IX Coach

The SBI Feedback Model, Made Practical

Situation, Behavior, Impact — the structure that takes the sting out of honest feedback

The SBI Feedback Model: feedback without defensiveness

The SBI model (Situation–Behavior–Impact), developed by the Center for Creative Leadership, structures feedback in three parts, delivered in that order. Situation anchors the conversation to one specific, verifiable moment — "in yesterday’s client call, about ten minutes in" — so you are discussing an event you both remember rather than a pattern the other person can dispute. Behavior describes only what a camera would have captured — "you cut across Priya twice while she was answering the pricing question" — not the motive you inferred from it ("you were dismissive"), because motive is the part people defend. Impact names the concrete effect, on the work or on you — "the client never got her answer, and I saw Priya stop volunteering after that." A worked example, end to end: "In yesterday’s client call (S), you cut across Priya twice while she was answering the pricing question (B). The client left without an answer, and Priya went quiet for the rest of the meeting (I)." This structure reduces defensiveness and attribution errors precisely because it never asserts anything the other person can honestly deny. The model is established clinical practice in leadership development, though direct RCT comparisons to other feedback approaches are limited.

Guided practice

Prefer to listen?

Transcript

In this explainer, you will prepare one real piece of feedback you owe someone using the situation, behavior, and impact structure, and it's going to take about 5 minutes. Now, you won't actually deliver it during this session. Instead, you'll walk away with a finished three-part draft that you can say out loud when the time is right. This is especially helpful if your feedback usually triggers defensiveness because it can often sound like a judgment of character.

Setup and safety. Before we begin this guided methodology from IX Coach, please go ahead and choose one recent real situation where you need to give someone feedback just so the details are really clear in your mind. Also, make sure your environment is ready. Have a timer where you can see it and keep your phone on silent. No writing is needed here, though you absolutely may say your draft out loud as we go.

Just a quick safety check before we dive in. If this brings up more than you can hold, please stop. This is coaching. It's not therapy. If the situation involves someone who is still causing you harm, put your own safety first. You absolutely do not have to prepare feedback about it.

Today, we'll be using a simple three-step framework: situation, behavior, and impact. This works so well because it removes the exact things that make people defensive. You know, things like the vague "you always," guessing their motives, or making them feel accused.

Let's move to step one, the situation. We're going to begin by anchoring your feedback to a single very specific situation. I want you to think of one recent moment. Name the time, the place, and what the event was. For example, in the team meeting yesterday morning or maybe the email you sent at 3:00 in the afternoon. Stay with that one moment. Resist the urge to reach for you always or you never. I'm going to give you 30 seconds to think and make that moment concrete. That's 15 seconds. You're halfway there. Just make sure the moment is concrete enough that the other person would clearly remember it, too. And time. Now, take another 30 seconds to say your situation sentence out loud to yourself. Remember to keep it to just that one single moment. Go ahead. 15 seconds left. And time. We're just going to take a brief 10-second pause to let that settle.

All right, step two, the behavior. Next up, we need to describe the behavior, not the character. Think about how to describe only what a video camera would have captured. What did the person actually do or say? Use actions, not labels. So, you want to completely remove words like rude, careless, or dismissive. Those words describe who they are. And that's the exact part people get defensive about. You really have to watch out for a label hiding inside an action. Saying something like, "You are being passive aggressive by..." still judges who the person is, even if it kind of sounds like a description. Find the pure action underneath that label and say that action instead.

Take 30 seconds to draft this behavior sentence in your head. We're halfway there. Check your sentence real quick. Could someone watching the moment actually see this happen? And time. For the next 30 seconds, I want you to say your behavior sentence out loud. If you catch yourself saying you are, just gently change it to you did. 15 seconds left and time. Let's take 10 seconds to just breathe.

The impact. Now, we're going to name the specific impact of that behavior. State what the effect was. Whether that's on the work, on the team, or on you. Be specific and honest about the exact outcome. Maybe it was a missed deadline, a colleague who stopped speaking up, or a client who didn't get an answer. Be honest here and be careful not to make the impact bigger than it actually was. Take 30 seconds to form this thought. Halfway there. Remember, if you're speaking about your own experience, make sure to use the word I to name it. And time. Now, take 30 seconds to say your impact sentence out loud. 15 seconds left. And time. Let's take another 10-second pause just to let that settle in.

Solid. Putting it together. It's time to bring all three parts of your draft together. Say your draft in order out loud. The situation, the behavior, and then the impact. It should sound something like this. In one moment, you did this. The effect was that. As you say it, notice exactly how the sentence sounds. Completely free of any verdict or judgment. We'll pause for 20 seconds so you can review your final completed feedback statement. 10 seconds left and time.

