Posted in LinkedIn · 8 min read
The anatomy of a LinkedIn comment that adds one real point
Great post! gets ignored. A single specific observation gets a reply from the author — and gets seen by everyone else in the thread. Here are the five moves that reliably earn one.
Farhad
In short
A LinkedIn comment earns a reply when it gives the author something they can respond to — a specific example, a boundary on their claim, or a question only they can answer — and agreement alone gives them nothing, which is why praise comments sit unanswered. Five moves cover almost every good comment: the counter-example, the boundary, the mechanism, the specific question, and the extension. The author's reply is what matters more than the comment itself, because it is what puts the thread back in front of their audience with your name attached.
Key takeaways
- Agreement is a dead end. An author cannot reply to praise with anything except thanks.
- The unit that works is one point, made specifically. Two points in a comment halves the chance of a reply.
- Comment early, while the author is still in the thread. A perfect comment on a two-day-old post is invisible.
- Naming the case where the post is wrong is the highest-response move, and the one people are most afraid of.
- Length is not the variable. A two-line comment with one real observation beats a paragraph of agreement.
A post does well on LinkedIn — four hundred reactions, sixty comments. Read the comments. Roughly forty are some variant of agreement: great post, so true, thanks for sharing, this resonates. Fifteen restate the post in slightly different words. Five say something.
The author replied to four of the five.
That ratio is stable across almost any post you check, and it is not about who is well-known or who writes well. It is mechanical: an author cannot reply to agreement with anything except thanks, and thanks is a conversation-ender. Give them something they can respond to and they will, because replying is genuinely more interesting for them too.
The author's reply is the thing worth optimising for, and not for their sake. A reply pushes the thread back into the feed of everyone who engaged with the post, with your comment attached and your name on it. One good comment on someone else's well-read post reaches more relevant people than most of what you will post yourself this month.
This is a narrower problem than not sounding salesy in a Facebook group. In a group the risk is reading as an advert. On LinkedIn nobody minds that you are a business — everyone is. The risk is being filler, and filler is invisible rather than unwelcome.
Why does agreement fail?
Because it hands the author a dead end.
Consider what they can actually type in response to "great post, so true". Thanks. That is the whole option set. They are not being rude by not replying; there is nothing to reply with.
Now consider "this matches what I have seen, except in [specific situation], where the opposite tends to happen". Now they have something. They can agree with the exception, argue with it, or ask what you meant. All three are replies, and all three are more interesting than thanks.
The mechanism is the same one that makes a good question work in conversation. You are not adding value in the abstract; you are giving the other person a next move. A comment with no next move in it is a comment that ends.
The question is not whether your comment is good. It is whether the author can reply to it with something other than thanks.
What are the five moves?
Almost every comment that earns a reply is one of five things.
| Move | What it does | When it works best | Risk |
|---|---|---|---|
| Counter-example | Names the case where the post's claim breaks | Confident, broadly-stated posts | Reads as contrarian if you have no specific |
| Boundary | Agrees, then marks where it stops being true | Advice posts, frameworks | Low. The most reliable of the five |
| Mechanism | Explains why the thing they observed happens | Posts describing a pattern without a cause | Being wrong in public |
| Specific question | Asks the one thing only they can answer | Story posts, results posts | Sounds like an interview if too open |
| Extension | Takes their point one step further with your own data | Posts you genuinely agree with | Becomes a mini-post if you overdo it |
The boundary is the workhorse. It is low-risk, it is almost always true — most LinkedIn advice is correct within a range and stated without one — and it is genuinely useful to everyone reading, because the range is the part the post left out.
The counter-example gets the most replies and it is the one people avoid, because disagreeing in public feels risky. It is only risky when done without a specific. "I disagree" is an opinion and invites a defence. "That held for us until we crossed about thirty people, and then it inverted" is an observation and invites a question.
What do the five look like written out?
Adapt the specifics — the shape is the reusable part.
Counter-example
This holds up in [context] — we saw the same thing.
Where it broke for us was [specific situation]: [what happened instead]. My guess
is [one-line theory], but I am not certain that generalises.
The hedge on the last line is load-bearing. It converts a contradiction into a shared question, which is what makes it comfortable to reply to.
Boundary
Agree, with one boundary: this works while [condition]. Past that, [what changes],
and the same approach starts producing [the opposite outcome].
Mechanism
The reason this works is probably [mechanism] — [one sentence on why]. Which
would also explain why [related thing people notice but do not connect].
Specific question
Curious about [the specific decision point in their story]. Did you [option A], or
did [option B] turn out to be the thing that mattered?
The specificity is the entire difference. "What was the biggest lesson?" is an interview question and gets an interview answer. "Did you cut the second product line before or after the revenue recovered?" gets a real one, because it can only be answered by someone who was there.
Extension
Adding one thing to this: [your own specific — a number, a case, a constraint].
Same pattern, [slightly different context], which suggests it is [broader
observation] rather than [narrower one].
What is the one-point rule?
One point per comment. Not two, and not one point with a supporting point.
This is the most common failure among people who are good at this. You read a post, three genuine observations occur to you, and all three go in. The comment becomes a paragraph, the author now has to choose which part to respond to, and the friction of choosing is enough that many will not.
It also changes how the comment reads. One observation reads as a person reacting. Three reads as a person performing expertise, which everyone on LinkedIn recognises instantly.
If you have three points, use the best one. The other two are a post of your own, and quoting their post in yours is a better outcome for both of you than burying two-thirds of your thinking under someone else's.
Why does timing matter more than quality?
Because the audience for your comment is created by the author's reply, and the author is only in the thread for a few hours.
