Manager guide
A practical, research-grounded guide to conversations that change behavior
The goal of a difficult feedback conversation is simple to say and hard to do: the person should leave knowing what needs to change, why it matters, and how you will support them. You can be clear about expectations and respectful of the person at the same time. In fact, the research suggests you have to be both.
This guide walks through the steps, from preparing to following up. Each step pairs practical actions and sample language with a “Why this works” note that explains the behavioral science behind the recommendation.
This guide walks through
What this guide covers
Manager-to-employee conversations about a shortfall in work behavior or performance that may be uncomfortable to deliver or receive. Keep in mind that corrective content is not the same thing as a harsh delivery. You can deliver a clear, serious message in a way that keeps the person focused on improving.
Serious misconduct, safety concerns, formal disciplinary steps, and accommodation questions sit outside this guide and need your organization's established processes. There is more on this at the end.
The big idea: feedback can help or hurt
Most managers assume that giving candid feedback is better than saying nothing. That makes sense. We expect leaders to speak up, and staying quiet can feel like avoiding the issue. But the research gives us a reason to be thoughtful about how we do it. Feedback improves performance on average, yet a meaningful share of feedback interventions actually make performance worse. The same information can help someone improve or leave them feeling defensive, discouraged, or unsure what to do next.
A lot depends on where the feedback directs the person’s attention. When we focus on the work, including what happened, what’s expected, and what to try next, we help someone think through how to improve. When feedback feels like a judgment about their competence or character, their attention can shift toward defending themselves or questioning their ability. They may spend more time replaying the conversation than figuring out how to do the work differently.
Consider two ways to deliver the same feedback. “Your updates keep missing risks. You need to be more careful” can leave someone wondering, “Do you think I’m careless?” Compare that with: “The last two updates didn’t include the delivery risks. Let’s talk about how to identify and flag those before the deadline.” Now there’s a clear problem to work through and a next step to discuss.
That’s why this guide emphasizes observable behavior, clear expectations, and practical next steps. These help people understand what needs to change and how to move forward, even when the message is difficult to hear.
Why this matters
Kluger and DeNisi's (1996) meta-analysis showed that feedback improved performance on average, yet more than one-third of the effects were negative, meaning performance declined after feedback.
Their Feedback Intervention Theory offers an explanation: feedback shifts attention. When attention goes to the task and to strategies for doing it better, performance tends to improve. When attention is pulled toward the self ("What does this say about me? Am I failing?"), it competes with the effort needed to improve.
The practical takeaway: every step below is designed to keep attention on observable behavior, changeable actions, and the work itself, and away from judgments about the person's worth. And because single conversations sit inside an ongoing relationship, the more regularly you give useful feedback, the less loaded the hard conversations become.
Part 1
Prepare
Most of what makes a hard conversation go well happens before it starts.
Step 1. Get clear before you start
Choose the most important behavior to address. Trying to cover five issues at once dilutes the message and increases the odds the person feels attacked. Bring one or two recent examples, the relevant expectation, and the impact you can actually support with evidence. A single incident can warrant feedback; describe it as a pattern only when you have evidence of a pattern.
Check your assumptions. Before the conversation, ask yourself:
- What did I directly observe, and what am I inferring?
- Was the expectation clear to begin with?
- Would I apply the same standard to someone else on the team?
- Have priorities, workload, tools, or my own instructions contributed?
- What evidence would challenge my first impression?
Describe what the person did. Labels like “negative,” “careless,” or “not committed” leave people guessing about what needs to change. Name the specific action and how it affected the work. Be thoughtful about interpreting eye contact, emotional expression, or personality. Focus on behavior that matters for the job and explain why it matters.
| Instead of a label | Describe the behavior |
|---|---|
| "You're careless with project updates." | "In the last two project updates, the delivery risks were missing." |
| "You don't seem committed." | "The last three deadlines were moved without a heads-up to the team." |
| "You're too negative in meetings." | "In Tuesday's planning meeting, the new timeline was dismissed before we discussed options." |
"In the last two project updates, the delivery risks were missing. The expectation is to flag risks before the deadline, so we have time to adjust."
Why this works
Labels can put people on the defensive. Feedback Intervention Theory (Kluger & DeNisi, 1996) helps explain why. Calling someone “careless” can shift their attention toward defending themselves. Describing what happened helps them focus on something they can change.
Check the story you’re telling yourself. It’s easy to explain someone’s mistake by their attitude or character and overlook the circumstances. Before giving feedback, ask: “What might I be missing? What evidence would challenge my interpretation?” Consider whether unclear expectations, competing priorities, or limited resources contributed.
