (the effect it had, including on you). The power is in separating behaviour from interpretation: 'you were unprofessional' is an interpretation that invites argument; 'in yesterday's meeting, you interrupted three times, and the presenter stopped sharing ideas' is observable behaviour plus impact, which is far harder to dispute and far easier to act on.
Situation — the specific when/where (not 'always').
Behavior — the observable action, what a camera saw, no judgement.
Impact — the effect it had, including on you.
Separating observable fact from interpretation is what defuses defensiveness.
Feedback that leads with interpretation — 'you were unprofessional,' 'you don't care' — triggers defensiveness and argument, because the recipient can dispute the label even when the underlying behaviour is real.
| The shortcut | What it costs | What it gives you instead |
|---|---|---|
| Leading with interpretation | 'You're disorganised' is a label that invites argument. | Stating observable behaviour is hard to dispute. |
| Vague generalisations | 'You always...' is unfair and unactionable. | A specific situation makes it concrete and fair. |
| Omitting impact | Behaviour with no stated effect seems like nitpicking. | Naming the impact shows why it matters. |
| Judgement, not observation | Mixing what you saw with what you assume. | Separating fact from interpretation keeps it fair. |
Name the precise when and where — 'in yesterday's standup,' not 'you always.' Specificity makes the feedback concrete, fair, and impossible to wave away as a vague impression.
State only what you actually saw or heard — what a camera would have recorded — with no judgement or interpretation. 'You interrupted three times,' not 'you were rude.'
Explain the effect the behaviour had, including on you or others. This is what makes the feedback matter — it connects the behaviour to a consequence worth addressing.
If you have an interpretation ('I wondered if you were frustrated'), offer it explicitly as your interpretation, distinct from the observed facts — not as a stated truth about them.
Deliver the SBI, then listen. The structure opens a fair, factual conversation about behaviour and impact — it's not a verdict delivered and walked away from.
A PM wanted to give a teammate feedback and led with interpretation: 'you've been really dismissive in meetings lately.' The teammate immediately got defensive and argued the label — 'I'm not dismissive, I just disagree' — and the real issue never got addressed. The interpretation was disputable, so it got disputed.
Restructured as SBI, it landed: 'In yesterday's planning meeting (Situation), when Sam proposed the new flow, you said “that won’t work” and moved on without asking why (Behavior). Sam stopped contributing for the rest of the meeting, and we may have missed their input (Impact).' There was nothing to argue — the behaviour was observable fact and the impact was real. The conversation moved straight to how to do it differently, instead of stalling on a contested label.
The deliverable is feedback structured Situation–Behavior–Impact — observable fact separated from interpretation, anchored to a specific moment.
| Interpretation (avoid) | SBI (use) |
|---|---|
| 'You're dismissive' | 'In yesterday's meeting (S)…' |
| 'You always interrupt' | '…you interrupted three times (B)…' |
| 'You don't care' | '…and the presenter stopped sharing (I)' |
| A disputable label | Observable fact + real impact |
SBI makes feedback land by separating what you observed from what you concluded — because observable behaviour is hard to dispute, while interpretation invites the defensiveness that derails most feedback.
The fact-versus-interpretation separation is the whole mechanism. When feedback leads with an interpretation — 'you were unprofessional,' 'you're disorganised,' 'you don't care' — the recipient can argue the label even when the underlying behaviour is undeniable, and the conversation stalls on whether the characterisation is fair. SBI sidesteps this by stating only what a camera would have recorded (the behaviour) anchored to a specific moment (the situation), then naming the real consequence (the impact). There's far less to argue with: the behaviour happened, the impact was real, and the conversation can move straight to change. If you do have an interpretation, SBI's discipline is to offer it explicitly as yours ('I wondered if…'), separate from the facts — which is both fairer and more persuasive than presenting your assumption as their truth. It's the same observable-behaviour discipline as the stakeholder archetypes' 'behaviour not people.'
Labels invite argument. State observable behaviour instead.
'You always' is unfair and unactionable. Anchor to a specific situation.
Behaviour with no stated effect seems like nitpicking. Name why it matters.
If you interpret, flag it as your interpretation, separate from what you observed.
SBI is the structure; Radical Candor (Tool 15) is the caring-and-direct spirit it's delivered in.
The recipient's side of an SBI conversation is Tool 16.
Separating observable behaviour from interpretation mirrors the archetypes discipline (Tool 05).
SBI is a core tool for surfacing issues in conflict (Tool 19).
Take a piece of feedback you might give as an interpretation ('you're disorganised', 'you were rude'). Rewrite it as SBI: the specific situation, the observable behaviour, and the real impact.
Read both versions — which one could the recipient argue with, and which gets straight to the change?
If the SBI version leaves nothing to dispute while the label invited an argument, you've felt the mechanism that makes SBI work: separating what you saw from what you concluded.