The Feedback Leaders Give — and Why It Rarely Changes Anything

The Problem

Most leaders give feedback regularly. Most of it does not produce the behavioral change they intend. This is not a reflection of effort — it is a structural problem with how the feedback is delivered.

The most common feedback leaders give is evaluative: it describes what happened and whether it was good or bad. What it rarely includes is the specific behavioral alternative — the precise, observable change the leader actually wants to see. Without that, feedback is a verdict, not a direction.

What the Research Shows

Research in organizational psychology distinguishes between two types of feedback: outcome feedback, which tells someone how they performed, and process feedback, which addresses the specific behaviors that produced the outcome. Studies by Kluger and DeNisi (1996) — among the most cited in the feedback literature — found that feedback interventions improve performance in less than half of cases, and in roughly a third of cases, feedback actually decreases performance.

The mechanism is not mysterious: feedback that focuses on the self rather than the task activates defensive processing rather than behavioral learning. Evaluative feedback about who someone is, or even how they did, tends to produce self-protection. Specific process feedback — describing exactly what happened and what could be done differently — tends to produce change.

Feedback that is accurate but not actionable does not change behavior. It only confirms a judgment.

 Three Structural Adjustments

1. Name the specific moment. Rather than 'your presentation wasn't landing,' identify the precise point of breakdown: 'In the first three minutes, the main point wasn't stated — the audience was still waiting for it when you moved to the data.' Specificity removes defensiveness because it is harder to argue with a moment than with a characterization.

2. Describe the alternative behavior, not the desired outcome. Telling someone to 'be more concise' describes an outcome. Telling someone to 'state the recommendation in the first sentence, then support it' describes a behavior. The first requires interpretation. The second requires only implementation.

3. Separate observation from evaluation in the delivery. The sequence matters: describe what you observed before offering any judgment. 'I noticed you qualified the recommendation three times before stating it — that pushed the main point to the two-minute mark' is easier to receive than 'You were too hesitant.' One is a data point. The other is an identity statement.

The Practical Takeaway

Think of the last piece of developmental feedback you gave. Can you identify the specific behavioral alternative you communicated? If the answer is unclear, the person who received the feedback is probably still waiting for it.