Step 2. Choose a setting where they can engage
Speak privately, reasonably soon after the event, and at a time when you can both give the conversation your attention. If you are frustrated, give yourself time to settle first. Share the topic in advance when practical, and avoid an unexplained "we need to talk," which tends to generate anxiety without information.
"I'd like to talk about how we flag project risks and agree on what will work better for the next update."
Begin directly. You don't need a long preamble or praise wrapped around the concern (the so-called feedback sandwich). Wrapping criticism in praise can bury the message and teach people to distrust your compliments. Genuine appreciation can stand on its own, at another time.
"I want to address something that is affecting the team's ability to plan. I'll share what I've noticed, and I want to understand your perspective before we decide the next steps."
Why this works
People learn more when they feel safe to speak openly. Psychological safety helps people ask for help, acknowledge mistakes, and share their perspective. How you start the conversation can make it easier for someone to engage or leave them worried about what will happen if they speak up (Edmondson, 1999).
A safe conversation can still be uncomfortable. Psychological safety means people can ask questions, disagree, or admit a mistake without fear of being embarrassed or punished for speaking up. You can hold someone accountable and give them room to speak candidly (Edmondson, 1999).
Your emotional state affects how feedback lands. Baron's (1988) findings on destructive criticism are a reason to wait until you are calm. Frustration makes it more likely that your feedback becomes general, harsh, or attributed to the person's character, exactly the features that trigger resistance.
Part 2
Make it a two-way conversation
Deliver the message clearly, then listen as if you might be missing something. You often are.
Step 3. Describe the behavior, impact, and expectation
Use a simple sequence: "Here is what happened. Here is the impact. Here is what is expected." Keep the impact proportional to what actually occurred. Separate what happened from what you worry could happen, and distinguish concerns others have reported from your own observations.
"The last two updates did not include risks you already knew about. We learned about them on the due dates and had to change the handoffs that day. I need known risks surfaced early enough for us to respond."
Then stop talking. Give them space to respond. Silence can feel awkward, but filling it tends to soften or muddy the message.
Avoid: "always" and "never," comparisons with coworkers, and guesses about intent such as "you don't care."
Why this works
People are more likely to take ownership when they understand why. Explaining how a behavior affects the team, customer, or work gives someone a reason to change beyond simply being told to. That helps them see the value in the request and make it their own (Gagné & Deci, 2005).
Feedback is more useful when it offers a way forward. Negative feedback can reduce motivation, but how you deliver it matters. Explain what needs to improve and offer guidance on how to get there. That gives the person something to work toward (Fong et al., 2019).
Absolutes and comparisons shift attention to the self. "You always" is rarely accurate and invites a debate about exceptions. Comparisons with coworkers make the conversation about social standing rather than the task, which is the kind of self-focused attention Kluger and DeNisi (1996) associate with worse outcomes.
Step 4. Ask what you are missing
"Walk me through what was happening. What made it difficult to raise those risks earlier?"
Listen for information that could change your understanding. Reflect it back to check that you heard it correctly:
"You were waiting until you had a solution because you thought that was the expectation. Have I understood?"
Correct your account if new evidence warrants it. If facts are disputed, name what you will verify and when you will reconnect.
Why this works: voice
Having a chance to be heard matters. People tend to respond more positively to performance conversations when they can share their perspective, even if it doesn’t change the outcome. Being heard can matter as much as having a say in the decision (Cawley, Keeping, & Levy, 1998).
Make room for their perspective and listen with an open mind. You may learn something that changes your understanding or the plan. Even when the expectation stays the same, listening helps people feel the conversation was fair.
Diagnose the barrier before you prescribe the fix
A performance gap is not automatically a motivation gap. The COM-B model (Michie et al., 2011) proposes that behavior results from the interaction of Capability, Opportunity, and Motivation. Use these three questions to match your support to what's actually getting in the way:
| Barrier | Ask yourself | What helps | Try asking |
|---|---|---|---|
| Capability | Can they do it? Is the expectation or method clear? | Explain it, show an example, practice together | "What would an early risk update look like?" |
| Opportunity | Does the work environment allow it? Are priorities, workload, access, or team norms in the way? | Change the conditions: tools, time, access, priorities | "What would we need to adjust to make this realistic?" |
| Motivation | What makes it hard to follow through? Confidence, habits, perceived consequences, competing incentives? | Discuss meaning, expectations, habits, and concerns | "What concerns you about raising a risk before you have a solution?" |
Why this works: COM-B
When managers assume every shortfall is about effort, they default to pressure, which does nothing for someone who lacks the know-how or the conditions to succeed. COM-B structures a broader diagnosis so the intervention fits the cause: instruction and practice for capability gaps, environmental changes for opportunity gaps, and conversation about meaning and concerns for motivation gaps.
