Since July 27, 2026, LinkedIn posts that ask people to "comment a keyword" to get something (like "Comment GUIDE and I'll send it to you") have been getting about half the reach they used to. That's according to a study of 684,000 posts. Before that date, the same kind of post did slightly better than average. A plain question at the end of a post wasn't affected.
LinkedIn hasn't announced this change. But it fits with what it did say in March 2026 about cleaning up the feed. Here's what the data shows, what LinkedIn has actually said, and what to do instead.
The numbers come from a study by MagicPost, a LinkedIn tool company. It looked at 684,000 posts from 16,000 creators between June 2025 and September 2026, and compared each post with that same creator's average post (MagicPost).
| Type of post ending | Before July 27 | After July 27 |
|---|---|---|
| "Comment KEYWORD" to get something | 9% above average | 52% below average |
| Other requests to comment or repost | Around average | About 20% below average |
| A plain closing question | 4% below average | 4% below average (no change) |
| Link in the first comment | 11% below average | 11% below average (no change) |
A few more details from the study:
Keep in mind: this is one study, from a company that sells a LinkedIn tool. It measures impressions only, and it can't see inside LinkedIn's ranking system, so it doesn't claim to know why it happened. Still, it's a big sample, and the drop is large and sudden.
LinkedIn hasn't said anything specific about "comment a keyword" posts. But in March 2026, it published two updates about the feed:
A post that asks for a one-word comment to unlock a freebie is designed to get comments, not to start a conversation. It makes sense that a feed trying to understand what posts are "actually about" would stop rewarding that.
The same company's bigger study looked at more than just endings (MagicPost). Since July 27:
In plain words: posts that say something specific, from something you actually did or saw, are doing better. Generic advice is doing worse.
Many AI writing tools end posts with "Comment YES if you agree" or "Drop a đ„ below" by default. Check every AI draft for these lines and delete them.
It's also a reason to give AI tools your own material, like a voice note, a video, or rough notes with real details, instead of just a topic. Posts with real detail are the ones getting more reach, and AI can't make up your experience for you (and shouldn't try). That's the idea behind tools like Reshaper AI, which write from your own videos, voice notes, or rough notes rather than a blank topic.
No. They aren't banned, and LinkedIn hasn't announced anything about them specifically. But data from 684,000 posts shows they've been getting about half their usual reach since July 27, 2026.
Yes. In the same study, posts ending with a plain question performed the same before and after July 27, at about 4% below the creator's average post.
Links still cost some reach. Link cards and links in the post text get fewer impressions than posts without links. A link in the first comment cost about 11% and didn't get worse after July 27.
Posts with concrete details and personal experience. Since July 27, those got about 15% more impressions than the creator's average, while posts with neither got about 23% fewer.
No. LinkedIn said in March 2026 that engagement pods are not allowed and that it has removed groups showing signs of pod behavior.
Transformez un seul contenu en posts pour chaque plateforme en quelques secondes.