“You need to be more strategic” can sound like meaningful feedback until the recipient tries to use it on Tuesday morning. What should they do differently? Which decision was weak? What would a better attempt look like? Without answers, the comment leaves them with a judgment and very little direction.
The person giving feedback may feel they’ve done the difficult part by speaking honestly. But honesty is only part of the job. If the purpose is improvement, the information needs to help someone change what happens next. A difficult truth delivered vaguely can be both painful and operationally useless.
I think of useful feedback as a contribution to the next attempt. It connects something observable with a consequence and a feasible adjustment. That doesn’t guarantee agreement or better performance. It does give the conversation a more useful destination than deciding whether the recipient is a good person.
Suppose a hypothetical sales representative gives a long product demonstration before asking about the customer’s situation. “You talk too much” turns the issue into a broad personal description. “In that demonstration, we reached minute twenty before learning which workflow the customer wanted to improve” identifies a moment that both people can examine.
The distinction matters because the recipient needs to locate the behavior. An example gives them somewhere to look. A character judgment invites arguments about all the occasions when it doesn’t apply. You can spend the entire meeting debating whether someone is generally considerate while never discussing the interruption that prevented a colleague from finishing a critical update.
Prepare one or two representative examples rather than an ambush assembled from months of irritation. Check that the examples are accurate and relevant to the standard. If you weren’t present, distinguish what you observed from what someone reported. Don’t present an inference about intent as if it were a recording of the event.
Avraham Kluger and Angelo DeNisi’s meta-analysis of feedback interventions found an average improvement alongside substantial variation, including interventions that reduced performance. Their proposed account emphasized where feedback directs attention, including attention toward the task or toward the self. The research is a reason to consider the design of feedback carefully, rather than assume any candid comment must help.
My practical interpretation is to keep returning to the work. Instead of trying to produce a particular emotional reaction, clarify what the person can notice and change. Embarrassment isn’t proof that the message landed. A calm response isn’t proof that it didn’t. The next attempt will tell you more about usefulness than the intensity of the meeting.
This also changes how you give positive feedback. “Great job” may be welcome, but it leaves the useful behavior unnamed. “Your summary put the customer’s decision and deadline in the first paragraph, so we could resolve the question without another meeting” identifies something worth repeating. Specific appreciation can teach without making every compliment a management technique.
A behavior matters because of what it changes. When you describe that consequence, stay proportionate. If the report arrived late, explain which decision was delayed. Don’t leap from a missed deadline to “You don’t care about the team.” The first is a consequence you can investigate. The second claims knowledge of a motive you may not possess.
Distinguish known effects from possible risks. “The missing attachment required two follow-up messages” describes something that happened. “It could have caused the customer to cancel” describes a possibility. Both may deserve attention, but treating the possibility as an accomplished disaster can make your account easier to dismiss.
Ask what the recipient saw from their position. Perhaps they prioritized an urgent customer issue because they thought the internal deadline was flexible. That doesn’t automatically excuse the outcome. It may reveal that your priority rules were unclear. A feedback conversation should be capable of uncovering information that changes the manager’s understanding too.
Once the gap is clear, agree on what to try. For the salesperson, the adjustment might be asking three diagnostic questions before opening the demonstration. For the report writer, it might be confirming the decision deadline and flagging a missing input a day earlier. Choose an action the person can recognize while doing the work.
Keep the size appropriate. If a presentation needs stronger structure, a request to “be more compelling” gives little assistance. A request to state the audience’s decision in the first minute provides a manageable test. After that improves, you can work on the evidence, pacing, or examples. Trying to correct everything at once may make it difficult to tell which change mattered.
Invite the recipient to propose an approach. They may understand the workflow better than you do. You can hold the standard while allowing flexibility in the method. “What would help you surface that risk earlier?” often produces a more realistic agreement than prescribing a process without knowing what resources or permissions it requires.
This is one reason accountability needs clear agreements. If you haven’t established the expected result, available authority, and review point, feedback may be correcting a standard that existed only in your head.
A person can know exactly what to do and still lack the time, information, or permission to do it. Imagine a support employee criticized for slow replies while covering twice the planned workload. A lesson on efficiency won’t by itself settle the capacity problem. You need to examine the conditions as well as the performance.
Useful diagnostic questions include whether the person understands the standard, can perform the task, has the necessary resources, and sees conflicting incentives. These are different issues. Training might address an unfamiliar procedure. A priority decision might address competing deadlines. A changed approval process might address a delay created outside the employee’s control.
Sometimes the standard is clear, resources are adequate, and someone repeatedly chooses not to meet an agreed obligation. That calls for a direct conversation about responsibility and consequences. You don’t have to make every performance issue a coaching opportunity forever. Fairness requires being clear about the situation you’re addressing rather than disguising a formal warning as a casual suggestion.
Recipients may dispute your example or your conclusion. Decide in advance whether your purpose is to understand and improve the work or to protect your initial wording. If the example is inaccurate, correct it. If the example is accurate but the interpretation is incomplete, refine the account. Changing your mind in response to information makes the conversation more credible.
You can also distinguish agreement about facts from agreement about standards. An employee might agree that an email went unanswered for two days while believing that response time was acceptable. The next discussion concerns expectations, not the existence of the email. If they believe the expectation is impossible, inspect that claim and decide what should change.
The broader principles of having difficult conversations early apply here. Choose a suitable setting, address the relevant issue, and give the other person room to respond. A rushed comment in front of colleagues may create a social problem that competes with the performance problem you intended to solve.
Before ending, summarize the agreement in ordinary language. What will the person try? What support will you provide? When will you review the result? The summary needn’t become a bureaucratic document. It should be specific enough that neither person has to reconstruct the conversation from a vague memory a month later.
At the review, look for evidence of the agreed behavior and its effect. If the new approach helped, say what improved. If it didn’t, inspect why. Perhaps the diagnosis was wrong. Perhaps the change was too large. Perhaps the person never attempted it. These explanations lead to different next steps, so don’t collapse them into a general impression of attitude.
Be careful not to move the target without acknowledgment. If you asked for concise updates and now want additional detail, explain what you learned and revise the standard openly. Otherwise, the recipient may reasonably conclude that compliance only generates a new reason for criticism. Stable expectations make it easier to learn from a genuine attempt.
Review your own contribution as well. If you promised an example, an introduction, or access to a tool, check whether you provided it. An employee shouldn’t have to meet a revised standard while quietly compensating for the support you forgot. Naming your missed commitment doesn’t erase theirs; it makes the complete agreement visible.
You can ask a simple closing question: “Was anything about my instruction or availability making this harder?” You may hear criticism you disagree with. Listen for the specific condition underneath it and decide what, if anything, needs adjustment. A leader who requests changes should also be prepared to inspect the conditions they control.
If you’re on the receiving end, you can help make vague feedback usable. Ask for an example, the consequence, and what a stronger attempt would look like. You don’t need to accept every interpretation to investigate whether something in the message is useful. You can acknowledge a specific mistake without endorsing a sweeping judgment about your character.
For example, you might say, “I understand that the recommendation wasn’t clear. Which decision did you need it to support?” That question takes the conversation toward the work. If the answer reveals that the assignment was ambiguous, clarify the next version rather than treating ambiguity as proof that nobody bears responsibility.
Feedback earns its place through what it helps people do. Give someone a clear view of the relevant behavior, a proportionate account of its consequences, and a next attempt they can actually make. Then remain involved long enough to learn whether the change helped. That’s a more demanding form of honesty, and a considerably more useful one.