The Art of Feedback: How to Give Feedback That Helps Your Team Grow, Not Quit

20 August 2026
Imagine: a specialist misses a deadline. For the second time. You see the problem but don't want to turn the conversation into a conflict. So you decide to say nothing for now. Maybe the person will understand it on their own. Or you'll discuss it at the next performance review.

A month passes. Deadlines continue to be missed, the team picks up other people's tasks, your irritation builds up — and ultimately, instead of one specific remark, the employee hears a whole list of complaints.

To them, it sounds something like this: "I'm working as hard as I can here (and I have no energy left), and they tell me everything is wrong!".

That is why good feedback is not just a manager's soft skill. It is one of the team management tools.

And its task is not to point out a person's mistake. The task is to make it clear what exactly should be continued, what needs to be changed, and how to move forward.

Feedback ≠ criticism

The problem begins with the perception of the word "feedback" itself.

The manager thinks: "Now I will have to criticize". The employee hears: "Now they will tell me that something is wrong with me".

As a result, one postpones the conversation, and the other prepares to defend themselves before it even starts.

But high-quality feedback works differently.

Criticism evaluates. Feedback provides information with which a person can do something.

Compare:

"You are constantly inattentive".

What to do with such information? Unclear. It is a characteristic of a person, not a situation.

"There were data errors left in the last two reports. Because of this, the team spent extra time double-checking. Let's add a final check for key metrics before sending."

Here we have a specific action, its consequence, and a clear next step.

The main rule: talk about the action, not the person

"You are irresponsible" — a label.

"We agreed to have the result on Friday, but the task was ready on Monday, and the team didn't know about the delay" — a fact.

This difference seems small, but it determines whether a person will start analyzing the situation or defending themselves.

Therefore, before a difficult conversation, ask yourself a simple question:
Can I currently name a specific behavior that the person is able to change?
If not — your feedback is probably not specific enough yet.

Not all feedback has to be about mistakes

Another management trap is talking to a person only when something goes wrong.

In such a team, an invitation like "let's talk for 15 minutes" quickly starts to sound like bad news.

In reality, feedback is needed in at least three situations:

1. Supportive: "Keep it up"

Did an employee handle a difficult client meeting well? Took responsibility in a critical situation? Helped a colleague resolve a problem?

Don't limit yourself to "good job." Say exactly what worked:

"You structured the arguments very clearly at the meeting and didn't take a defensive stance when the client started objecting. This helped bring the conversation back to a resolution. Such an approach is definitely worth maintaining."

The person understands not just that they were praised, but which behavior is worth repeating.

2. Developmental: "You are ready for more"

You don't have to wait for a mistake to talk about development. Sometimes good feedback sounds like this:

"You're already consistently handling your part of the work. As a next step, I see you taking more responsibility for the entire process. What do you think about managing the next project on your own?"

This is especially important for strong specialists. If a person has long outgrown their role but the manager doesn't notice it, sooner or later they will start looking for growth elsewhere.

3. Corrective: "Here is what needs to be changed"

This is the feedback that is most often avoided. But in vain.

Corrective feedback should not punish for a mistake. It should help to understand:

what happened → why it is a problem → what we do differently next time

Then the mistake stops being a reason to find someone to blame and becomes material for improving the work.

When to give feedback: not in three months, but not at the peak of anger either

Imagine that at a performance review a manager says: "By the way, back in April there was a situation that I didn't like..."

For the employee, this is almost useless information. They might not remember the context anymore. They can't go back to that situation. And most importantly, for three months they continued to work the same way because no one told them that something needed to be changed.

That is why the "hot oven" rule works: the reaction should be close to the event itself. Optimally — on the same day or within the next few days.

But there is an exception.
If you just got out of a problematic release, the client is furious, and you yourself are ready to say much more than you should — don't start the conversation right away. Feedback given to vent your own irritation is no longer a development tool. Let the emotions subside. Return to the facts. And only then speak.

