How do I give feedback about someone’s behaviour rather than their personality?

Labelling someone as “unprofessional,” “disorganised,” or having a “bad attitude” rarely leads to performance improvement. It usually triggers defensiveness, inflames conflict, and shifts the conversation away from work outcomes and toward character defence.

The issue is almost never that your underlying concern is invalid. The issue is that you are delivering an inference rather than an observation.

When leaders confuse their interpretations with facts, feedback ceases to be a tool for development and becomes a verdict on character. To give behavioural feedback that drives actual change, you must learn to separate the raw facts of an event from the judgements you assign to them.

Why leaders default to personality critiques

If focusing on behaviour is so effective, why do most leaders naturally default to personality labels?

First, personality labels are cognitive shorthand. It takes very little mental effort to say, “She’s unreliable” or “He’s disengaged.” Synthesising weeks of complex workplace interactions into a single adjective is easy, but it strips away all actionable context.

Second, emotional frustration creates narrative certainty. When a colleague misses a critical deadline or disrupts a client meeting, you feel the immediate pressure. In that moment, your brain seeks an explanation, often falling into what psychologists call the fundamental attribution error, attributing other people’s actions to character flaws while attributing our own to circumstantial pressures.

Finally, we often mistake our conclusions for common sense. You assume that because you saw someone check their phone during a presentation, everyone agrees they are “disrespectful.” But “disrespectful” is an inference. The phone check is the observation. When you share the inference, the other person debates the label. When you share the observation, you establish a factual starting point.

The core discipline: Observation vs. inference

Effective feedback relies entirely on the discipline of separating facts from interpretations.

• Observation (Objective feedback): Factual, neutral, and verifiable events. This is what you explicitly heard or saw with your own senses.
• Inference (Judgemental feedback): The conclusions, labels, motives, or assumptions you assign to those facts.

The human brain is a meaning-making engine. When sensory data enters our conscious awareness, such as a colleague walking into a meeting late, an email sent without an attachment, or a quiet response during a brainstorming session, our brain immediately overlays a story onto that data. It does this automatically to save processing energy and predict threat levels.

Where leaders run into trouble is through a phenomenon called inference stacking:

1. Raw observation: Alex submitted the client report at 4:30 PM instead of 10:00 AM.
2. First inference: Alex rushed the report.
3. Second inference: Alex doesn’t care about client quality.
4. Third inference (Trait verdict): Alex is sloppy, careless, and unprofessional.

By the time the leader steps into the room, they are no longer reacting to a timeline delay at 10:00 AM. They are reacting to the verdict that Alex is “unprofessional.” When you deliver that verdict, Alex’s brain perceives an attack on their professional identity and core worth. The amygdala activates, triggering a fight-or-flight response. The conversation shifts instantly from how to fix report timelines to defending one’s character.

Conversely, when you strip away the stack and stay firmly anchored at step one, the raw observation, the brain registers a problem to be solved rather than an identity threat. You preserve emotional safety while maintaining high performance standards.

The video camera test
Before stepping into a feedback conversation, apply the Video Camera Test. Ask yourself: “Could a video camera record this?”

A camera can capture a colleague entering a room 15 minutes late, or interrupting a speaker three times during a pitch. It cannot record “laziness,” “disrespect,” or a “bad attitude.” If a camera cannot capture it, you are sharing an inference.

Translating personality labels into behavioural feedback

Moving from judgment to observation requires stripping away emotional shorthand. Words like “unprofessional,” “sloppy,” or “passive-aggressive” inflame conversations. It takes deliberate effort to translate personality critiques back into the observable behaviours that triggered them in the first place.

Here are three common feedback traps and how to reframe them effectively:

1. Labelling the person vs. describing the event

• Instead of: “You are disorganised and sloppy with your deliverables.”
• Try: “The quarterly report was submitted two days after Friday’s deadline, and three key metric sections were left blank.”
• Why it works: The first statement attacks their competence, forcing them to defend their self-image. The second points to objective gaps in a document that can be fixed immediately.

2. Guessing motives vs. stating actions

• Instead of: “You don’t care about this project or the team’s success.”
• Try: “In the last three sprint reviews, you haven’t shared updates on your assigned tasks or raised any potential risks.”
• Why it works: You cannot read minds. Attempting to diagnose someone’s level of “care” invites an argument about intentions. Stating their participation history focuses on what you can see.

3. Using abstract descriptors vs. concrete examples

• Instead of: “You were aggressive and dismissive in that meeting.”
• Try: “During the presentation, you raised your voice and spoke over Sarah three times while she was sharing her slides.”
• Why it works: “Aggressive” is a subjective label that people will fight. Raising your voice and interrupting three times are specific behaviours you can acknowledge and adjust.

