Every year LinkedIn's feed changes a little, and every year a wave of outdated advice keeps circulating anyway. Post at 8am. Comment on ten posts before writing your own. Use exactly three hashtags. Some of it was never true. Some of it used to work and does not anymore.
This guide separates three things: what LinkedIn itself has said, what independent studies have measured, and what is only rumor. We link every number to its source.
How does the LinkedIn algorithm work in 2026?
LinkedIn's feed tries to show each member posts they are likely to find useful and conversations they are likely to join. It has never published its ranking formula, so anyone who claims to know the exact weights is estimating. What we do know comes from two places: LinkedIn's public statements and outside studies of large numbers of posts.
Taken together, they point one way. Reach now follows conversation. A post that starts a real exchange travels further than one that collects quick reactions, and activity that looks manufactured is held back.
What has LinkedIn officially said?
LinkedIn has been unusually direct about what it does not want. In March 2026 it published What We're Doing to Support Authentic Content and Conversations on LinkedIn. The key lines:
"We actively limit the reach of inauthentic activity, including engagement pods and automated comments, by using technology and human review to help detect suspicious behavior, sometimes reducing reach, and taking action when policies are violated."
On engagement pods, which are groups that agree to like and comment on each other's posts, it says this activity "is not allowed on LinkedIn". Its Professional Community Policies say the same in plainer words: "don't agree with others ahead of time to like or re-share each other's content."
On automated comments, the March 2026 post defines them as comments posted "using a browser extension, script, or third party tool" that "don't include human action or clicking the 'comment' button", and says they "aren't allowed on LinkedIn". When it detects them, LinkedIn says it may leave them out of the "Most Relevant" comments, and may restrict the member's ability to use LinkedIn.
LinkedIn's VP of Product Management, Gyanda Sachdeva, put the goal bluntly in late 2025, as reported by Social Media Today: "Our goal is to make engagement pods entirely ineffective."
What does this mean for tools that comment for you, including Allen?
It means you should know how any tool you use posts, including ours. Allen has two modes. With Autopilot on, Allen posts the day's comments for you through LinkedIn's official API after emailing you the plan. With Autopilot off, Allen finds the posts and writes the comments, and nothing goes out until you post each one yourself.
LinkedIn's statement is about comments posted by a tool with no human clicking comment. We are not going to tell you how LinkedIn will treat any tool, ours included. LinkedIn writes and enforces its own rules. If you want every comment to go out by your own hand, leave Autopilot off. Allen's help center covers this in is Allen safe for my LinkedIn account?
Whichever mode you use, the content rule is the same. A comment has to be specific to the post and worth reading, or it helps nobody.
Which signals matter most, according to independent studies?
Outside studies agree on the order, even where they disagree on the size. Remember these are estimates from observed posts, not LinkedIn's own figures.
| Signal | What studies report | Source |
|---|---|---|
| Comments | One widely cited analysis scores a comment at 12 points against 1 for a like, with longer comments counting for more | Richard van der Blom's Algorithm Insights report, as summarized by Limber |
| Saves | A save drives about five times the reach of a like and about twice that of a comment, in an analysis of more than 3 million posts | Hootsuite, citing AuthoredUp |
| Author replies | Replying to comments on your own post is linked to about 30% higher engagement, across 72,000 posts | Buffer |
| Early activity | The first 90 minutes after publishing are described as the critical window | Limber, citing the same report |
| Likes | The weakest of the common signals in every analysis | All of the above |
The pattern is consistent. Signals that take effort and show real interest count for more than signals that take one tap.
Is engagement on LinkedIn going up or down?
Both, depending on what you count. Metricool's 2026 LinkedIn study looked at 673,658 posts from 63,108 accounts and found likes down 13%, comments down 17% and shares down 10% year over year, while overall engagement rose nearly 14%. Visible reactions are falling while quieter actions, such as clicks, are rising.
For anyone commenting, that is useful. Fewer comments under each post means a good one stands out more. It is easier to be the comment an author remembers when you are one of eight than one of eighty.
Do personal profiles beat company pages?
Yes, by a clear margin. The same Metricool study found personal profiles get a 63% higher engagement rate on average than company pages. People respond to people.
This is why commenting as yourself works better than commenting as a brand, and why Allen works from your personal profile only. If you want your company to be seen, the practical route is one or two real people at the company showing up in their own names.
Do comments on other people's posts help your reach?
Commenting on other people's posts helps in a different way from posting. It does not push your own posts up the feed. It puts your name and your thinking in front of the author and everyone reading that thread, most of whom do not follow you.
That is the mechanism behind comment-led growth. Your own posts reach your existing network. Your comments reach someone else's. We explain the trade in why LinkedIn comments matter more than posts.
It only works if the comment earns its place. Buffer's finding about author replies cuts both ways: authors who reply to comments do better, so authors have a reason to reply to a comment that gives them something to answer.
Which LinkedIn algorithm advice is outdated or unproven?
| Claim you may have heard | What we can verify |
|---|---|
| "Generic comments are penalized by a comment quality score" | LinkedIn has announced no such score. Its public statements target automated comments and pods, not short ones. Generic comments fail for a simpler reason: nobody replies to them. |
| "Comments under 10 words are ignored" | Not stated by LinkedIn. One third-party report suggests longer comments carry more weight. Treat it as a hint, not a rule. |
| "You must reply in the first 60 minutes" | The only window we could trace to a named report is 90 minutes, and it is a third-party estimate. Replying the same day is the safe habit. |
| "Pods still work if the group is small" | LinkedIn says pods are not allowed and that it is increasing the ways it detects them. |
| "Post at exactly 8am" | Best times vary by audience. No fixed hour is supported by LinkedIn. |
| "Hashtags drive reach" | We found no current, reliable evidence either way, so we make no claim. |
Where we could not verify something, we have said so. That is also our advice for reading any algorithm guide, including this one.
What should you do differently?
- Comment less and better. Five specific comments beat fifty that could sit under any post. See the anatomy of a LinkedIn comment that gets you noticed.
- Stay in one niche. Comment on the same kind of person about the same kind of problem, so your name becomes familiar to one group.
- Reply to every comment on your own posts, the same day.
- Leave pods. The upside is shrinking and the downside is a restricted account.
- Show up on a schedule you can keep. Buffer found that posting two to five times a week brings about 1,182 more impressions per post than posting once a week, in a study of more than 2 million posts. Regular beats occasional.
- Comment as a person, not a page.
- Know how your tools post. If a tool acts on your account, find out whether a person presses the button.
How does Allen fit with how the feed works now?
Allen is built around the parts of this that take time. It finds fresh posts from the past week written by people who match your audience by job title, so your comments stay in one niche. It writes each comment in your voice from examples you wrote, and it refuses to write on spam, empty or offensive posts. With Autopilot it spaces comments a few minutes apart and caps the day at 5, 10, 15 or 20.
When someone replies, it emails you and drafts the direct message, because replies are where the value is. You can run it with Autopilot on, or keep Autopilot off and post every comment yourself. The details are on the how it works page, and every plan is free for 7 days.