1. LandOffer
  2. Blog
  3. All articles
  4. Give a Coworker Difficult Feedback Using Observations, Impact, and a Clear Request

Workplace Communication

Give a Coworker Difficult Feedback Using Observations, Impact, and a Clear Request

Two fictional peer conversations turn a design-file handoff problem and an unclear chart objection into specific mutual actions, with bounded follow-up messages.

10 min readLandOffer team

Give a coworker difficult feedback by naming one specific situation, describing what you observed, explaining its effect on your work, and asking for their perspective. Then propose a change you can both understand and act on. Keep the request within your peer relationship. Ask for a clearer handoff or a different collaboration practice without diagnosing the person's character or assigning a performance rating.

This guide from LandOffer.ai develops two fictional cross-functional conversations and their follow-up messages. The people, files, observations, and replies are invented teaching inputs. No coworker was contacted or workplace record accessed. The examples address ordinary collaboration, rather than formal personnel decisions or complaint procedures.

Two coworkers compare their views of one shared work artifact, keeping an observation separate from an assumed motive.

Choose a work event you can describe accurately

Start with the event that makes a change necessary. “You keep making my work harder” is a conclusion spanning an undefined period. A particular file replacement, review comment, or handoff gives the other person something they can recognize and explain. If you cannot identify what happened, first clarify the record rather than constructing a confident accusation.

Separate three things in your notes: what you saw, the effect you experienced, and the explanation you are currently assuming. You might have opened an older attachment and repeated a comparison. That does not prove the designer intended to exclude you. The same work problem can arise from an unclear notification practice, a misunderstood agreement, or a mistake.

The Center for Creative Leadership's Situation–Behavior–Impact guidance, published August 19, 2025, describes identifying a specific situation, observable behavior, and its impact, then asking about intent. The framework supplies a way to structure a conversation; it does not establish what caused your colleague's action or guarantee a cooperative response.

Decide whether you are ready to hear an answer that changes your understanding. A question about intent is useful only if you leave room for that answer. If the evidence is uncertain, say which part you need to clarify. You can still explain a real impact without presenting your interpretation as a settled account of the whole event.

Make a request that a peer can answer

Ask for a specific action connected to the problem. “Be more considerate” leaves the coworker guessing. “When you replace the agreed handoff file, add a version note and mention me in the issue” describes a practice the two of you can discuss. Check whether that practice fits the actual tools, responsibilities, and pace of the work.

MIT's coaching and feedback guidance recommends descriptive observations, specific changeable behavior, open questions, and next steps. It is an employer's public guidance, primarily written for management conversations. The transferable principle here is specificity and dialogue; citing it does not give a peer managerial authority.

Choose a setting where the person can respond without being made into an example for an audience. Ask for a short private conversation and name the work event you want to discuss. If writing is the normal collaboration channel, keep the message bounded and invite a reply. Choose the event that needs a different action now, and keep unrelated frustrations out of that request.

Also identify your own part in the proposed change. If you want a replacement notice, will you acknowledge it and update the reference you are using? If you ask for precise review comments, will you supply enough context for the reviewer to locate the work? Naming both actions lets you check where the practice failed later: whether the notice was missing, the acknowledgment was pending, or a different step needed clarification.

Conversation one: a replaced design file and a stale reference

This is a fictional teaching case. Engineer Noor and product designer Lena are peers working on a desktop application's status badges. On Tuesday, Noor downloaded the badge specification attached to their handoff issue. On Wednesday, Lena replaced that attachment with a revised specification but did not add a comment or mention Noor. Noor continued comparing the implementation with Tuesday's local copy, then repeated the comparison after discovering the replacement.

The supplied facts establish that Noor repeated the comparison; the example does not quantify the time spent. They do not establish a release delay, customer harm, or Lena's motive. The team's handoff practice had not specified how attachment replacements should be announced.

Noor: Could we talk about the badge handoff? I downloaded the issue's specification on Tuesday. The attachment was replaced Wednesday, but there was no update comment, and I continued using my downloaded copy. I repeated the comparison after I found the new version. What was your understanding of how I would see the change?

Lena: I thought you were reading the attachment from the issue each time. I didn't realize you had a local reference. I replaced it to avoid leaving two files that looked current.

Noor: That explains the replacement. I had assumed you knew I was working from a download, but I hadn't said so. I still need a way to notice a changed handoff before I use the old copy. Could you add a short version comment and mention me when the agreed file changes?

Lena: Yes. I can include what changed and whether it affects the badge implementation. Could you reply when you've moved to the new version, so I know the notice reached you?

Noor: Agreed. I will name the version in my implementation note and acknowledge a replacement notice. If I cannot review it before our next checkpoint, I'll say that rather than leave you assuming I have.

Lena: I'll add the missing note for Wednesday's revision. For the next handoff change, we'll use the same issue and check whether this is enough to keep the reference clear.

The conversation changes Noor's explanation of the event without erasing the impact. Lena was trying to maintain one current attachment; Noor was trying to work from a stable reference. Neither had communicated the assumption. The agreed notice and acknowledgment address that gap, while the example makes no claim that every later handoff will succeed.

Notice the limit of the request. Noor does not prohibit design revisions or require permission before Lena changes her work. They agree on how to identify a changed shared artifact. A proposal to freeze all specifications would be a different decision involving project responsibilities, rather than a natural consequence of this feedback.

Conversation two: a review objection that cannot be located

The second fictional case concerns feedback on a review practice, not a file handoff. Analyst Sumi posts a draft chart in an internal review thread. Engineer Ellis replies, “These numbers are wrong.” Sumi cannot tell which values or comparison Ellis means. She asks about the comment privately while retaining the substantive concern for investigation.

Sumi: In this morning's chart thread, your comment said the numbers were wrong without identifying a point or expected value. I couldn't locate the discrepancy, so I haven't resolved it. Could you show me which part you were comparing?

Ellis: I meant the last point on the weekly chart. It differs from the daily table. I assumed the two were meant to agree, but I hadn't checked whether they used the same time window.

Sumi: That is specific enough for me to investigate. The windows may explain it, but I need to check the definitions before saying either output is correct. For future objections, could you name the chart point, the comparison source, and the difference you see? An uncertainty is fine if you label it.

Ellis: Yes. I can change the thread comment to say which point and table I compared, and that the time-window match is unverified. Please include the chart's window in the caption so I don't have to infer it.

Sumi: I can add the window definition after checking it. I'll reply in the same thread with the comparison result or the remaining question. I am asking for a locatable objection, not asking you to stop flagging suspected errors.

Ellis: Understood. I'll update my comment today. We can use the next chart review to check whether the added context is sufficient.

Here the technical objection remains open. A clearer conversation has not validated the chart, resolved the difference, or proved Ellis was mistaken. The next technical check still needs the chart point, the daily-table source, and the definitions of both time windows. The improvement is a defined way to investigate and communicate the concern. Both people take a next action, and neither needs to agree that the original wording was malicious.

This matters when feedback concerns the way criticism is delivered. Do not make politeness a condition for examining a potentially important defect. Address the work concern through the appropriate review, while separately explaining what would make the comment actionable. Sumi's request protects that distinction by explicitly preserving Ellis's ability to raise an error.

Two design-sheet versions use a version-note ribbon and an acknowledgment bookmark to illustrate the agreed handoff practice.

Record the agreement without rewriting the conversation as a verdict

CCL's active-listening guidance, published September 1, 2026, recommends clarifying ambiguous points and checking your understanding. Apply that to the next action as well as the explanation of the event. A short recap can prevent different memories of the next action. Record the practice you both accepted and where you will try it. Avoid claiming that the person admitted fault or that you settled a broader disagreement when that did not occur. An unanswered recap remains your proposed understanding, not confirmed agreement.

Noor's completed fictional recap reads: “For changes to the badge handoff attachment, Lena will add a version comment explaining the implementation-relevant change and mention Noor. Noor will acknowledge the notice and identify the version used in the implementation note. If a review is pending, Noor will state that. We will check the practice at the next changed handoff; this does not freeze design changes.”

Sumi's completed fictional recap is different: “Ellis will identify the chart point, comparison source, and observed difference when flagging a suspected numerical error, and label an unchecked assumption. Sumi will verify and display the chart's window definition, then report the comparison result or unresolved question in the review thread. The current numerical concern is still open.”

These are mutual work agreements, not performance records. They keep a file-version problem separate from a numerical-review problem and preserve the unfinished work in each. Use whatever record and access rules apply to your team; the examples do not require creating a separate personnel file or sharing private conversation details broadly.

Follow up on the action you actually observed

At the next relevant event, compare the agreed practice with what happened. Acknowledge a specific change without announcing that the relationship is repaired. If the practice was not used, ask what prevented it and whether the agreement needs adjustment. Repeating a character judgment adds little information about the failed step.

For Noor, a bounded fictional follow-up is: “I saw your version note on today's badge revision and acknowledged it before continuing the comparison. The implementation note now names that version. That made the current reference clear for this handoff. Let's keep using the notice and acknowledgment when this file changes.” One handoff supports that observation; it does not prove a permanent pattern.

Sumi cannot yet send a message saying the numbers are fixed. Her fictional follow-up instead reads: “Your revised comment identifies the final chart point and the daily-table comparison. I can now trace the question. I have not finished checking the window definitions, so the discrepancy remains open. I will put the comparison result or the remaining blocker in the review thread at tomorrow's checkpoint.”

If the coworker disagrees with your account, return to the event and the requested action. Correct a mistaken detail openly. If you still disagree about the effect, say which part is your experience and look for a workable next step. You do not need agreement on every interpretation before agreeing to include a version identifier or a comparison source.

Before the next difficult conversation, write the observed event, its concrete effect, and one proposed action. Underline any explanation you have assumed rather than observed. Ask for the other person's perspective, revise the request where their information warrants it, and confirm each person's next action before treating the conversation as an agreement. Use this LandOffer.ai guide to prepare that conversation, then close with the action both people actually agreed to try.

Sources and Further Reading