Publicly or one-on-one?

The basic rule is simple: reinforcing the good can be done publicly, correcting — is better done privately.

Public recognition can reinforce the desired behavior not only of a specific specialist but of the whole team.

Public criticism does the opposite: the person's attention switches from "what can I fix?" to "how do I save face in front of my colleagues right now?".

And one more nuance that managers often miss: check if the person is even ready to hear you right now. If the employee has just received tough news, is under severe stress, or is obviously emotionally exhausted, sometimes the best management decision is to reschedule the conversation.

6 feedback techniques: which one to use and when

There is no universal formula for good feedback. A conversation with a junior specialist after their first mistake and a conversation with a senior who systematically violates agreements require a different approach. Similarly, you shouldn't give feedback the same way to a person who doesn't understand the consequences of their actions and to one who already sees the problem and is looking for solutions.

Therefore, a strong manager doesn't try to use one favorite model in every situation. They have several tools — and choose the right one depending on the context.

1. Start / Stop / Continue — when specifics are needed

One of the simplest models, especially convenient for regular 1-on-1s, retrospectives, or working with mature specialists. Together you determine:

  • Stop — what should be stopped;
  • Continue — what already works and needs to be kept;
  • Start — what new behavior or practice should be added.
For example:
Stop: reporting the risk of missing a deadline on the due date.
Continue: independently looking for solutions before escalating the problem.
Start: warning the team as soon as you see that deadlines might shift.

When to use: when the person understands their role well enough and needs clear guidelines without a lengthy analysis of the situation.

2. SBI: Situation → Behavior → Impact — when subjectivity needs to be removed

An especially useful model for difficult corrective feedback. Instead of "You communicate poorly with the team", we break the situation into three parts:

  • Situation (where and when did it happen?): "At yesterday's planning..."
  • Behavior (what exactly did the person do?): "...you interrupted your colleagues several times when they were explaining their estimates..."
  • Impact (what effect did it have?): "...as a result, part of the team stopped joining the discussion, and we missed several important risks."

There is no "you are a bad communicator" label here. There is behavior that can be observed and changed.

When to use: when the conversation can potentially cause a defensive reaction and it is especially important to rely on facts.

3. Action → Impact → Change — when it is important to agree on what to do next

This model is similar to SBI but has a fundamental difference: it does not end with explaining the consequence but immediately shifts the conversation to the future.

For example:
"Before the release, you decided to skip part of the testing. Because of this, an error got into production, and the team spent the evening on an urgent fix. Let's agree that before the next release we go through the full checklist, even if time is tight."

The person receives not just an answer to "what was wrong", but an understanding of "what to do differently now".

When to use: when the problem is clear, but you need to change a specific behavior or process.

4. COIN: Context → Observation → Impact → Next steps — when the situation is more complex than a single mistake

COIN helps to conduct a more structured conversation, especially if the problem is recurrent:

  • Context — in what context did the situation arise?
  • Observation — what did you actually observe?
  • Impact — what consequences did it lead to?
  • Next steps — what do we do next?
For example:
"Over the last two sprints, we changed priorities several times after the work had already started. I noticed that you took on new tasks without syncing with the team. Because of this, part of the planned work was carried over to the next sprint. Let's determine how we will align on such changes going forward."

Unlike a brief remark, COIN allows you to show a pattern, its consequences, and move on to agreements.

When to use: when you need to analyze recurrent behavior or a situation with a broader context.

5. Feedforward — when the past cannot be changed

Sometimes a detailed analysis of a mistake provides less benefit than a conversation about the next attempt.

Then instead of "Why was the presentation overloaded with details again?" you can ask: "What can we change in the next presentation so that management sees three key conclusions within the first five minutes?"

Feedforward shifts the focus from "what you did wrong" to "how to get a better result next time". This doesn't mean ignoring responsibility for mistakes. But if the cause is already clear, there is no point in endlessly returning to the past.