Finally, our final reflection. Take one deep breath and quietly consider this question. Which part of the draft was hardest to keep to observable facts? When you eventually deliver this feedback, just remember to leave room for the other person to respond. I invite you to prepare your next piece of feedback tomorrow using these exact same three steps.

Most feedback fails not because it is wrong but because it is vague, personal, or arrives without context — triggering defensiveness instead of reflection. The Center for Creative Leadership’s SBI model solves this by separating what happened (situation), what was done (behavior), and what resulted (impact). Each element does specific psychological work. Situation removes the "when? that never happens" escape hatch by pinning the conversation to one shared, checkable moment. Behavior removes the thing people actually defend — not the act, but the motive someone attributed to it — by restricting you to observable description. Impact supplies the reason to change, which pure description lacks on its own: without it you have reported an event, not made a case. The order matters as much as the parts, because each one earns the right to the next: agree on the moment, agree on what happened in it, and only then discuss what it cost. Reversing that order — leading with impact — is how SBI collapses back into the accusation it was designed to replace. Below are the core practices, each with the mechanism behind it and a calibrated read on the evidence.

Practices

How these practices connect

Every practice in this model manages the same variable: how much threat-response the receiver's brain generates before it can process the content of the feedback. Specificity (of situation, of behavior, of impact, of request) narrows the claim being made and gives the receiver something factual to evaluate rather than a global judgment to defend against. Dialogue and timing manage the same threat response from the outside, by controlling when the conversation happens and whether the receiver has a voice in it. The whole model is a set of independent levers on the same nervous-system bottleneck.

Do you have to use all the SBI steps every time, or can you just describe the behavior?

Situation, Behavior, and Impact function as a chain, not three optional add-ons. Naming the situation narrows a global-sounding complaint down to one moment; describing the behavior instead of the character keeps that moment factual instead of an identity attack; naming the impact gives the receiver a reason to care that isn't just compliance. Dropping the situation reintroduces the vague-generalization problem that makes people defensive in the first place, and dropping the impact leaves the behavior description sounding like an arbitrary rule. You can use behavior-only in a very low-stakes moment, but the three core steps are load-bearing together.

Why does it matter if I say 'you're being defensive' instead of 'you interrupted me three times'?

"You're being defensive" is a character attribution — a claim about who the person is — and character attributions trigger identity-threat, which activates the amygdala and narrows the receiver's bandwidth for reflection. "You interrupted me three times" is a behavioral description: it names an observable action the person can agree or disagree with on the facts, without it being an attack on their self-concept. This is also the most common way people accidentally break SBI — dressing a character judgment in behavioral-sounding language, like "you were being passive-aggressive by," which is still a character claim underneath.

Should I ask the other person what they think after giving feedback, or just say my piece?

Ending with an open question is a distinct, necessary step, not a politeness add-on. Feedback delivered as a monologue leaves the receiver in a passive, one-down position that blocks real reflection, while inviting a response activates their sense of agency and turns the exchange into a two-way calibration — you may have the situation, the behavior, or the impact slightly wrong, and their response is the only way you find that out. The catch is that the invitation has to be a genuine open question; asking "does that make sense?" signals you want agreement rather than their actual perspective, and shuts the dialogue back down.

Is it enough to just tell someone what they did wrong, or do I need to tell them what to do differently?

Impact and forward-request answer two different questions, and both are usually needed. Naming the impact tells the receiver why the behavior mattered — the causal story of what happened as a result — but on its own it can leave them knowing something was wrong without knowing what right looks like. Pairing the impact with a specific, doable request for next time gives the receiver a concrete target they can mentally rehearse, which is where implementation-intention research says behavior change actually happens. An open-ended request like "just think about it" fails for the same reason a missing request does: it puts the burden of figuring out the fix entirely on the receiver.

Can you use SBI for praise, or is it just for criticism?

SBI works the same way for positive feedback, and using it that way changes how the critical version lands later. Generic praise like "great job" carries almost no information and its reward signal fades fast, while SBI-structured praise — this situation, this specific behavior, this specific impact — tells someone exactly what to repeat and why it mattered. If SBI is only ever used for correction, people learn to hear the structure itself as a threat signal, which undercuts the safety that makes the corrective version work well. Using it for both is what builds the trust that critical feedback depends on.

Does it matter where or when I give someone feedback if I get the wording right?

Wording and setting operate on the same threat-detection system but from different angles, so getting the words right does not compensate for the wrong moment. Threat-detection is context-sensitive: the same sentence said publicly instead of privately, or delivered right after a stressful event instead of in a calm window, produces very different levels of amygdala activation regardless of how carefully it's phrased. A receiver in that activated state cannot access the reflective processing that a well-built SBI statement is designed to reach, so a poorly timed, well-worded statement can still fail. The common failure is giving corrective feedback in the room where the event happened, in front of others, for the sake of efficiency — trading the receiver's openness for an audience's approval.

