Feedback That Lands

You’ve built the container — a recurring one-on-one that belongs to the person you manage — and given it real content. Now comes the thing that container was built to carry. Feedback is the daily currency of helping someone grow, and it’s the move most managers do badly, not for lack of caring but for lack of technique.

This page is about the giving half of feedback. It makes one claim: feedback is information the receiver can act on — and if they can’t act on it, it isn’t feedback, it’s just a feeling you dumped on them. Everything else here follows from taking that seriously. We’ll diagnose why the usual attempts fail, name the three properties that make feedback useful, and give you one concrete structure — Situation-Behavior-Impact — to build every piece of feedback on.

Why most feedback fails

Picture the two most common kinds of feedback a manager actually gives. The first is vague: “You need to be more proactive.” “Great job this quarter.” “Just be more of a team player.” The second is late: a manager sits a person down in a review and lists things that happened weeks or months ago.

Both fail for the same underlying reason — the receiver can’t do anything with them.

Take “be more proactive.” What, precisely, should the person do differently tomorrow morning? They don’t know. They might guess you mean “email me updates,” when you actually meant “flag risks before I have to ask.” The words feel like feedback, but they carry no instruction. The receiver is left to reverse-engineer your meaning, usually badly, and often anxiously. Vague praise fails the same way in the other direction: “great job” feels nice for an afternoon but teaches nothing — the person can’t repeat a success they can’t identify.

Now take the delayed feedback. Even if it’s specific, three months is too long. The person barely remembers the meeting you’re describing. They can’t inspect the moment, because it’s gone. Worse, they’ve been doing the thing wrong — or right — for three months, so either a bad habit has set or a good one went unreinforced. Late feedback also lands as an ambush: if this mattered, why am I only hearing it now? Trust takes the hit.

So the two most common feedback failures aren’t about tone or courage. They’re about actionability. Vague feedback gives nothing to act on; late feedback gives nothing you can still act on. Fix those two and you’ve fixed most of the problem.

The common trap — mistaking venting for feedback

The moment you’re most tempted to “give feedback” is right after someone frustrates you — and that’s the moment it’s least likely to be feedback at all. What comes out is usually a release of your feeling (“this is sloppy,” “you always do this”) dressed as help. The test is simple: could the receiver take your words and know exactly what to do differently? If not, you’re venting, not coaching. Venting relieves you and costs them. Wait, sharpen it into something actionable, then deliver it.

The three properties of feedback that lands

Feedback the receiver can act on has three properties. Miss any one and it slips.

1. Specific, not general

Name the exact thing. Not “your communication needs work” but “in yesterday’s client call, you answered the pricing question before letting them finish asking it.” The more concrete you are, the less the receiver has to guess, and the less room there is to argue about what you really meant. Specificity is a gift: it hands the person a single, clear thing to change or keep, instead of a fog to worry about.

This applies to praise just as much as correction. “Nice work” teaches nothing. “The way you summarized everyone’s position before we voted — that’s what kept the meeting from going in circles” tells the person exactly which behavior to do again.

2. Timely, not stored up

Give it close to the event, while the moment is still fresh for both of you. Hours or a day or two — not weeks, and definitely not “at the review.” Timeliness means the person can still inspect the moment, the habit hasn’t hardened, and the feedback reads as ordinary maintenance rather than a stored-up grievance. A rough rule: the half-life of useful feedback is short. If you find yourself saving items for a big conversation later, you’ve already let most of their value evaporate.

Timely does not mean instant or in front of others. Sometimes the kind, clear move is to wait until the one-on-one that afternoon, or to pull the person aside. Timely means soon, not reckless.

3. About behavior, not the person

This is the one that separates feedback that helps from feedback that wounds. Aim at what the person did, never at who they are.

“You were unprofessional” is an attack on identity — it labels the person, and identity is not something anyone can change by Thursday. “You interrupted the client twice while they were describing the problem” is about behavior — it’s a specific action the person can choose to stop. Behavior is changeable; character feels fixed. The instant feedback becomes a verdict on someone’s character, they stop listening and start defending, because you’ve made it about their worth rather than their work.