When to use: for development, skill building, and situations where it is more important to form new behavior than to analyze the old one again.

6. Sandwich: Praise → Correct → Praise — when the remark is really minor

The classic scheme: positive → correction → positive.

For example:
"The report's structure has become much clearer. I would just move the key numbers to the beginning so the client doesn't have to look for them in the text. Overall, the direction is very good — especially the conclusions block."

For a small tweak, this might work. But it's better not to wrap a serious problem in two compliments. If a manager starts a difficult feedback session every time with "you are doing a great job", employees quickly learn to wait: "Okay, so where's the 'but'?". As a result, even sincere praise begins to be perceived as preparation for criticism.

When to use: for small specific adjustments, when the positive part of the feedback is truly sincere and relevant.

Table for those who like specifics

Situation

Recommended technique

Need short and clear guidelines?

Start / Stop / Continue

Need to discuss a complex situation as objectively as possible?

SBI (Situation → Behavior → Impact)

Need not only to show the problem but also to change a specific behavior?

Action → Impact → Change

The problem recurs and requires a broader analysis?

COIN (Context → Obs. → Impact → Next steps)

The person has already understood the mistake and it is more important to prepare for the next situation?

Feedforward

Need a small, specific tweak?

Sandwich (Praise → Correct → Praise)

But the model itself will not make a difficult conversation good. It only provides structure.

The tone, context, trust between you, maturity of the specialist, and your readiness not only to speak but also to listen — all this will determine the result no less than the correctly chosen framework. A manager's task is not to learn six beautiful abbreviations, but to understand when which of them will help a person truly hear your feedback and do something about it.

What most often ruins even the right feedback

You can know all the models and still have a bad conversation. Most often, the reason is one of these mistakes:

  • Accumulating complaints. You shouldn't wait until five small problems turn into one big conversation.
  • Speaking in generalizations. "You always", "you never", "with you constantly" almost guarantee to cause a defensive reaction.
  • Attributing motives. You see the person's behavior, but you don't always know why they did it.
  • Turning a dialogue into a monologue. Good feedback implies an answer. Ask: "How do you see this situation?"
  • Ending the conversation on the problem. If after a 1-on-1 a person perfectly understands what they did wrong but doesn't understand what to do now, the job is not yet finished.

A simple test: was your feedback helpful?

After the conversation, the employee should have answers to four questions:

  1. What exactly happened?
  2. What impact did it have?
  3. What is expected of me in the future?
  4. What specifically can I do differently?

If instead, the person leaves with the thought "the boss is unhappy with me", the feedback did not fulfill its function.

A strong team is not one where no one makes mistakes

It is a team where mistakes can be discussed before they turn into a systemic problem.

Where strong work is noticed, and not taken for granted.

Where a manager can say "this needs to be changed" without humiliating the person. And the employee can reply "I see the situation differently" without fear of consequences.

Such a culture gradually creates an environment where people don't guess their manager's expectations, but understand them.

And this is no longer just about comfortable communication. This is about the quality of work, the speed of specialists' development, and the maturity of the team.

And how to give feedback when simple models are no longer enough?

Knowing Start / Stop / Continue is not difficult. It is much harder to apply the right approach when you have a real person in front of you, a real conflict, and no ready-made script.

How to talk to a strong specialist who systematically violates agreements? How to give feedback to a person who immediately goes into a defensive position? How to distinguish a problem in an employee's work from a problem in your own management? How to conduct a difficult 1-on-1 so that a solution emerges after it, rather than a new conflict?

This is exactly where management frameworks cease to be a theory and the real work of a leader begins.

In management programs and mentoring sessions at SkillsUp, we work with exactly such situations: we analyze real management cases, difficult conversations, and decisions that managers face in their teams.

Because a strong leader is not someone who never has difficult conversations. It is someone who knows how to conduct them in a way that the team becomes stronger after the conversation.