What's the actual difference between naming the situation and describing the behavior — aren't they the same thing?

They narrow different parts of the claim. Anchoring to a situation narrows the scope in time and place — "last Tuesday's all-hands" instead of "you always do this" — which prevents the receiver from hearing a sweeping generalization about their character over time. Describing the behavior narrows the content of the claim itself — the specific observable action, not an interpretation of it — which keeps the statement disputable on facts rather than being about the receiver's self-concept. A statement can get the situation right and still fail if the behavior description smuggles in a character judgment, or vice versa; they close two separate gaps in the same sentence.

How do you choose between these?

The right combination depends on what the moment requires, not on who you're giving feedback to. If the stakes are low and the moment is casual, situation plus behavior plus impact is often enough. If the behavior needs to change and not just be understood, add the forward-looking request, since impact alone can tell someone what went wrong without giving them anywhere to go. If the relationship or the topic is sensitive, invite-a-response matters more, because you're more likely to have part of the story wrong and a monologue forecloses correction. Timing and privacy sit underneath all of it as a precondition: no combination of wording fixes a conversation held in the wrong place or the wrong moment, so that choice comes before the others, not alongside them. And positive SBI isn't reserved for low-stakes moments — the practices note it should run continuously alongside corrective feedback, not just when something needs fixing.

What these approaches do not settle

Most of these mechanisms are described as mechanistic or clinical rather than confirmed by trials isolating that specific component of SBI — the amygdala and identity-threat research is well established in general, but whether "behavior vs. character" framing or situational anchoring specifically modulates it inside a feedback conversation is inferred, not directly tested. The forward-looking request has the strongest evidence base, drawing on implementation-intention research, though whether people apply that mechanism on their own or only when explicitly coached to hasn't been isolated either. The positive-feedback evidence is the thinnest of the seven: most specificity-of-praise research comes from educational settings, and its generalization to adult professional feedback is plausible but less directly studied.

Practice this with IX Coach

Practice this with IX Coach

Frequently asked questions

The SBI Feedback Model: feedback without defensiveness

The SBI model (Situation–Behavior–Impact), developed by the Center for Creative Leadership, structures feedback in three parts, delivered in that order. Situation anchors the conversation to one specific, verifiable moment — "in yesterday’s client call, about ten minutes in" — so you are discussing an event you both remember rather than a pattern the other person can dispute. Behavior describes only what a camera would have captured — "you cut across Priya twice while she was answering the pricing question" — not the motive you inferred from it ("you were dismissive"), because motive is the part people defend. Impact names the concrete effect, on the work or on you — "the client never got her answer, and I saw Priya stop volunteering after that." A worked example, end to end: "In yesterday’s client call (S), you cut across Priya twice while she was answering the pricing question (B). The client left without an answer, and Priya went quiet for the rest of the meeting (I)." This structure reduces defensiveness and attribution errors precisely because it never asserts anything the other person can honestly deny. The model is established clinical practice in leadership development, though direct RCT comparisons to other feedback approaches are limited.

How do you anchor feedback to a specific situation?

Feedback without a situational anchor forces the receiver to guess what you mean, which activates threat-response as they search for the "right" moment to defend. Anchoring to a specific context (last Tuesday’s all-hands, the email you sent at 3 pm) narrows the scope from "you always do this" to "this happened once," which is cognitively and emotionally easier to accept and act on.

How do you describe behavior, not character?

Character attributions ("you’re dismissive") trigger identity-threat, activating the amygdala and narrowing cognitive bandwidth for reflection. Behavioral descriptions ("you interrupted three times") are observable and disputable on facts rather than self-concept, which keeps the receiver in a cooler, more coachable state.

How do you name the actual impact on you or the team?

The impact statement converts abstract criticism into a causal story the receiver can evaluate: "when X happened, Y resulted." It also shifts the emotional register from judgment ("this was wrong") to consequence ("this affected me / the team in a concrete way"), which activates empathy rather than defensiveness and gives the receiver a reason to care that isn’t just rule-compliance.

How do you invite a response after delivering SBI?

Feedback delivered as a monologue leaves the receiver in a passive, one-down position that blocks genuine reflection. Opening space for response activates the receiver’s sense of agency and turns the conversation into a dialogue, which is essential for accurate two-way calibration: you may have the situation, behavior, or impact wrong, and a response gives you the data to correct this.

How do you add a forward-looking request to SBI?

Feedback without a clear forward request gives the receiver nowhere to go: they know what was wrong but not what "right" looks like. A specific behavioral request activates implementation-intention processes — the receiver can mentally rehearse the new behavior and plan for situations that trigger it, which is where real change happens.

How do you use SBI for positive feedback, not just correction?