How to address “attitude” without critiquing personality

One of the most frequent questions leaders ask in coaching is: “How do I give constructive feedback about a bad attitude?”

“Attitude” is an internal state. You cannot observe it directly, nor can you measure it. However, an internal state always manifests in observable actions, such as tone of voice, body language, missed commitments, or dismissive comments.

To address an attitude issue without derailing the conversation, identify the specific actions that led you to your conclusion, and address those actions directly.

In practical terms: The dialogue

Leader: “When the project scope change was announced in today’s standup, you rolled your eyes, folded your arms, and stopped taking notes. When I asked for your input, you said, ‘Do whatever you want.’ That response leaves the team unclear on your project dependencies. What was happening for you in that moment?”

Team Member: “I wasn’t trying to be difficult. I’m just frustrated because we keep changing priorities every two weeks.”

Leader: “I appreciate that the changing scope is frustrating, and we can discuss how we manage scope changes. But in team standups, I need us to raise concerns constructively so we can map out dependencies. Can we agree on how to handle those moments going forward?”

Notice how this approach addresses the exact behaviours without ever using the word “attitude.” It separates the emotional reality (frustration over scope) from the workplace behaviour (rolling eyes, dismissive language), allowing you to solve both without attacking character.

Frameworks to anchor objectivity: A worked case scenario

To see how structured frameworks enforce behavioural feedback, let’s examine a single, realistic workplace scenario across three distinct models: AID, BOOST, and CEDAR.

1. The AID feedback model in practice

The AID Feedback Model forces you to anchor your opening statement in observable action before discussing consequences or next steps:

• Action (Observable behaviour): “Alex, yesterday’s performance report was submitted at 4:30 PM rather than the agreed 10:00 AM deadline, and the financial variance table was omitted from section three.”
• Impact (Objective result): “Because the report arrived late and incomplete, the leadership team didn’t have time to review the figures before the client call. We spent the first 20 minutes of the client session pulling up figures manually rather than leading the strategy discussion.”
• Development (Clear future expectation): “Going forward, I need final reports submitted by 10:00 AM with all required financial tables completed. If you run into data delays before deadline day, notify me by 9:00 AM so we can adjust.”

2. Stress-testing with BOOST

While AID helps deliver the message, the BOOST Feedback Model acts as a pre-conversation quality check to ensure your preparation hasn’t slipped back into personality judgements:

• Balanced / Observed: Did you observe this first-hand? Yes, you checked the email timestamp (4:30 PM) and opened the document yourself, rather than relying on second-hand office gossip.
• Objective: Have you stripped out character assumptions? Yes, you removed judgements like “Alex is unreliable” or “Alex doesn’t respect client time,” leaving only time stamps and document contents.
• Specific: Can you point to concrete facts? Yes, citing 10:00 AM vs 4:30 PM, yesterday’s meeting, and section three’s financial variance table.
• Timely: Are you addressing it promptly? Yes, holding the conversation within 24 hours of the client meeting while memories and context are fresh.

3. Navigating recurring patterns with CEDAR

If Alex’s late submission is not an isolated slip-up but part of a recurring pattern over the last two months, the CEDAR Feedback Model shifts the structure into a collaborative diagnosis:

• Context: “Alex, I want to talk about report delivery timelines and accuracy across our key client accounts over the last quarter.”
• Example: “Yesterday, the monthly performance report arrived at 4:30 PM without the financial variance table. Two weeks ago, the quarterly review deck was delivered three hours late with missing appendix data.”
• Diagnosis: “We’ve discussed deadline expectations before, but the pattern is continuing. What is happening in your workflow that is causing these delays on delivery day?”
• Actions: (Alex identifies a bottleneck waiting on raw data from the finance team). “Let’s set up an automated data pull from finance 48 hours prior, and agree that if data isn’t received by noon on Wednesday, you escalate it directly to me.”
• Review: “Let’s review how this new workflow functions after the next report cycle on Thursday the 15th.”

(For a comprehensive guide on selecting the right framework for your situation, explore my primary Feedback Models Hub.)

4 reflection questions for leaders before speaking

Before initiating any behavioural feedback conversation, pause and run through these self-coaching questions:

1. Am I describing a specific event, or passing a verdict on who they are?
2. Can I point to a concrete date, time, document, or interaction?
3. Am I using descriptive words (interrupted, late, blank) or judgment words (rude, sloppy, disengaged)?
4. What story have I told myself about their motives, and am I leaving room for context I might be missing?

Feedback is a tool for professional development, not a mechanism for emotional discharge. By stripping away judgements and focusing purely on observable behaviour, you preserve dignity, reduce defensiveness, and create the psychological safety required for high performance.

Recent articles in this leadership series

• How do I give feedback to someone I don’t manage?
• How to give difficult feedback to someone more senior
• The best feedback model for a difficult conversation