Feedback is one of the most trained and least effective things leaders do. Everyone has done the course; everyone knows a model or two. And most feedback still changes nothing, because it founders on two distinctions that the models rarely make: the difference between an assertion and an assessment, and the difference between a grounded assessment and an ungrounded one. Get these right and feedback starts to work. Get them wrong and you're just transmitting your mood with a serious face on.
An assertion is a claim about a fact: “the report came in three days after the deadline.” It's true or false, and you can check it. An assessment is a judgement: “you're unreliable.” It isn't true or false in the same way — it's an opinion, and the person on the receiving end knows it. The trouble starts when leaders deliver assessments as though they were assertions: “you're not strategic enough,” stated as plain fact. The other person feels handed a verdict, gets defensive, and nothing moves — not because they're fragile, but because you've given them a character ruling to argue with rather than anything they can act on.
A grounded assessment is one backed by the assertions underneath it: the specific, checkable observations the judgement rests on, and the standard you're measuring against. “You're disorganised” is ungrounded and useless. Compare: “In the last month, three deliverables came in after the deadline without a heads-up — the offsite deck, the budget, and the client note — and what I need is to be able to plan around your commitments.” That names the evidence, names the standard, and leaves the person with something concrete instead of a label. They might still disagree, but now they're disagreeing about facts and expectations, which is a productive conversation, rather than about whether they're a good person, which is not.
The discipline is simple to state and genuinely hard to practise: before you offer any assessment, know the assertions that ground it. If you can't name the specific observations behind “not strategic enough,” you aren't ready to say it — not because it would be unkind, but because it won't land. Ungrounded feedback is just an opinion delivered with authority, and people discount it accordingly. Half the feedback that “didn't take” failed for this reason alone: there was nothing underneath it to hold onto.
If you can't name the specific observations behind your judgement, you're not ready to give it. Ungrounded feedback is just your mood, transmitted.
Even perfectly grounded feedback changes nothing if it stops at the judgement. The final move is the one most often missing: a clear request for what you want instead. Not “you need to be more proactive” — that's another ungrounded assessment in disguise — but “going forward, when a deadline is at risk, I'd like to hear from you the day before, not the day after.” That is a request someone can actually act on, agree to, or renegotiate. It converts feedback from a statement about the past into a commitment about the future, which is the only place behaviour change can happen.
Put the whole sequence together and watch how different it feels. Instead of “you've been unreliable lately” — a verdict that invites a fight — you get: “Three deadlines slipped this month without warning [the assertions]; I need to be able to depend on your commitments to plan the team's work [the standard]; so from now on, flag a slipping deadline the day before, and we'll re-plan together [the request].” Same concern, completely different outcome. The first version makes someone feel bad for an afternoon. The second gives them something specific to do differently on Monday.
Everything here applies just as much to positive feedback, which fails for exactly the same reason criticism does. “Great job” is an ungrounded assessment, and people discount it just as fast, because it tells them nothing about what to repeat. “The way you handled that client's objection — naming the concern before they'd fully raised it, then putting the data in front of them — is what kept the deal alive” is grounded praise, and it does real work: it tells the person precisely which behaviour earned the result, so they can do it again on purpose. Most leaders are as lazy with praise as with criticism, scattering vague approval that feels pleasant and teaches nothing. Ground it, and praise quietly becomes one of the most efficient development tools you have — you are, in effect, telling someone exactly which of their instincts to trust.
Grounded assessment, plus a clear request. Do those two things and feedback stops being an uncomfortable ritual that changes nothing, and becomes what it was always supposed to be — the ordinary mechanism by which capable people get better.