Generic praise ("great job") carries almost no information and its reward signal fades quickly. Specific SBI-structured positive feedback tells the person precisely what to repeat and why it mattered, which reinforces the behavior through both informational learning and the stronger affective hit of understood appreciation.

How do you choose the right moment and setting for feedback?

Threat-detection is context-sensitive: the same words said publicly versus privately, or right after a stressful event versus in a calm window, trigger very different levels of amygdala activation. A receiver in threat mode cannot access the reflective processing needed to act on feedback, so optimal timing reduces the neurological barrier before the content even lands.

Do you have to use all the SBI steps every time, or can you just describe the behavior?

Situation, Behavior, and Impact function as a chain, not three optional add-ons. Naming the situation narrows a global-sounding complaint down to one moment; describing the behavior instead of the character keeps that moment factual instead of an identity attack; naming the impact gives the receiver a reason to care that isn't just compliance. Dropping the situation reintroduces the vague-generalization problem that makes people defensive in the first place, and dropping the impact leaves the behavior description sounding like an arbitrary rule. You can use behavior-only in a very low-stakes moment, but the three core steps are load-bearing together.

Why does it matter if I say 'you're being defensive' instead of 'you interrupted me three times'?

"You're being defensive" is a character attribution — a claim about who the person is — and character attributions trigger identity-threat, which activates the amygdala and narrows the receiver's bandwidth for reflection. "You interrupted me three times" is a behavioral description: it names an observable action the person can agree or disagree with on the facts, without it being an attack on their self-concept. This is also the most common way people accidentally break SBI — dressing a character judgment in behavioral-sounding language, like "you were being passive-aggressive by," which is still a character claim underneath.

Should I ask the other person what they think after giving feedback, or just say my piece?

Ending with an open question is a distinct, necessary step, not a politeness add-on. Feedback delivered as a monologue leaves the receiver in a passive, one-down position that blocks real reflection, while inviting a response activates their sense of agency and turns the exchange into a two-way calibration — you may have the situation, the behavior, or the impact slightly wrong, and their response is the only way you find that out. The catch is that the invitation has to be a genuine open question; asking "does that make sense?" signals you want agreement rather than their actual perspective, and shuts the dialogue back down.

Is it enough to just tell someone what they did wrong, or do I need to tell them what to do differently?

Impact and forward-request answer two different questions, and both are usually needed. Naming the impact tells the receiver why the behavior mattered — the causal story of what happened as a result — but on its own it can leave them knowing something was wrong without knowing what right looks like. Pairing the impact with a specific, doable request for next time gives the receiver a concrete target they can mentally rehearse, which is where implementation-intention research says behavior change actually happens. An open-ended request like "just think about it" fails for the same reason a missing request does: it puts the burden of figuring out the fix entirely on the receiver.

Can you use SBI for praise, or is it just for criticism?

SBI works the same way for positive feedback, and using it that way changes how the critical version lands later. Generic praise like "great job" carries almost no information and its reward signal fades fast, while SBI-structured praise — this situation, this specific behavior, this specific impact — tells someone exactly what to repeat and why it mattered. If SBI is only ever used for correction, people learn to hear the structure itself as a threat signal, which undercuts the safety that makes the corrective version work well. Using it for both is what builds the trust that critical feedback depends on.

Does it matter where or when I give someone feedback if I get the wording right?

Wording and setting operate on the same threat-detection system but from different angles, so getting the words right does not compensate for the wrong moment. Threat-detection is context-sensitive: the same sentence said publicly instead of privately, or delivered right after a stressful event instead of in a calm window, produces very different levels of amygdala activation regardless of how carefully it's phrased. A receiver in that activated state cannot access the reflective processing that a well-built SBI statement is designed to reach, so a poorly timed, well-worded statement can still fail. The common failure is giving corrective feedback in the room where the event happened, in front of others, for the sake of efficiency — trading the receiver's openness for an audience's approval.

What's the actual difference between naming the situation and describing the behavior — aren't they the same thing?

They narrow different parts of the claim. Anchoring to a situation narrows the scope in time and place — "last Tuesday's all-hands" instead of "you always do this" — which prevents the receiver from hearing a sweeping generalization about their character over time. Describing the behavior narrows the content of the claim itself — the specific observable action, not an interpretation of it — which keeps the statement disputable on facts rather than being about the receiver's self-concept. A statement can get the situation right and still fail if the behavior description smuggles in a character judgment, or vice versa; they close two separate gaps in the same sentence.

Research

  • Ross (1977), the fundamental attribution error, Advances in Experimental Social Psychology
  • Eisenberger & Lieberman (2004), social pain and physical pain share neural substrates, Science
  • Gollwitzer & Sheeran (2006), implementation intentions meta-analysis (d ≈ 0.65), Advances in Experimental Social Psychology About this work

Related concepts

Articles that cover this

Why this matters

The guided video and audio on this page are narrated by AI voices.