Giving Feedback on Work: Making It Useful Instead of Just Kind

Feedback fails in two directions. Too vague and it changes nothing. Too much and the learner cannot act on any of it.

The useful middle is narrower than most people expect: specific, limited, and pointing at a next action.

What makes feedback actionable

• Point at the specific place — the line, the section, the moment.
• Say what the issue is, not just that something is wrong.
• Say what to do instead. Feedback without a direction leaves the learner stuck with a problem they already sensed.
• Explain why it matters — this is what lets them apply the principle to the next piece of work rather than just fixing this one.

The test: could the learner make a specific change based on what you wrote? If not, it is a judgement, not feedback.

How much to give

• Three points is usually the limit of what someone can act on.
• Prioritise ruthlessly. Fix what matters most, ignore the rest for now.
• Address the structural problem before surface details. Correcting wording in a section that needs deleting wastes both your time and theirs.
• A page of comments produces less improvement than three good ones — and often produces none, because the learner freezes.

Structure that works

• Say what worked, specifically. Not as a cushion, but because they need to know what to keep — learners often remove their best work in revision because nobody told them it was good.
• Give the main issues, with location and direction.
• Say what to do next — one concrete action.
• Keep the focus on the work, not the person. Describe the work, not their ability.

Tone

• Be direct. Softening feedback until the message is unclear is not kindness — the learner submits the same problem again.
• Avoid sarcasm entirely in written feedback; it reads far harsher than intended.
• Be careful with questions used as criticism — asking why they did something can read as an accusation. Say what you observed instead.
• Assume good intent. Most problems in submitted work are misunderstandings, not carelessness.
• Written feedback lands harder than spoken. Reread it once before sending, imagining you received it.

Feedback at scale

When you have many learners and limited time:

• Build a bank of common comments you can reuse — but always add one line specific to that person, or it reads as a template and they disregard all of it.
• Record a short audio or video comment — often faster than writing and it carries tone better.
• Give group feedback for common issues and individual feedback only for what is individual.
• Publish a worked example after an assignment — this answers many questions at once.
• Use a checklist for self-assessment first so learners catch the obvious problems before you spend time on them.

Peer feedback

• It works, but only with structure. Unstructured peer review produces vague praise.
• Give specific prompts — ask reviewers to identify one thing that worked and one thing unclear.
• Teach how to give feedback before asking for it.
• Giving feedback teaches more than receiving it — reviewing others' work is one of the strongest learning activities available.
• Check the quality of peer feedback early rather than assuming it is happening.

When someone reacts badly

• Give it a moment. A defensive first reaction often settles.
• Restate the specific observation, calmly, without adding new criticism.
• Ask what part they disagree with — sometimes they have a reason you did not know.
• If you were wrong or unclear, say so.
• Do not soften the substance to end the discomfort — that helps nobody.

Timing

• Sooner is better — feedback on work from three weeks ago arrives after the learner has moved on.
• State your turnaround time and meet it. Unpredictable timing is worse than slow timing.
• Feedback during the work beats feedback after it where possible — a check-in at the halfway point prevents a lot of wasted effort.

Feedback on early attempts

• Be gentler on first attempts — not less specific, but more selective. One or two points.
• Name what they got right, because beginners genuinely cannot tell.
• Set the expectation that revision is normal — many learners think needing changes means failure.
• Watch for someone who stops submitting after feedback. That is a signal worth following up on privately.

One thing worth remembering

The purpose of feedback is not to evaluate work that is already finished. It is to change what the learner does next.

Read what you wrote and ask whether it does that. If it does not, it is a grade with extra words.

FAQ