What did LinkedIn actually change?

It stopped showing every reader the same list. Social Media Today reported in August that LinkedIn now orders the comments under a post by relevance to whoever opens it, weighing signals like professional interests, connections and engagement activity. 2 buyers can open the identical post at the identical moment and land on 2 different top comments, because the ranking runs off the reader.

You already ran into a version of this once. Narrowing your content to 1 real buyer moved your signal from likes to profile views and DMs, for the same reason. The reader decides what counts as relevant. You don’t get a vote.

The old system rewarded speed. Comment inside the first few minutes of a fresh post and you had a real shot at the top slot for everyone who scrolled past after you. The new one cares whether that specific reader’s interests line up with what you wrote.

Why did LinkedIn make this change now?

Because a third of what looked like conversation was AI. An analysis from AI detection company Pangram Labs, reported by Social Media Today, looked at 57,000 public LinkedIn posts and found that 30% of the comments posted between April and June this year were entirely AI-generated. That’s a THIRD. When you can spot AI like I can, it’s cringingly bad. It’s literally AI just having conversations with AI, and I’d dare to say it’s more than 30% on any given post. 

Comment time on LinkedIn was already up 18% year over year by the platform’s own numbers, and overall content consumption was up 10%. Buyers are reading comments more than they used to. A third of what they were reading came from AI.

LinkedIn has gone after volume dressed up as engagement before. The same enforcement logic showed up in outreach automation earlier this year. The comment section just caught up finally. I haven’t always been the biggest fan of LinkedIn’s moderation, but they have my full support on this one 

The signals LinkedIn named, interest, connection and engagement history, are exactly what a template comment or a fully automated one struggles to fake at scale. That’s the quiet part of a ranking change dressed up as a UX tweak.

What actually earns a comment a spot now?

Specificity to the reader beats speed to the post.

Comment on the actual point in the post. Don’t comment just because the poster happens to be on your target list. If your comment could sit under any post in the feed, word for word, the ranking has nothing distinctive to match it against. Commenting “congrats” or “great idea,” or anything that’s less than, give or take, 18 ± insightful words is a complete waste of your time… and the author’s.

Don’t be lazy… Or do what we do and develop an 800 to 1,000-insight Claude insights bank so commenting becomes a breeze. Copy and paste the post into your commenting Claude project with about 4 pages’ worth of things to avoid saying, things to say, and speaking references, most importantly so it captures your tone. 

The point is removing the brain fart so you can get your knowledge out faster and have your comment actually mean something. You will always, now and forever, need to put your personal touch on each one. Done right, your comments can literally hit 50,000 to 70,000 impressions depending on the success of the original post. In some cases your comments can actually outperform the post it’s on from an impressions perspective.

Freshness still matters, but the reason changed. Sort your search results by latest instead of top when you’re hunting for something to comment on. A fresh post can put your comment in front of roughly 150 people, while the identical comment on a post from 3 weeks ago reaches around 4.

AI can help you write it. What can’t survive the new ranking is AI writing the whole thing with nothing of yours in it, aka AI slop. A human still has to write and post the comment, even when a tool pulls raw material from something you’ve actually said elsewhere. That’s a difference in architecture rather than degree, and it’s what the ranking is now built to notice.

Skip it when you’ve got nothing real to say. Forcing a comment on pure sales content, or anything outside your lane, reads as generated whether a person or a tool wrote it, and it marks your account the same way. The more visibility you’ve already built, the more that one bad comment costs you.

Moral of the story: Don’t be a tool just for the sake of forcing a sale. If you need that touchpoint that badly, just leave a like.

3 moves before your next comment

• Base every comment on the specific point a post makes rather than on whether the poster’s on your target list

• Sort your search results by latest instead of top, so your comment still has time to rank before the post goes stale

• Skip the comment when you’ve got nothing genuine to add. A forced one costs your own account, more than it costs that one thread

Steve Lewis
CEO
TORC Executive Marketing

Ready to turn a comment into an actual conversation? 

Sign up for our FREE 7-day tech founder LinkedIn personal brand positioning and lead generation course

Frequently asked questions

If you can’t control who LinkedIn ranks your comment for, is commenting before the DM still worth doing?
Yes. The old system never guaranteed visibility either. It only guaranteed a fair shot if you showed up early. A comment that actually matches your buyer’s own signals can now outrank 100 generic ones competing for the same slot. You can still use AI commenting tools like Extrovert to source your comments faster by keywords, Sales Nav lists, etc.

Does commenting on the same buyer’s posts repeatedly improve your odds under the new ranking?
It should, since engagement activity is 1 of the 3 signals LinkedIn named. Commenting on 1 person’s posts more than once builds exactly the kind of history the ranking reads for that specific relationship. That’s different from commenting more often across your whole target list, which does nothing for relevance and reads as automated fast. Depth with 1 reader beats spread across 100.

Should a tech founder avoid commenting on posts that already have a lot of comments, since the good spots seem taken?
No longer, and that’s actually the upside of this change. Under the old chronological order, a busy post buried you fast no matter how good your comment was, because every viewer saw the same fixed list. Under per-viewer ranking, your comment can sit at the top for the specific buyer whose interests and connections line up with it, regardless of how many comments came before yours. Busy only hurts you when your comment has nothing distinctive to say.