Person (fails)                    Behavior (lands)
------------------                --------------------------------
"You're careless."          →     "The last two reports had the
                                   client's name misspelled."
"You're not a team player." →     "In standup you talked over
                                   Sam twice before he finished."
"You're a great leader."    →     "You made sure the quietest
                                   person got heard before we
                                   decided."

Same information, completely different reception. The left column starts a fight; the right column starts a change.

The structure: Situation-Behavior-Impact

You now know the three properties. The problem is remembering all three in the moment, especially when you’re nervous or annoyed. That’s what a structure is for. The most useful one is Situation-Behavior-Impact, or SBI — three sentences, in order.

• Situation — anchor the moment. When and where did this happen? This makes the feedback specific and timely automatically, because you’re pointing at one concrete event, not a pattern in the abstract.
• Behavior — describe what you observed. What did the person actually do or say? Only observable actions go here — the things a camera would have caught. This is where you keep it about behavior, not character.
• Impact — explain the effect. What happened as a result — on the work, the team, the client, you? This is the part that answers the receiver’s natural “so what?” and tells them why it’s worth changing.

Situation and Behavior keep you honest and concrete; Impact supplies the reason to care. Here’s a full example.

In practice — the same feedback, before and after SBI

A designer, Priya, keeps delivering work an hour before deadline, leaving no time for review. The manager’s first instinct — the version that fails:

“Priya, you’ve got to be more reliable with deadlines. It’s becoming a problem.”

Vague (“more reliable”), a label (“a problem”), no anchored moment. Priya doesn’t know what to fix and feels judged. Now the SBI version:

Situation: “On the Henderson brief yesterday —” Behavior: ”— you sent the final files at 4pm for a 5pm deadline.” Impact: “That left no time to catch errors before it went to the client, and I ended up reviewing it in a rush and missing the wrong logo.”

Now Priya knows the exact event, the exact behavior, and precisely why it matters. There’s nothing to argue about and something clear to change — build in a review buffer. Notice the manager didn’t even have to say “be more reliable.” The impact makes the point, and Priya gets to draw the conclusion herself, which she’ll own far more than one you hand her.

SBI works identically for praise — and you should use it there too, because specific praise is how you get more of a behavior. “In this morning’s handover [situation], you read back the patient’s medication list to the incoming nurse [behavior], which is exactly what stops a missed dose [impact].” That nurse now knows precisely what to keep doing.

Observation, not interpretation

The single most common way SBI goes wrong is that people smuggle a judgment into the Behavior step. They think they’re describing behavior, but they’re actually reporting the story they told themselves about it.

The distinction is between observation — what a camera would have recorded — and interpretation — the meaning or motive you inferred from it. “You didn’t respond to my three messages” is an observation. “You ignored me” is an interpretation: ignored assumes intent you can’t see. Maybe they were heads-down, out sick, or never got the messages. When you lead with the interpretation, you’re handing the receiver a motive to deny, and they will — the conversation becomes a debate about what was in their head instead of what actually happened.

Interpretation (arguable)        Observation (hard to argue)
--------------------------       ----------------------------------
"You ignored me."          →     "I didn't get a reply to my
                                  three messages this week."
"You don't care about       →    "You joined the client call
 quality."                        without the numbers we'd agreed
                                  to bring."
"You were being defensive." →    "When I raised the timeline, you
                                  said 'that's not on me' before I
                                  finished."

Stick to observation and your feedback becomes almost impossible to dispute, because you’re only claiming what both of you can see. The receiver may still explain why they did it — good, that’s the conversation you want — but they can’t tell you it didn’t happen. Save your interpretation for a question: “I noticed X — what was going on?” That invites the reason instead of assuming it.

The evidence — behavior beats the person

Research on feedback and motivation generally converges on an uncomfortable point: feedback aimed at the person can lower subsequent performance rather than raise it, because it triggers self-protection instead of problem-solving. A widely cited review of feedback interventions (Kluger and DeNisi, 1996) found that a large share of them failed to help, and some actively hurt — with feedback that drew attention to the self (praise, blame, ego) among the worst performers, while feedback focused on the task tended to help. Treat the exact numbers loosely, but the direction is robust and matches the SBI logic: aim at the work, not the worth.