Post goes up. The author is watching, engagement is being distributed, and comments are visible. Within the first hour or two your comment is likely to be read by the author and by people still arriving at the post. A comment on that same post two days later is read by nobody, including usually the author.
The practical version of this: it is better to leave a decent comment on a post from an hour ago than a perfect one on a post from Tuesday. Which means the routine matters more than the craft — being present daily beats being brilliant occasionally, and that is the entire argument for the ten-minute habit.
A decent comment in the first hour beats a perfect comment on day two, every time. The thread is the audience, and the thread is short-lived.
What does a bad comment look like when it is trying?
Four patterns, all of them written by people making an effort.
The restatement. Summarising the post back with slightly different words. It reads as engagement and adds nothing, and the author can tell instantly, because they wrote it.
The unearned pivot. Two lines of agreement followed by a hard turn into what you do. Everyone recognises the pivot, including the people you hoped would not.
The essay. Two hundred words of genuinely good thinking in someone else's comment section. It was a post, and putting it here means fewer people read it and the author feels slightly crowded.
The fluent nothing. This one has become much more common and is now the fastest way to be scrolled past. Well-structured, correctly punctuated, three balanced clauses, no specific observation anywhere in it. LinkedIn readers have recalibrated hard, and a comment with no fact that could only have come from experience now reads as generated regardless of who wrote it. The fix is not shorter or more casual — it is one concrete thing. A number, a case, a constraint you actually hit.
What should you deliberately not automate here?
The observation. Everything else in this article — the structure, the hedge, the length, the register — is mechanical. The specific thing you have seen that the post did not account for is the comment, and it cannot come from anywhere except your own experience.
Which posts to comment on. Choosing the thread where you genuinely have something is most of the skill, and it is the decision that a volume mindset destroys first.
And the reply to the reply, when the exchange has become genuinely interesting. At that point you are having a conversation with someone in public and it should read like one.
What does Reply Pilots do here, and what does it not?
Reply Pilots drafts in the LinkedIn comment box from the post and the comments already visible on your screen — which is what lets it write the comment that adds to the thread rather than repeating what six people have already said. It applies your voice settings and your banned-words list, so the fluent-nothing tells are gone before you see the draft, and per-platform rules keep LinkedIn length and formality distinct from how you would reply on X. The thread summary tool is useful on long posts where the discussion has moved on from the post itself.
What it does not do: it cannot supply the observation. It has your business context, not your experience, so it will produce a well-shaped comment with a general point in it — and a general point is exactly what gets scrolled past on LinkedIn. The working pattern is to let it handle the structure and add the specific yourself, which takes about ten seconds and is the entire difference between a comment that gets a reply and one that does not. It also does not post, and it does not comment on a schedule.
Your next step
Open LinkedIn, filter to posts from the last two hours, and find one where you know a case the author did not mention. Write the boundary version: agree, then mark where it stops being true, with a specific.
Then watch whether the author replies. That reply — not the comment — is the thing you are building, and once you have seen it happen twice the rest is repetition.
Related reading
- Ten good comments a day, in ten minutes — the routine that makes timing possible
- Comment or DM: where your next client actually comes from — what to do when the author replies
- Writing in your own voice: a 10-minute setup — removing the fluent-nothing tells
See how Reply Pilots works for the product this article is about, end to end.
Frequently asked questions
How long should a LinkedIn comment be?
Two to four sentences for almost everything. Long enough to make one point with a specific in it, short enough that it does not read as an attempt to write a post inside someone else's thread. If you have genuinely got several hundred words to say, that is a post of your own that quotes theirs.
Does commenting early actually matter?
A great deal, for two reasons. The author is still in the thread and likely to reply, and their reply is what puts the conversation back in front of their audience. A comment left two days later is usually read by nobody except the author, and often not by them.
Is it risky to disagree with someone in their own comments?
Disagreeing with the claim is fine and is the highest-response move there is. Disagreeing with the person is not. The distinction is whether you are naming a case where the pattern breaks or implying they have not thought about it — the first invites a reply, the second invites a defence.
Should I comment on posts from people I do not know?
Yes, and it is most of where the value is. Commenting on your own network mostly reaches people who already know you. A comment on a post outside your network is how the people who do not know you come across your name for the first time.
How many comments a day is reasonable on LinkedIn?
Five to ten good ones, consistently, is better than thirty in an afternoon. The audience overlap between posts in one industry is high, so the same people see you repeatedly, and a burst reads differently from a habit — one looks like presence, the other looks like a campaign.
Do hashtags or tagging people in comments help?
Hashtags in comments do nothing useful and read as marketing. Tagging is worth it only when the person genuinely has the answer and would want to be pulled in. Tagging to get attention is noticed by everyone, including the person tagged.
What should I do if the author replies?
Answer the specific thing they raised, once, and then stop. The value has already been created — their reply put the thread in front of their audience with your name on it. A third and fourth comment turns a good exchange into a conversation other people are scrolling past.
Can AI draft LinkedIn comments that do not sound like AI?
It can draft the structure reliably, and the tells are removable with a banned-words list — but the specific observation that makes the comment work has to come from you, because it is drawn from something you have actually seen. A draft that reads well and contains no specific is the exact comment LinkedIn readers have learned to scroll past.
Related articles
A Simple ROI Calculation for Comment & DM Specialists
This role's ROI math is different from a general time-saved calculation. Here's the version that actually fits how this role gets measured.
Read article →
A Simple ROI Calculation for DM-to-Book Agencies
This agency's real ROI number isn't time. It's booked calls. Here's how to calculate it that way.
Read article →
A Simple ROI Calculation for Freelance Social Managers
You don't need a spreadsheet model to know if this is worth it. Here's the two-number calculation that settles it.
Read article →