Skip to main content

Feedback: Giving, Receiving, and the SBI Model

What This Concept Is​

The SBI model (Center for Creative Leadership) is a three-part structure for giving feedback:

  • Situation. When and where it happened. Specific time, specific context.
  • Behavior. What the person did, in observable terms. Not intent, not personality, not inference.
  • Impact. What effect it had on you, the team, or the outcome. Named specifically.

The sharpest version of the model adds a fourth step, Intent (SBI-I), where you ask the receiver what they were trying to do, rather than assume. Feedback is then a conversation, not a verdict.

Receiving feedback is a separate skill. The move there is: listen, thank, separate content from delivery, act. Do not defend in the moment; defending prevents the next piece of feedback from arriving.

Why It Matters Here​

Feedback in engineering teams fails in predictable ways:

  • Identity feedback. "You're not detail-oriented enough." Not actionable. Triggers defense. Gives the receiver nothing to change.
  • Vague praise. "Good job on the launch!" Feels nice; teaches nothing.
  • Drive-by negativity. A sharp comment in Slack, no context, no impact named. Receiver infers hostility.
  • Feedback that never arrives. Manager or Staff+ engineer observes a problem, says nothing, writes it into the performance review six months later. Maximum damage, zero growth.

A Staff+ engineer is a net exporter of feedback - formal and informal, up and down the org. If the feedback is not well-structured, it is either useless or actively harmful.

Concrete Example​

Weak feedback.

"You were really disruptive in the design review yesterday. Try to be more collaborative."

Problems: no situation (which review?), no behavior (what exactly did they do?), no impact (what got worse?), identity-flavored ("disruptive"). The receiver cannot act on this and is likely to feel attacked.

SBI feedback.

Situation. "In yesterday's design review for the payment-gateway RFC, around the 20-minute mark." Behavior. "You interrupted the author three times in five minutes and rewrote their proposed solution on the whiteboard before they'd finished describing their design." Impact. "The author stopped presenting and the rest of the review turned into a debate between you and them. The two junior engineers on the call didn't speak again. I think we lost their input on the consistency tradeoff." Intent (ask). "I don't think that was your intent - what were you going for?"

This is a conversation. The receiver can respond, "I was trying to prevent us from going down a dead end I'd seen before," and you can respond, "That's useful context. Could you hold that until the author finishes next time?" Now the feedback is actionable.

Receiving feedback​

When receiving feedback, the moves in order:

  1. Listen all the way through. Do not interrupt or defend.
  2. Thank them. Short. "Thank you for telling me." The thanks is not agreement; it is acknowledgement that they spent social capital.
  3. Ask clarifying questions. "Can you name another specific time?" "What impact did you see?" The goal is clarity, not counter-argument.
  4. Separate content from delivery. Bad delivery of true feedback is still true feedback. Note the delivery problem separately.
  5. Act on it. Name one change you will try. Come back in two weeks and report.

The senior engineers people trust most are the ones who can receive hard feedback without a defensive reaction. That trust is earned once and spent for years.

Common Confusion / Misconception​

"Feedback must be positive to be safe." Unbalanced feedback, in either direction, is not feedback; it is marketing. Engineers know when they are being praised to avoid difficult conversation. Honest mixed feedback is higher-trust than one-sided positive feedback.

"The 'feedback sandwich' (praise-critique-praise) is the safe format." The sandwich is usually transparent and therefore condescending. SBI is a more respectful structure: you treat the receiver as a peer who can handle the behavior and impact directly.

"Feedback should be private always." Private is the default, but not absolute. Public recognition of behavior worth modeling is also feedback and is appropriate in team settings. Never deliver critical feedback publicly.

"Impact = my feelings." Impact can include your feelings but should usually include something observable: a meeting that stalled, a decision that was delayed, a junior engineer who stopped speaking up. "I felt bad" is weaker than "the reviewers could not finish the agenda."

How To Use It​

Giving feedback: S - B - I - ask for I (intent).

  1. Name the specific situation. If you cannot, the feedback is not ready.
  2. Describe behavior, not person. Verbs, not adjectives. "You interrupted three times" not "you're rude."
  3. Name a specific impact. Observable. Not attributed to their intent.
  4. Ask about intent. "What were you going for?" Converts verdict to conversation.

Receiving feedback: listen - thank - clarify - separate - act.

Check Yourself​

  1. Name the three letters of SBI and what each one requires.
  2. Why is "you're disruptive" worse feedback than "you interrupted the author three times in five minutes"?
  3. What is the correct first move when receiving hard feedback?

Mini Drill or Application​

Take one piece of feedback you have given in the last month (or should have given and did not). Rewrite it in SBI form:

  • Situation: specific time and setting, one sentence
  • Behavior: observable verbs, no adjectives about the person, 1-2 sentences
  • Impact: named effect on people, decisions, or outcomes, 1-2 sentences
  • Intent ask: one question

Then deliver it. Observe what happens. Compare the conversation to what would have happened with your original phrasing.

Upward Feedback: The Hardest Case​

Giving feedback to your manager or a more senior leader is where SBI most often fails. Three rules make it workable:

  • Ask before delivering. "I have some feedback from the last planning meeting; is this a good time to share it?" gives your manager control and stops them from being ambushed.
  • Keep the Impact section business-focused, not personal. "When the roadmap was published without the team's input (B), two engineers flagged privately that they didn't understand how their Q2 work fit (I)." This is easier to act on than "You made me feel excluded."
  • Offer a path. "One option would be to circulate the draft for 48 hours before publishing." You are not their manager, and this is not a demand - it's a suggested experiment, which leaves their authority intact.

Upward feedback given well is one of the rarest behaviors in engineering organizations; it is also one of the most trust-building. Managers who receive crisp, specific, behavior-level feedback from their senior ICs quickly learn to rely on it and often invite more.

Downward feedback to direct reports has the inverse risk: they may feel they cannot push back. Pair SBI with the explicit invitation to disagree: "That's what I observed; I could be wrong about the impact. What's your read?"

Transfer / Where This Shows Up Later​

Feedback is the repair mechanism for every other concept in the module:

  • Archetype and altitude (Concepts 1-2). Drift is diagnosed through feedback. "You've been spending most of your time on code-altitude work this quarter; the team missed the systems-altitude reviews" is role-shaping feedback.
  • Strategy and roadmaps (Concepts 4-6). Review-trigger firings are feedback to the strategy's authors. Roadmap slippage is feedback about capacity estimation.
  • Stakeholder mapping and disagree-and-commit (Concepts 8-9). Stakeholder complaints are feedback about the map; disagreement memos are feedback in structured form.
  • Communication (Concepts 10-12). "I didn't understand the exec summary" is feedback; a careful SBI version - "In the Monday exec review (S), when the first sentence named three decisions (B), two VPs asked follow-up questions we had to schedule a second meeting for (I)" - is how you improve.
  • Sponsorship (Concept 13). Sponsors give feedback that mentors cannot afford to give. A sponsor says "In yesterday's review, you let the architect drive when the RFC was yours; you need to own the room" because the sponsor has already staked credibility on the sponsoree.
  • Personal strategy (Concept 15). Feedback is input to your own review triggers; the annual strategy reopens when the feedback you receive contradicts the plan.
  • Semester 8 and beyond. M1 interview practice is a feedback-dense loop; M2-M4 retrospectives are SBI opportunities; Semester 10 M5 uses feedback as the primary input to staff+ career navigation.

Read This Only If Stuck​

No local book chunk covers SBI directly. Closest local chunks: