You’re reading the staging page and a sentence stops you. Not because it’s wrong, exactly — the page still works, the button still goes where it should — but because the sentence doesn’t sound like your company. Or it’s technically accurate and somehow still confusing. Or it’s fine on its own and completely wrong next to everything around it.

You have two bad options in front of you. Rewrite the sentence yourself, which means you’ve now written a second draft that someone else has to reconcile with the first. Or write “this doesn’t sound right,” which gives your copywriter or developer nothing to act on — the same unresolved feeling you started with, now in their inbox instead of your head.

There’s a third option, and it’s the same shift that works for visual feedback: describe the problem precisely instead of prescribing the fix. Copy feedback just needs its own version of that technique, because the thing you’re describing isn’t a visual element — it’s meaning, tone, and brand voice, and those need a different vocabulary than “make it bigger” or “move it up.”

Why is copy feedback harder to give than visual feedback?

Copy feedback is harder because the problem isn’t where something is — it’s what something means, and meaning doesn’t have a pixel you can point to. With a button that’s the wrong color, you can click on it and say “this.” With a sentence that reads too salesy, there’s no visual artifact to anchor the comment to; you have to name an impression instead of an object.

That gap matters more than it looks like it should, because copy is doing more work on the page than most founders give it credit for. Nielsen Norman Group’s research on web reading behavior found that visitors read only about 20–28% of the words on an average page — on a page with roughly 600 words, that’s a couple hundred words a visitor actually absorbs before they decide whether to keep going. Every sentence that makes the cut is disproportionately load-bearing, because most of the sentences around it never get read at all. A visual bug on a page nobody looks at closely is forgivable. A confusing sentence in the 20% visitors actually read is the whole experience.

It also compounds with brand consistency, which copy carries more directly than layout. Lucidpress and Demand Metric’s benchmark research on brand consistency found that companies with consistently applied brand voice saw revenue increases up to 33% compared to companies with inconsistent branding — and separately, 81% of the companies surveyed admitted to dealing with off-brand content regularly. Copy is one of the easiest places for that drift to start, because unlike a layout change, a tone shift can slip through review without anyone flagging it as a “design” issue at all.

Should you rewrite the confusing sentence yourself, or describe what’s wrong with it?

Describe what’s wrong with it — rewriting it yourself just adds a second draft that someone else now has to reconcile with the first. If you’re not the person who’s going to own the copy going forward, your rewrite isn’t a fix. It’s an alternate version that the actual writer or developer now has to compare against the existing draft, decide between, and probably rewrite again to match the voice used everywhere else on the site.

This is the same off-brand-drift risk the Lucidpress research flags, just introduced from the review side instead of the writing side. A founder’s improvised rewrite, written in the moment, under none of the constraints the original copy was written under, is a likely source of exactly the inconsistency that research found cutting into revenue. The founder isn’t wrong that something needs to change — they’re just not the best-positioned person in the loop to decide what the replacement sentence should literally say.

What you are best positioned to do is notice the problem and name it precisely. You read the page as a visitor would. That’s valuable information. It stops being valuable the moment it turns into your own first draft of a replacement, because now the writer isn’t fixing the original problem — they’re mediating between two versions of the page that didn’t exist an hour ago.

How do you describe a tone problem without rewriting the sentence yourself?

Name the gap between the tone the copy is reading as and the tone you want it to read as. “This reads too corporate” is a feeling. “This reads too corporate — I want it to sound like a founder explaining this to another founder, not a press release” is an instruction, because it gives the writer a reference register to aim for instead of just a direction to move away from.

The reason this works is the same reason vague corporate language fails broadly, not just in your specific case. A survey of workplace communication by the transcription company Notta, which analyzed over 5,000 Reddit comments about corporate jargon, found the most-hated phrases — “circle back,” “synergy,” “leverage,” “touch base” — weren’t disliked for being pretentious so much as for being vague. Notta’s COO summed up the finding directly: the jargon exposes “a disconnect between corporate language and effective communication,” and the underlying frustration is a “lack of actionable communication.” The same mechanism applies to feedback about jargon: “make this sound less corporate” is itself the kind of vague directive the study describes. Naming the specific register you want — founder-to-founder, plainspoken, direct — gives the writer something concrete to write toward.

There’s research backing the other side of that swap too. A study published in Personality and Social Psychology Bulletin, covered by CBS News, found that concrete, visualizable language reads as more credible than vague or abstract phrasing — readers process concrete statements faster and associate that speed with truthfulness. If the copy you’re flagging leans on abstract words like “optimize” or “streamline” instead of concrete ones like “cuts your review time from a week to a day,” that’s not just a style complaint. It’s a credibility gap you can name specifically: “this is too abstract — tell them what it actually does.”

How do you give feedback on clarity without just saying “this is confusing”?

Name the specific question a reader has at that point in the page that the copy doesn’t answer. “This is confusing” tells the writer that something failed without telling them what question went unanswered — they have to guess what a reader was trying to figure out and failed to. “A visitor reading this line doesn’t know if the $29.95 is monthly or a one-time setup fee” tells them exactly what to add.

This matters more than it feels like it should given how little of the page actually gets read. Tying back to the Nielsen Norman Group reading data above — if a visitor is only going to read a fraction of the page, the sentences they do read need to resolve the questions they’re actually asking in that moment, not just be grammatically fine. A sentence can be clear, correctly spelled, and still fail completely if it answers a question the reader isn’t asking while leaving the one they are asking untouched.

