A writer hands me a draft and says, “Be honest, tear it apart.” They almost never mean it. What they mean is closer to “tell me it’s good, and if it isn’t, tell me gently, and also tell me I’m capable of fixing it.”
I learned to hear that gap early. The instruction to be brutal is rarely an invitation to be brutal; it is a pre-emptive flinch, a way of bracing for criticism by pretending to want more of it than anyone could stand. Read the request literally and you lose the writer. Read what sits underneath it and you have a chance of being useful.
I have edited thousands of manuscripts since I started The Expert Editor in 2012, and the thing that took me longest to understand was that most advice about giving feedback aims at the wrong target. It treats the problem as one of tact: soften the blow, sandwich the criticism, lead with something nice.
That framing misses what actually goes wrong. The damage rarely comes from a single harsh sentence. It comes from the writer deciding, somewhere in the exchange, that you do not respect them, or that the work is hopeless, or that the relationship now contains a scorekeeper. Those conclusions are mostly avoidable, and avoiding them has little to do with niceness.
The first question a writer asks is not “is this true” but “why are you saying it”
Before a writer can act on a note, they have to decide what the note means about them and about you. Douglas Stone and Sheila Heen, who teach negotiation at Harvard Law School, call this a relationship trigger: in Thanks for the Feedback, they argue that the identity of the person giving feedback shapes how it lands at least as much as the content does. The same sentence — “this section drags” — reads as helpful from someone the writer trusts and as an attack from someone they suspect of indifference or competition.
It is why a note from a stranger on the internet bounces off, while the same observation from an editor the writer trusts reorganises a manuscript. The writer is not weighing the words. They are weighing the source. When I was new to this and led with a barrage of corrections, however accurate, I watched them get ignored — not because the writer disagreed, but because they never got past the question of motive. The corrections were right and useless at the same time.
Praise is not the opposite of criticism, and treating it that way backfires
The compliment sandwich assumes feedback works on a scale from harsh to kind, and that the job is to dose the harshness. But one of the largest analyses of feedback ever assembled suggests the variable that matters is something else entirely. In a 1996 meta-analysis in Psychological Bulletin covering 607 effect sizes, Avraham Kluger and Angelo DeNisi found that feedback improved performance on average — but that more than a third of feedback interventions actually made performance worse. The deciding factor was not whether the feedback was positive or negative. It was where it pointed the recipient’s attention. Feedback that drew attention to the self — to the writer’s ability, standing, or worth — tended to hurt. Feedback that stayed on the task tended to help.
This reframes the whole problem. “You’re a strong writer, but this chapter is weak” is not a gentle note; it is two self-directed statements that invite the writer to think about their rank rather than their paragraph. “This chapter loses the thread in the third section because the timeline jumps” points at the page. The second is harder to write and easier to use. The notes I see acted on are almost always the ones a writer can carry out without first having to decide how they feel about themselves.
High standards and confidence in the writer have to arrive together
There is a specific combination that makes criticism usable, and it has been tested directly. In a 2014 paper in the Journal of Experimental Psychology: General, David Yeager, Geoffrey Cohen and colleagues studied how to deliver critical feedback to adolescent students in a context shaped by mistrust. For one group, the critique on their essays came with a short added message: that the reviewer had high standards and believed the student was capable of meeting them. Among the students most inclined to distrust the feedback, that added line more than doubled the rate at which they revised their work, and the revisions were better.
The mechanism is worth dwelling on, because it is easy to mistake for flattery. The message did not soften the criticism; the criticism stayed exactly as pointed. What changed was the writer’s reading of why it was being given — not because the reviewer thought little of them, but because the reviewer thought enough of them to hold the bar high. A good editor does a version of this almost reflexively: the unspoken contract is “I am marking this up precisely because I think you can clear it.” My mistake, for years, was assuming that contract was obvious. It usually has to be said out loud.
Description travels further than judgement
Peter Elbow, who spent a career studying how writers respond to readers, argued that the most useful thing a reader can offer is often not evaluation at all but an honest account of what happened to them while reading. In his essay “About Responding to Student Writing,” Elbow makes a claim that should unsettle anyone confident in their feedback: students learn a great deal from writing, he notes, but surprisingly little from the comments written on it. The comment that helps is the one that fits this draft, this writer, at this moment — and that depends on knowing what the writer was trying to do.
So I try to tell the writer where I got confused, where I stopped believing the argument, where I skimmed — rather than ruling on whether the prose is good. “I lost track of who was speaking here” is something a writer can investigate. “This dialogue is confusing” is a verdict they can only accept or resist.
Description hands the problem back to the person who owns it. Judgement keeps it on my side of the desk, which is exactly where it cannot be fixed.
The relationship is information, not an obstacle
Feedback gets framed as a tension between honesty and tact, as though protecting the relationship means diluting the truth. That is the wrong model, and it is the one I most often have to talk editors out of. The relationship is what makes honesty possible in the first place. A writer who trusts that you are on their side will absorb a level of criticism that would devastate them coming from someone they suspect of contempt. The trust is not a softening agent applied on top of the feedback; it is the channel the feedback travels through. Spend it carelessly — by scoring points, by being right at the writer’s expense, by mistaking bluntness for rigour — and the channel narrows, and the next true thing you say arrives distorted or not at all.
None of this is an argument for pulling punches. The writer who said “tear it apart” does want the weak chapter named; they simply also need to finish the exchange believing the work is salvageable and that I think so too.
Those two things are not in conflict, though the worst feedback treats them as if they were — and the surest way to ruin a draft is to be completely right about it in a way that leaves the writer unable to do anything but agree.
