Begin with your experience, not their mistakes

When someone hands you a draft, they are not asking you to become an editor with a red pen. They are asking what happened to you while you read. That distinction is the foundation of generous, genuinely useful feedback.

So start there. "I reread the opening twice because the voice pulled me in." "By the fourth page I'd lost track of who was speaking." Both sentences are about you, the reader, and both give the writer something they cannot get anywhere else: an honest record of the effect their prose had on a mind that isn't theirs.

Try to avoid verdicts dressed up as analysis — "this is confusing", "the pacing is bad". Trace instead the moment the confusion began, or the paragraph where your attention drifted. Writers can act on a located problem. They cannot act on a verdict.

Ask questions rather than issuing commands

Commands close a conversation. "You need to cut this scene." "Change that metaphor." Even when you are right, the writer is left holding a correction rather than a craft problem they now understand.

Questions open one. They also have the useful property of being honest about the limits of your reading:

• What did you want me to feel here? — useful when a scene lands flat.
• Whose head are we in during this paragraph? — useful when point of view slips.
• Is this image doing something the plain sentence couldn't? — useful when a metaphor feels decorative.
• What happens if this arrives two pages later? — a way of testing structure without prescribing it.

Each question hands the decision back. The writer stays the author of their own book, which is exactly what you want. Your task is to widen their sense of the options, not to narrow it.

Name the specific passage, and say what it did to you

Vague praise is as unhelpful as vague criticism. "I loved the atmosphere" tells the writer nothing they can repeat next time. "The line about the tide going out like a breath held too long — I stopped and read it again" tells them precisely which instinct to trust.

Do the same with difficulty. Rather than "the middle dragged", quote the words: "I lost my footing in the paragraph beginning 'The kitchen still smelled of him' — I think because we jump from her memory of the funeral into the present without a signal."

Three habits make this easier:

• Quote the exact words. Paraphrase blurs the evidence.
• Report the effect, not the diagnosis. "I felt shut out" beats "the prose is opaque".
• Distinguish "I didn't understand" from "this isn't clear". The first is a fact about you; the second is a claim about the text, and one the writer may reasonably dispute.

Feeding back on metaphor specifically

Metaphor is where feedback most often goes wrong, because readers tend to offer taste — "I don't like it" — when what they mean is "this doesn't quite work yet". Separate the two.

Ask first what the image is trying to do. A metaphor carries a vehicle (the thing invoked) and a tenor (the thing meant). When they pull in opposite directions, or the vehicle drags in associations the story doesn't want, you will feel it as a jolt. Things worth noticing and naming:

• Clashing vehicles. Two images yoked together that fight each other, so the reader ends up picturing neither.
• Dead metaphors. "Heart of stone", "a sea of faces", "light at the end of the tunnel". The eye slides straight off them.
• Over-extension. A sustained metaphor is a gift when it develops. When it merely restates, it becomes a lecture.
• Mismatched register. A lyrical image wedged inside a clipped, hard-boiled paragraph.
• Abstraction smuggled in as image. "A profound sense of emptiness" is not a picture; it is a summary of one.

Then check where the imagery comes from. A metaphor drawn from a character's own working life — a joiner thinking in terms of joints and grain, a nurse in terms of triage — is doing two jobs at once. If the figurative language could be lifted into any other novel unchanged, that is worth mentioning gently.

Prioritise, and say what is working

A reader who returns forty comments is a reader the writer cannot hear. Choose the two or three things that would most change the piece and say those clearly. List the rest briefly as smaller observations the writer can take or leave.

Say what is working, too — precisely, and not as a cushion before the criticism. Writers need to know which instincts to trust, and they will only believe you if the praise is as specific as the notes.

Finish by asking about intention: "What were you trying to do with the ending?" The answer often tells you more about the draft than anything you have written yourself. And remember to name your own position. This is one reader's response, offered honestly — not a ruling. The writer decides what to do with it.