The fix is to reverse-engineer the confusion into a question. Read the flagged sentence as if you were seeing the page for the first time, and ask yourself what you’d want to know right there that the copy doesn’t tell you. Then write that question down as the comment instead of the vague verdict. “This is confusing” becomes “a visitor doesn’t know if they need a developer to set this up” — a sentence the writer can resolve in one pass instead of three rounds of “what did you mean by confusing?”

How do you flag copy that’s off-brand without listing every example yourself?

Point to a page that already sounds right and name the gap between that page and the one you’re reviewing. “The homepage hero sounds like us — confident, a little blunt, no fluff. This pricing page doesn’t; it reads like it was written by a different person” gives the writer two real anchors to reconcile instead of an abstract instruction to “match the brand voice,” which means something different to everyone who hears it.

This reference-point technique does the same job the Lucidpress research points to as the gap in most brand-consistency failures — most companies have brand guidelines, but far fewer have guidelines that are consistently enforced day to day. A style guide is abstract. A real page that already sounds right is concrete, and concrete references are easier to write toward than adjectives like “punchy” or “professional,” which five different people on your team would each interpret differently.

You don’t need to catalog every off-brand phrase on the page. Flag the two or three sentences that read most differently from the reference page, name the specific gap — more formal, more hedging, more passive — and let the writer apply that correction across the rest of the page themselves. They’ll catch more drift doing a full pass with your reference point in mind than you will doing a line-by-line audit.

How do you give feedback on a CTA or headline without prescribing the exact words?

State what the CTA or headline needs to accomplish, not the words it should use. “Change this to ‘Start Your Free Trial’” prescribes a specific sentence that might be wrong in a dozen small ways you haven’t considered — length, which benefit it leads with, how it reads against the button next to it. “A visitor should know this is free before they click” states the job the button has to do and lets the person who writes buttons for a living solve for it.

Button copy is a place where small, unclear wording compounds fast. Baymard Institute’s research on button design found that a meaningful share of ecommerce sites — the study puts it at 48% — get basic clarity elements of their primary buttons wrong, including vague or non-descriptive copy that doesn’t tell the visitor what will actually happen when they click. Their recommendation is specific: button text should communicate the result of the click, not a generic instruction. “Continue to Review Page” outperforms “Continue” for the same reason “see pricing” outperforms “learn more” — the visitor knows what they’re about to get.

That’s the outcome you should be naming in your feedback, not the sentence. If the current CTA doesn’t tell a visitor what happens next, say that: “this doesn’t tell someone what they’ll see if they click it.” The writer or developer can test two or three options against that bar faster than they can untangle a version you drafted that solves a slightly different problem than the one you meant to flag.

How do you pin copy feedback so your developer doesn’t have to guess which sentence you mean?

Click directly on the line you mean before you write anything, and quote the exact phrase inside your comment. A page usually has more than one sentence that could plausibly match a vague description like “the paragraph under the pricing table” — pinning removes that ambiguity the same way it does for a visual element, and quoting the phrase removes any doubt about exactly which words you’re talking about even if the pin shifts after a later edit.

This is where a website annotation tool earns its keep for copy specifically, not just for layout. The habit most teams default to — a screenshot in Slack with an arrow pointing vaguely at a block of text — works worse for copy than it does for visual bugs, because a screenshot can show you where a sentence is but not which exact words inside a paragraph someone means. Simpl_Markup vs. Screenshots in Slack covers why that gap shows up across feedback types; copy is one of the clearest cases, because “fix this paragraph” could mean one word or four sentences.

Simpl_Markup’s comments are plain text — no bold, no markdown, no rich formatting — which sounds like a strange limitation to mention in a post about copy feedback, but it’s actually the right constraint. The pin and the quoted phrase are doing the pointing; your comment doesn’t need formatting to be unambiguous, it needs to name the gap precisely, the same way it would for a button color. Click the paragraph, quote the phrase that’s off, state the tone or clarity gap, and the developer or writer sees exactly which words you mean in the Slack thread Simpl_Markup posts to, without a follow-up asking “which sentence?”

Unlike feedback on a visual element, copy feedback can’t rely on a pin alone to do all the disambiguating — a pin on a paragraph still leaves “which word in this paragraph” open in a way a pin on a button doesn’t. The quote inside the comment is what closes that last gap.

Put the technique together

Copy feedback breaks down into the same shape every time: notice the gap, name it precisely, and let the person who writes copy for a living close it.

Notice the gap — tone, clarity, brand voice, or missing answer. Name which one it is before you write the comment.
Name it concretely — “reads corporate, want founder-to-founder” beats “sounds off.” “Doesn’t say if it’s monthly” beats “confusing.”
Point to a reference, not a replacement — a page that already sounds right, or the specific question a reader needs answered — instead of a sentence you drafted yourself.
State the outcome for CTAs and headlines — what the line needs to accomplish, not the exact words it should use.
Pin it and quote it — click the line, quote the phrase, remove any question about which words you mean.

None of this requires you to be a copywriter. It requires the same discipline that works for visual feedback: describe the problem precisely enough that the person doing the fix doesn’t have to guess — and isn’t stuck reconciling your draft with theirs.

End the back-and-forth. Give feedback that lands — on the design and on the words.