Be kind and clear, not kind or clear

There’s a temptation, once you understand how easily feedback wounds, to soften it into nothing. The classic form is the feedback sandwich: bury the real message between two slices of praise. “You’re doing great, [the actual issue], but overall really strong work!”

It fails for a precise reason. The receiver either misses the message entirely — they hear the two compliments, relax, and walk out thinking everything’s fine — or they learn to distrust your praise, because they now know a compliment is just the wrapper around bad news. Either way you’ve corrupted both signals: the correction is muffled and the praise is poisoned. Other softeners do the same damage — “this is probably nothing, but…”, “I might be totally wrong, but…”, hedging that trains the person to discount whatever follows.

The instinct behind softening is decent: you don’t want to be cruel. But the fix isn’t to blur the message — it’s to be kind and clear at the same time, not to trade one for the other. Kindness is in your manner: private, calm, respectful, genuinely on their side. Clarity is in your message: the actual thing, stated plainly, unhedged. A person can absolutely hear “here’s a real problem” delivered warmly and directly. What they can’t work with is a real problem hidden so well they never find it.

Kind OR clear (pick one — both fail):
  Cruel & clear:   "This is sloppy work, honestly."
  Kind & vague:    "Great stuff! Maybe tighten it up a bit? Amazing though!"

Kind AND clear (the target):
  "I want this project to go well and I think you can nail it —
   so I want to be straight with you: the last two reports went
   out with errors that reached the client. Let's figure out a
   check step before they ship."

Warmth in the delivery, no fog in the message. That’s the whole discipline. Give praise separately and often (that’s its own skill), and when there’s a problem, name it cleanly.

How it connects

Everything on this page assumes the container from the one-on-one: feedback given inside an established, trusting relationship lands where the same words dropped on someone cold would sting. Feedback is also only half a loop — how you take it, covered in Receiving Feedback Well, decides whether anyone dares give you the truth. And SBI is the same backbone you’ll lean on when the stakes rise in Hard Conversations Without Flinching — a hard conversation is mostly just feedback that matters more.

Try this

Take one piece of feedback you’ve been sitting on — the one you noted in the part’s opening inventory, if you did it. Before you deliver it, write it out as three literal sentences: Situation, Behavior, Impact. Then read your Behavior line and cross out anything a camera couldn’t have recorded — every “ignored,” “didn’t care,” “was rude” — and replace it with the plain action you actually saw. Deliver it this week, warmly and without the hedges. Notice how much shorter and calmer the conversation is when there’s nothing to argue about.

Reflect

1. Think of the last feedback you gave. Was it specific enough that the person knew exactly what to do differently — or did you leave them to guess what you meant?
2. What feedback are you currently storing up for “the right moment”? What is the delay costing — in the behavior continuing, and in how it’ll land when you finally say it?
3. Look at a recent piece of feedback you gave (or wanted to). Was it aimed at what the person did, or at who they are? How would you rewrite the character version as a behavior?
4. Where do you catch yourself reporting interpretations (“she was being defensive,” “he doesn’t care”) as if they were facts? What did you actually observe in those moments?
5. Do you soften real messages until they disappear — sandwiches, hedges, “it’s probably nothing”? Whose feedback have you misread because it was buried that way?

Show reflections

1. If you can’t reconstruct one concrete action from the feedback you gave, the receiver couldn’t either — which means it did little. A strong answer catches the vagueness and rewrites it as a single, nameable thing to change or repeat.
2. Stored feedback is decaying feedback. The honest cost is double: the behavior keeps happening because you never flagged it, and when you finally do, the pile lands as an ambush that damages trust. The fix is to give it small and soon, not big and late.
3. This is the highest-leverage rewrite in the whole page. “You’re careless” has no exit; “the last two reports had errors” has an obvious one. If your feedback names a trait, you’ve likely started a fight instead of a change — convert it to an action.
4. Almost everyone narrates motives as facts. The useful move is to strip your feedback back to the camera-level observation and hold the interpretation as a question (“what was going on?”) — which invites the real reason instead of assuming a bad one.
5. If you soften habitually, you’re probably being misheard more than you know — people walk away thinking things are fine. And recognizing feedback you received that was buried is the clue: kind-and-clear would have helped you more than kind-and-vague did. Aim for both, not either.