Peer feedback only works when it is specific, fair, and tied to real work. In performance management, the goal is not to sound polished; it is to help a colleague improve outcomes, remove friction, and keep expectations clear. When giving feedback to peers, I think the best conversations are short, evidence-based, and useful enough that the other person can act on them immediately.
What matters most when feedback has to improve performance
- Keep it about observable work, not personality, motives, or assumptions.
- Use one clear example, explain the impact, and name the next step.
- Choose the right moment; private, recent, and calm usually works better than public or delayed.
- Match the format to the issue; a quick course correction needs a different approach from a repeated performance problem.
- Make it regular, because surprise feedback tends to create defensiveness instead of improvement.
- Keep the UK context in mind; performance conversations should support development as well as accountability.
What peer feedback should actually do in performance management
I do not treat feedback as a mini verdict on someone’s ability. In a healthy performance management process, it does three things: it clarifies expectations, it corrects course early, and it helps people get better at the job they already have. That is very close to how ACAS frames performance reviews in the UK: discussion should cover strengths, areas for improvement, support or training, and development objectives.
The other point I keep in mind is cadence. ACAS recommends regular performance reviews for employees, ideally at least once a year, but also informal feedback in between. I have found that the in-between conversations matter more than the formal one, because they prevent the “why am I hearing this now?” moment that usually kills trust.
CIPD’s UK research points in the same direction. It found that just over a quarter of employees experienced systematic performance management, and those people were more likely to rate their performance positively. To me, that says the structure matters, but so does the quality of the feedback inside it.
| Useful outcome | Weak outcome |
|---|---|
| Clearer expectations for the next piece of work | More confusion about what good looks like |
| A practical improvement the person can try straight away | A vague sense that something was “off” |
| Better support, coaching, or training if there is a skill gap | Shame without help |
| Stronger performance over time | Defensiveness, silence, or resentment |
That is why I think peer feedback belongs inside performance management, not outside it. The point is not to be harsher; it is to make work easier to improve. Next, I will show the simplest structure I use when I need to be direct without sounding blunt.

A simple framework that keeps the message specific
My default structure is simple: situation, behaviour, impact, next step. It keeps the conversation grounded in facts, which matters because vague criticism is easy to ignore and hard to act on. The more concrete the feedback, the less likely it is to feel like a personal attack.
| Step | What to include | Example language |
|---|---|---|
| Situation | Name the exact moment or piece of work | “In yesterday’s client update...” |
| Behaviour | Describe what you saw or heard, not your interpretation | “the figures on slide 4 changed after the meeting had started” |
| Impact | Explain what that changed for the team, client, or project | “that made the team pause and reduced confidence in the numbers” |
| Next step | Suggest a practical change or ask for one | “next time, can we lock the final version earlier or flag late changes before the call?” |
I like this structure because it avoids two common traps. First, it prevents the conversation from drifting into personality language such as “careless” or “difficult.” Second, it forces me to say what should happen next, which is the part many people forget. Feedback without a next step is just commentary.
If you want to make this even more useful, keep it to one issue at a time. People can absorb a short, well-chosen correction. They usually tune out when they feel a whole backlog of complaints is being dumped on them in one go.
Words that help and words that usually backfire
The wording does most of the heavy lifting. I have seen the same message land well or badly depending on whether it sounded factual or accusatory. When I need to give someone useful feedback, I focus on language that describes work, impact, and options.
| Better wording | Why it works | Avoid this |
|---|---|---|
| “The handover note was missing two owners, so the team had to chase details later.” | Specific, evidence-based, and tied to impact | “You weren’t organised.” |
| “The update arrived after the deadline, which meant the client saw an incomplete version.” | Shows the consequence without guessing intent | “You are unreliable.” |
| “The numbers on slide 5 do not match the spreadsheet.” | Clear enough for the person to verify and fix | “This looks sloppy.” |
| “What would help you get the draft finished on time next week?” | Invites problem-solving instead of defensiveness | “Why can’t you manage your time?” |
There are also a few words I try to strip out of my own first draft. “Always,” “never,” and “everyone says” usually make the conversation worse because they sound exaggerated and impossible to prove. Sarcasm does the same thing. Even when the criticism is valid, a sarcastic tone tells the other person that the goal is embarrassment, not improvement.