Step 5. Acknowledge emotion without losing the message
Give people room to react while keeping the conversation moving. An emotional reaction is not proof that someone is unwilling to improve. They may feel surprised, embarrassed, frustrated, or worried about what the feedback means. Give them time to process it before expecting them to work through a solution.
Allow a pause, and ask whether they need a moment. Avoid telling them how they should feel, saying “don’t take it personally,” or requiring immediate agreement. You can acknowledge their reaction while keeping the expectation clear.
"I can see this is difficult to hear. Would a short pause help? We still need to work through what changes from here."
If you contributed, say so plainly:
"I was not clear about when to escalate. That is mine to fix. Let's make the expectation explicit now."
Understanding the context helps you choose a better response. It does not automatically remove accountability.
Why this works
People engage more when they have a voice, feel capable, and feel supported. Self-determination theory describes these needs as autonomy, competence, and relatedness. Listen to their perspective, explain why the expectation matters, and offer choices about how to meet it where you can. These practices help people take ownership of what’s being asked (Gagné & Deci, 2005).
Acknowledge how they’re feeling and own any part you played in the situation. That helps build trust and keeps the conversation focused on improving the work. Keep the expectation clear, and pair it with support that helps them see how they can succeed.
Part 3
Turn the conversation into action
Understanding what needs to change is not the same as changing it. Close the intention-behavior gap.
Step 6. Agree on a small, observable change
Choose one or two actions with a clear standard and a realistic opportunity to practice. Invite the employee to help choose the approach, while being explicit about what's required. Ask them to draft a risk message or rehearse how they would raise it in a meeting. Give feedback on the attempt, then let them try again.
"For each project update, include known risks, their likely impact, and any help needed. If a risk emerges between updates, flag it that working day, even if you do not have a solution yet. What would help you do that consistently?"
Why this works: goal setting
Clear goals give people something to work toward. “Do better” leaves too much open to interpretation. Specific goals that stretch someone’s abilities can help them focus their effort and keep going. They work best when the person is committed to the goal, has the ability and support to pursue it, and gets feedback along the way (Locke & Latham, 2002).
When the task is new or complex, include a learning goal. Asking for better results may not help if the person hasn’t figured out how to achieve them. Pair the expectation with something they can practice, such as, “Use this example to practice identifying three types of delivery risk.” That learning may need to come before a tougher performance target. Focus goals on what the employee can control, and clarify where success depends on other people or resources.
Step 7. Confirm understanding and follow through
Check understanding rather than relying on a nod or "I'll do better." Write a brief, factual recap and invite corrections. Include the behavior, the support you'll provide, what counts as evidence of progress, and the review date.
| Example agreement | |
|---|---|
| Employee | Include risks in the next three weekly updates; flag new risks the day they are identified. |
| Manager | Share an example today, clarify priorities, and help resolve escalated blockers. |
| Progress | Review whether known risks were raised promptly and with useful detail. Don't equate "no risks reported" with success. |
| Follow-up | Check in after the first update; review the next three together and adjust support as needed. |
Recognize the specific improvement when you see it. Specific recognition tells the person exactly what to repeat:
"You raised that dependency on Tuesday, which gave us time to adjust the handoff. Keep using that approach."
If progress stalls, revisit the barrier, your support, and the expectation, rather than repeating the same feedback more forcefully. If expectations remain unmet, be clear about next steps and use your organization's performance process. If this is already a formal performance conversation, say so. Don't present it as informal coaching.
Why this works
Follow-up helps people turn a goal into progress. People are more likely to reach their goals when they track how they’re doing, especially when they write it down or share it with someone. A brief written recap and planned check-ins make that easier (Harkin et al., 2016).
Choosing a meaningful progress indicator matters. "No risks reported" could mean the behavior improved, or that risks are still being hidden. Measure and recognize the behavior you actually want to see. Specific recognition closes the feedback loop. It reinforces the new behavior, supports the person's sense of competence (Gagné & Deci, 2005), and shows that your feedback isn't only about problems.
