How LinkedIn Ranks Comments Now.

LinkedIn has changed how comments are ordered under a post.

Instead of one fixed order that everyone sees, comments are now displayed by relevance to each reader, using signals such as professional interests, connections and previous engagement activity. LinkedIn is also surfacing more timely discussions in the feed to pull people into conversations rather than past them.

The update was reported on 9 August 2026 by Social Media Today, picking up creator economy analyst Lindsey Gamble. LinkedIn has not published a detailed newsroom post explaining it.

For a brand team the practical effect is short. The top comment is no longer a position you can occupy, because there is no single top comment any more. If the comment a reader sees is chosen for that reader, then a comment has to earn its place the same way a post does.

What LinkedIn has actually changed

Two things, and they work together.

The first is comment ranking. Comments under a post are ordered by how relevant they are judged to be to the person viewing them. The reported signals are:

• Professional interests
• Connections
• Previous engagement activity

The second is feed display. LinkedIn says it will surface more timely and relevant discussions in the feed, which means a live conversation under someone else’s post is now something the feed can show you, not just the post itself.

Together, the comment box stops being an appendix to a post and starts behaving like its own ranked surface.

What is not known yet

Coverage of this has been thin, and it is worth being precise about the gaps rather than filling them in.

• No stated rollout position. Nothing published says whether every member has this now, or whether it is arriving in stages by market or by surface.
• No detail on the model. Relevance covers a lot of ground, and LinkedIn has not said how heavily each signal weighs, or whether recency still holds a floor.
• No published measurement. There is no stated effect on how many comments get seen, or on how far down a comment can sit and still be shown to someone.

Treat the direction as certain and the mechanics as unclear. The direction is enough to act on.

Why LinkedIn is pushing on comments now

LinkedIn’s own numbers explain the priority. Following Microsoft’s earnings, LinkedIn chief executive Daniel Shapero said time spent in LinkedIn comments was up 18 percent year on year and content consumption up 10 percent, as reported by Social Media Today on 2 August 2026.

A surface people spend more time in is a surface worth curating. Once comment time is a growth line you report to the market, you have every reason to decide which comments each person sees, rather than letting arrival order decide it for you.

That raises the stakes for brands. The comment section was already a job most companies had never assigned, which is the case we made in who moderates your comment section now. Now it is an unassigned job with distribution attached to it.

The tactics this quietly retires

Several familiar LinkedIn habits depend on a fixed comment order. When order becomes personal, they stop working as designed.

• The seeded first comment, written by the poster to frame the discussion for everyone.
• The link parked below the post, on the theory that links in the body cost reach.
• The pod that piles in within the first ten minutes to hold the top slots.
• The one line reply left under a big account’s post to claim territory.

None of these break loudly. They simply stop being reliable, and reliability was the entire point of getting in first. A comment that only had position going for it now has nothing going for it.

The AI comment problem underneath this

The volume of low effort commenting is the other half of the story.

Detection research published by Pangram Labs in July 2026 put the share of fully AI written long form LinkedIn posts above 40 percent, and found comments slightly more likely to be AI written than top level posts once length is accounted for. Plenty of what sits under a LinkedIn post was not written by a person thinking about that post.

We would not assume LinkedIn is detecting and demoting AI text, and nothing published says it is. The problem is simpler than detection. A generic comment contains nothing specific for a relevance model to match a reader against.

That is the quiet cost of automating the reply. A comment written to fill a slot reads as filler to a person and scores as filler to the relevance model, which are the same failure arriving from two directions.

What a comment now has to do to get surfaced

If relevance is judged per reader, a comment competes on the qualities a post competes on. In practice:

1. It says something the post did not. Agreement adds no information, so it gives the relevance model nothing to work with.
2. It comes from an account with a real relationship to that audience, through connections, a shared field or a history of engaging there.
3. It is specific enough to be about something. A number, a counter example, a constraint you actually hit.
4. It gives someone a reason to reply, because replies are engagement activity and engagement activity is one of the stated signals.

That is a higher bar than most brand commenting is written to. It is also a bar a real practitioner clears without effort and a template never clears at all.

What good looks like

A brand that treats comments as distribution runs them like publishing, not like admin.

• One named owner per platform, with comment work inside their week rather than squeezed around it.
• A short target list. Ten to twenty accounts whose audience you genuinely want, reviewed monthly, instead of commenting wherever the feed lands.
• A written standard for what leaves the account. No agreement without addition, no compliment openers, no link drops.
• A response window on your own posts, because your replies under your own comments now have their own chance of being surfaced.
• A measurement line. Profile visits and follows you can trace to commenting, reviewed alongside the rest of the reporting.

None of that is exotic. It is ordinary social media management discipline applied to a surface most teams have left to whoever has a spare minute.

How NBK thinks about the comment section

We treat the comment section as a distribution channel with an owner, a standard and a slot in the week. Not a courtesy, and not a growth trick.

That framing matters more after a change like this one. When order was fixed, comment work could be won by speed and volume, so it was rational to hand it to whoever was fastest. When order is earned per reader, the only thing that scales is judgement, and judgement has to be somebody’s named job.

It is also worth saying plainly that this is organic distribution. Nobody can sell you a top comment. The way to be seen under a post is to write something worth showing, which is the same principle behind why personal brands need a LinkedIn content system.

Next step

Pull the last month of comments your brand left on LinkedIn and ask one question of each. Would a stranger with no connection to us have found this useful?

If most of them fail that test, the fix is a standard and an owner, not a higher comment count.

If your social output feels busy but not effective, an NBK audit is a sensible place to start.