For delicate issues, I prefer private conversation over email or chat. Text is fine for a small correction, but it is a poor channel for anything that could be read as criticism of competence, judgement, or attitude. If the subject matters enough to affect performance, it usually deserves a real conversation.
Choosing the right format for the moment
Not every performance conversation needs the same shape. Sometimes I need to correct a single event. Sometimes I need to explore a pattern. Sometimes the person wants ideas for what to do next. Picking the format well is often the difference between a conversation that helps and one that sounds irritatingly generic.
| Format | Best for | Main risk if misused |
|---|---|---|
| SBI, or situation-behaviour-impact | Quick, factual correction tied to one event | Can feel sharp if you forget tone or context |
| Start-stop-continue | Team retrospectives or broader development conversations | Can become too broad for a sensitive issue |
| Feedforward | When the person is already open to improvement ideas | Can skip accountability if you only talk about future ideas |
| Question-led check-in | When you are not sure whether the issue is skill, process, or workload | Can be too indirect if the problem is repeated |
My rule is simple. If the issue is isolated and clear, I use SBI. If the team needs to reflect together, I use start-stop-continue. If the person wants practical coaching, I lean toward feedforward. If I am still trying to understand what is going wrong, I ask questions before I make a judgement. That keeps me from overreacting to a symptom when I should be fixing a process.
This also matters in performance management because format shapes how fair the conversation feels. A person is more likely to accept hard feedback when the structure is clear and the intent is obviously constructive. That is one of the reasons regular check-ins work better than surprise feedback dropped in late.
Handling awkward situations without turning feedback into conflict
The hardest peer conversations are rarely about the content alone. They are about status, history, and trust. If the peer is senior to you, close to you, or simply hard to read, the feedback can become emotionally loaded very quickly. I find the safest approach is to stay factual, keep the language calm, and avoid turning the exchange into a debate about character.
- If the colleague is senior to you, stick to the facts and the shared goal. I would not apologise for being specific.
- If the colleague is a friend, separate the working issue from the relationship issue. “I want to be straight with you about the project” is better than hinting around it.
- If the problem has repeated, bring two or three examples, not a long list. A small set of recent examples is usually enough.
- If the issue is serious or affects conduct rather than day-to-day performance, escalate it instead of pretending a peer chat is enough.
- If the colleague is neurodivergent or may need support, adjust the delivery and check what would help before assuming resistance.
One more practical point: if you know a conversation could sting, do not spring it on someone in a group meeting. A private setting lowers the temperature and gives the other person room to respond honestly. Face to face or video usually works better than written messages for anything nuanced.
Make peer feedback part of the rhythm, not a surprise
The cleanest performance conversations are the ones people saw coming. I would rather give a colleague two short check-ins during a project than one heavy conversation after the deadline has already been missed. Regular feedback reduces surprise, and surprise is what usually turns a useful correction into defensiveness.
Here is the habit set I return to:
- Set expectations early, especially on deadlines, ownership, and quality.
- Share feedback while the work is still fresh enough to change.
- Keep a brief written record of what was discussed and what the next step is.
- Separate support from judgement; if a skill gap is the issue, name the support that would help.
- Close the loop in the next check-in so the conversation does not disappear into memory.
That rhythm also makes formal performance management less awkward. When informal feedback is already happening, reviews feel like an extension of the same conversation rather than a surprise audit. I think that is the healthiest version of performance management: regular, specific, and focused on improvement rather than blame.
The habit that makes the next conversation easier
If I had to reduce the whole approach to one habit, it would be this: before speaking, I write down the exact behaviour, the visible impact, and the change I want to see. If I cannot do that clearly, I am probably not ready to give the feedback yet. That one pause saves a lot of friction.
- Name the work, not the person.
- Use one recent example.
- State the impact plainly.
- Ask for, or suggest, a next step.
- Follow up once, rather than hoping the message sticks on its own.
That is the standard I would use in any UK performance conversation, whether I am supporting a teammate, coaching a peer, or helping a group improve how it works together. The best feedback is direct enough to change behaviour, respectful enough to preserve trust, and regular enough that nobody is blindsided by it later.