Language for difficult moments
Even well-prepared conversations take unexpected turns. These responses keep the conversation constructive without dropping the message.
| When they... | Try saying | Why it works |
|---|---|---|
| Disagree | "We may see the situation differently. Let's separate what we can verify from how we interpreted it, then clarify the expectation going forward." | Gives voice while separating facts from interpretation (Cawley et al.) |
| Point to someone else | "That dependency matters, and I'll follow up on it. Let's also identify what was within your control and what you could flag earlier." | Takes opportunity barriers seriously while keeping focus on controllable actions (COM-B) |
| Go quiet | "You do not have to respond immediately. What questions do you have? We can reconnect tomorrow after you have had time to think." | Silence isn't agreement; time to process supports genuine voice |
| Quickly promise to improve | "Thank you. Let's make that concrete. What will you do differently the next time this happens?" | Converts a vague intention into a specific plan (Gollwitzer & Sheeran) |
| Become upset | "We can take a pause. Would you like a few minutes, or should we agree on a time to continue?" | Offers choice and acknowledges emotion (SDT) without abandoning the topic |
| Say it's the first they've heard of it | "Thank you for saying that. I should have raised it earlier. Let's clarify the expectation now and agree on a fair opportunity to practice." | Owns your part and restores fairness before holding them to the standard |
When coaching isn't the right tool
Use coaching for development and everyday work behavior. Serious misconduct or safety concerns need your organization's appropriate process. If a need for an accommodation emerges during the conversation, involve the appropriate support (such as HR) rather than trying to diagnose the person yourself.
Your 60-second preparation check
- Can I name the behavior without a label?
- Is the standard clear and applied consistently across people?
- What context might I be missing?
- Which barrier is most likely: capability, opportunity, or motivation?
- What should happen next, specifically?
- What support will I provide?
- When will we review progress, and how will we know it's working?
A note on how to read this guide
These practices translate established research into manager actions. The sequence and sample language are practical applications of that research, not a single validated coaching protocol. Several of the most relevant studies examine adjacent questions (goal pursuit, training, team learning, appraisal reactions) rather than difficult feedback conversations directly. That is typical of applied behavioral science, and it is a good reason to adapt the guide to your context, pay attention to what actually changes afterward, and adjust.
References
Anseel, F., & Sherf, E. N. (2025). A 25-year review of research on feedback in organizations: From simple rules to complex realities. Annual Review of Organizational Psychology and Organizational Behavior, 12, 19–43. https://doi.org/10.1146/annurev-orgpsych-110622-031927
Baron, R. A. (1988). Negative effects of destructive criticism: Impact on conflict, self-efficacy, and task performance. Journal of Applied Psychology, 73(2), 199–207. https://doi.org/10.1037/0021-9010.73.2.199
Cawley, B. D., Keeping, L. M., & Levy, P. E. (1998). Participation in the performance appraisal process and employee reactions: A meta-analytic review of field investigations. Journal of Applied Psychology, 83, 615–633. https://doi.org/10.1037/0021-9010.83.4.615
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44, 350–383. https://doi.org/10.2307/2666999
Fong, C. J., Patall, E. A., Vasquez, A. C., & Stautberg, S. (2019). A meta-analysis of negative feedback on intrinsic motivation. Educational Psychology Review, 31, 121–162. https://doi.org/10.1007/s10648-018-9446-6
Gagné, M., & Deci, E. L. (2005). Self-determination theory and work motivation. Journal of Organizational Behavior, 26, 331–362. https://doi.org/10.1002/job.322
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69–119. https://doi.org/10.1016/S0065-2601(06)38002-1
Harkin, B., Webb, T. L., Chang, B. P. I., Prestwich, A., Conner, M., Kellar, I., Benn, Y., & Sheeran, P. (2016). Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin, 142, 198–229. https://doi.org/10.1037/bul0000025
Katz, I. M., Rauvola, R. S., & Rudolph, C. W. (2021). Feedback environment: A meta-analysis. International Journal of Selection and Assessment, 29(3–4), 305–325. https://doi.org/10.1111/ijsa.12350
Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological Bulletin, 119, 254–284. https://doi.org/10.1037/0033-2909.119.2.254
Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57, 705–717. https://doi.org/10.1037/0003-066X.57.9.705
Michie, S., van Stralen, M. M., & West, R. (2011). The behaviour change wheel: A new method for characterising and designing behaviour change interventions. Implementation Science, 6, 42. https://doi.org/10.1186/1748-5908-6-42
Roch, S. G., Woehr, D. J., Mishra, V., & Kieszczynska, U. (2012). Rater training revisited: An updated meta-analytic review of frame-of-reference training. Journal of Occupational and Organizational Psychology, 85, 370–395. https://doi.org/10.1111/j.2044-8325.2011.02045.x
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article