“Good job” is pleasant but weak as feedback because it does not tell the person what worked, what effect it had, or what to repeat next time. A useful feedback system needs to be fast enough for real classrooms and meetings while still giving the receiver something they can act on.
The CueLab E-I-N feedback method
Use three parts: Evidence, Impact, and Next move. The sequence keeps feedback grounded in something observable and avoids vague judgments about personality or talent.
| Part | Question | Example |
|---|---|---|
| Evidence | What did the person actually do? | “You paused after showing the chart and named the blue line before explaining the change.” |
| Impact | What did that choice help or hinder? | “That gave the audience time to locate the series before the comparison started.” |
| Next move | What should be repeated, changed, or tested next? | “Use the same pause-and-name pattern on the next two charts.” |
Why this is different from praise
Praise answers, “Did I like it?” Feedback answers, “What happened, and what should happen next?” Both can have a place, but they are not interchangeable. If the goal is improvement, the receiver needs a description they can recognize and a next action they can attempt.
This does not mean feedback should sound cold. A warm sentence can still be specific. “That opening was much easier to follow because you stated the decision before the background” is more useful than “Great opening.”
Use a four-level rubric without turning it into a grade
The rubric below is for diagnosis, not certification. Do not add the scores into a fake scientific total. Use the levels to identify where the next revision effort belongs.
| Dimension | 1 — unclear | 2 — partly working | 3 — clear | 4 — deliberate |
|---|---|---|---|---|
| Message | The main point is hard to identify. | The point becomes clear late or after explanation. | The main point is clear in the section. | The point is clear and reinforced by the structure. |
| Evidence | Examples or visuals do not clearly support the claim. | Some evidence helps, but its relevance needs explanation. | Evidence clearly supports the claim. | Evidence is selected and sequenced to make the comparison easy to judge. |
| Delivery | Pace, wording, or timing regularly interferes with understanding. | Several moments are clear, but difficult sections become rushed or flat. | Pace and emphasis generally support the explanation. | Timing changes deliberately around important ideas and audience tasks. |
| Visual guidance | The audience must search for what matters. | Some hierarchy exists, but competing elements remain. | The visual path is usually obvious. | Labels, contrast, spacing, and timing direct attention exactly where needed. |
| Audience check | The speaker finishes without testing understanding. | Questions are invited, but checks are broad or late. | The speaker uses at least one specific check. | Checks are placed where misunderstanding is most likely and influence what happens next. |
Worked example: feedback on a presentation opening
Weak: “Good introduction, but try to be more confident.”
Stronger: “You started with two minutes of background before saying what decision the team needed to make. I understood the context, but I did not know why the details mattered yet. Try stating the decision in the first thirty seconds, then keep only the background that helps the audience evaluate it.”
The stronger version avoids guessing at confidence. It points to an observable sequence and gives the speaker a revision they can test.
Worked example: feedback on a teaching explanation
Weak: “Students looked confused.”
Stronger: “After the second example, three students could repeat the procedure but could not explain why the second step was necessary. Before moving to the next problem, ask them to compare two cases and explain which rule changes.”
This keeps the feedback tied to evidence rather than treating facial expression as proof of understanding. CueLab’s guide on checking whether students understand an explanation gives more ways to test the learning rather than infer it.
Worked example: feedback on a chart slide
Weak: “The chart is too busy.”
Stronger: “The slide asks the audience to compare the red and green lines, but four other series use equally strong colours and the legend is separated from the plot. Fade the four background series and label the two important comparison lines directly.”
The revision identifies what “busy” means in this case. For more on that specific design choice, see direct labels versus legends.
Feedback questions that produce better answers
- What was the main point you took from this section?
- Where did you have to work hardest to follow me?
- Which slide helped the explanation most, and why?
- Which transition felt abrupt?
- What did you expect me to explain next?
- If I had to cut two minutes, what would you remove first?
These are more diagnostic than “Was that good?” because they ask the listener to describe the experience rather than approve the presenter.
Feedback phrases to avoid or repair
| Vague phrase | Why it fails | Repair |
|---|---|---|
| “Be more confident.” | It labels an internal state and gives no behavior to change. | Name the observable behavior: volume, pacing, eye position, hesitations, or wording. |
| “Make it engaging.” | It does not identify the audience task. | Specify whether the audience needs a question, example, comparison, or decision point. |
| “Too much text.” | It may be true but does not say what the text is doing. | Identify what should be spoken, shown, moved to notes, or removed. |
| “Great visuals.” | The receiver cannot tell what to repeat. | Name the useful choice: direct labels, hierarchy, spacing, annotation, or sequence. |
Peer-review sheet
| Moment | Evidence I noticed | Impact on me as listener | Next move I suggest |
|---|---|---|---|
| Opening | ________________ | ________________ | ________________ |
| Strongest explanation | ________________ | ________________ | ________________ |
| Hardest section to follow | ________________ | ________________ | ________________ |
| Most useful slide | ________________ | ________________ | ________________ |
| One revision to test | ________________ | ________________ | ________________ |
Give fewer comments, but make them usable
Five pages of comments can leave a presenter unsure where to start. After reviewing the work, choose one thing to preserve and no more than three changes to test. Rank them by effect on understanding rather than by how easy they are to notice.
If a speaker needs help with rehearsal after receiving feedback, use the CueLab Speaking Rehearsal Worksheet to turn the comments into a new run rather than a list of intentions.