Watch the video: Why Your LinkedIn Posts Reach the Wrong People (8 minutes, YouTube). This article covers the same ground in writing, with every source linked.
Ask ten LinkedIn "experts" how the algorithm works and you'll get ten confident answers. Post at 8am. Never use links. Use three hashtags. Get 50 comments in the first hour or your post is dead. Some of these claims are old, some are guesses, and some are flatly contradicted by what LinkedIn has published.
The good news is that LinkedIn has published a lot about its feed in 2026. Its engineers have described the retrieval and ranking systems in detail, its Pressroom has spelt out what it limits, and on 9 November 2026 it published a new post on how it recommends content from people you aren't connected to. So rather than add another opinion, we took ten of the most repeated claims and checked each one against LinkedIn's own words. Every source below was read on 10 November 2026.
What has LinkedIn actually published about its feed?
These are the primary sources this guide relies on. If a claim isn't supported by one of them, we say so.
- Engineering the next generation of LinkedIn's Feed, LinkedIn Engineering Blog, 12 March 2026 (LinkedIn Engineering). The most detailed description of retrieval and ranking.
- How LinkedIn Is Improving the Feed, LinkedIn Pressroom, 12 March 2026 (LinkedIn Pressroom). The plain-English version, including engagement bait.
- What We're Doing to Support Authentic Content and Conversations, LinkedIn Pressroom, 13 March 2026 (LinkedIn Pressroom). Pods and automated comments.
- How LinkedIn is Continuing to Tackle AI Slop, LinkedIn Pressroom, 15 September 2026 (LinkedIn Pressroom), and Tim Jurka's 8 October 2026 article on the technology behind it (LinkedIn).
- Profile Relevance: Boosting Feed Relevance with LLMs at LinkedIn, LinkedIn Engineering Blog, 9 November 2026 (LinkedIn Engineering). How posts reach people outside your network.
- How Does the LinkedIn Feed Work?, Tim Jurka, 11 August 2025 (LinkedIn). Jurka writes that he has spent over a decade working on the Feed.
How does the LinkedIn algorithm work, in plain English?
Think of it as a funnel with a quality filter running through it. LinkedIn says it ranks "millions of posts at any given time" while balancing "content from your network, who you follow, and the LinkedIn Economic Graph more broadly (i.e. suggested content)" (LinkedIn Engineering).
- Your post becomes text a model can read. For each post, LinkedIn includes "format, author information (name, headline, company, industry), engagement counts, article metadata, and post text" (LinkedIn Engineering).
- Retrieval builds a shortlist. LinkedIn replaced several separate sources with "a single, unified retrieval system powered by LLM-generated embeddings", which finds top candidates "in sub-50ms" (LinkedIn Engineering).
- Ranking picks what each member sees. LinkedIn's Generative Recommender "processes more than a thousand of your historical interactions" and predicts both passive actions (click, skip, long-dwell) and active ones (like, comment, share) (LinkedIn Engineering).
- Quality systems limit what travels. LinkedIn says it has "built backend content classifiers to spot and limit distribution of AI slop" (LinkedIn Pressroom) and that it limits the reach of pods and automated comments (LinkedIn Pressroom).
The three-label test we used
Every claim below gets one of three labels. It's a simple habit worth borrowing whenever someone tells you what "the algorithm" wants.
- Documented: LinkedIn has said this, and we link to where.
- Contradicted: LinkedIn has said something that conflicts with the claim.
- Not documented: we couldn't find it in LinkedIn's published sources. It might still be true, but it's a hypothesis to test, not a rule.
10 popular LinkedIn algorithm claims, checked
1. "LinkedIn only shows your posts to your connections." Contradicted.
LinkedIn's retrieval system explicitly balances your network, who you follow and suggested content (LinkedIn Engineering). Tim Jurka writes that "exceptional content may even be distributed broadly across LinkedIn to members who are interested in the type of content you post, even if they don't follow you" (LinkedIn). And the 9 November post describes Profile Relevance, a system built specifically to recommend posts from outside a member's network (LinkedIn Engineering).
One nuance: members can switch their feed to Recent, which LinkedIn describes as updates "in reverse chronological order from people you follow, without personalization" (LinkedIn Help). LinkedIn is also testing a dedicated Network feed of posts from connections and people you follow (LinkedIn Pressroom). So your network still matters. It just isn't the only route.
2. "LinkedIn matches posts to people by keywords and hashtags." Contradicted for keywords; hashtags not documented.
LinkedIn's own example is a member interested in "electrical engineering" who engages with posts about "small modular reactors". Keyword systems "might miss the connection", it says, while its language model understands the topics are related (LinkedIn Engineering). Matching works on meaning, not exact words. We found nothing in these sources on hashtags either way, so any "use exactly three hashtags" rule is not a LinkedIn statement.
3. "Your profile has nothing to do with how your posts perform." Contradicted.
Your name, headline, company and industry travel with every post into the model's prompt (LinkedIn Engineering). Jurka says the system "factors in the author's professional identity, such as experience, content shared on their profile, and skills" (LinkedIn). Profile Relevance also uses members' headlines, summaries, skills and work history to match them with posts, and appends member and author country names so people see more local professional content (LinkedIn Engineering).
4. "The first hour decides everything." Not documented.
We found no fixed "golden hour" in LinkedIn's sources. What is documented is speed: when a post is published or gains engagement, LinkedIn updates its prompt, and "the entire process happens within minutes" (LinkedIn Engineering). Profile Relevance adds one specific rule: its recommended posts "must have received a single viral action (like, comment, or share) from an unconnected member" first (LinkedIn Engineering). Early genuine engagement clearly helps. A hard 60-minute window is folklore until LinkedIn says otherwise.
5. "Engagement pods and comment swaps boost reach." Contradicted.
LinkedIn calls an engagement pod "a coordinated group of people that like, comment and share posts to attempt to boost visibility and virality" and says it "is not allowed on LinkedIn". It says its teams work on "flagging posts that look artificially boosted, and limiting their reach in the feed", and that detected automated comments may not show in Most Relevant comments (LinkedIn Pressroom). Its Help Centre adds that it doesn't permit "any tools or services that try to manipulate LinkedIn's content algorithms" (LinkedIn Help). We go deeper in are LinkedIn engagement pods allowed?.
6. "'Comment YES and I'll send you the guide' helps the algorithm." Contradicted.
In March, LinkedIn said it is reducing "repetitive, low-substance posts and engagement bait, such as 'comment to agree' prompts or videos that don't match the text", adding: "We want the Feed to be relevant to your interests, not a popularity contest" (LinkedIn Pressroom).
7. "LinkedIn penalises anything written with AI." Contradicted as stated.
LinkedIn's Help Centre says: "LinkedIn's focus is not on how content is created, but whether it adds value", and that content which "feels generic, repetitive, or lacks a clear point of view is less likely to be widely distributed" (LinkedIn Help). The tool isn't the test. Substance is.
That test has teeth. LinkedIn says views of content classified as AI slop fell by 40% (LinkedIn Pressroom), and Jurka wrote in October that its classifiers now cover "all posts that are distributed beyond your immediate network with 94% precision at detecting AI-slop" (LinkedIn). Our guide to LinkedIn's AI slop check covers how to pass it.
8. "Time spent reading your post matters." Documented.
LinkedIn's ranking model predicts passive actions including skips and long dwells, alongside likes, comments and shares (LinkedIn Engineering). This isn't new: a 2020 engineering post described adding a "P(skip)" objective that predicts the probability a member will skip an update (LinkedIn Engineering). A post that rewards reading is doing real work, even if nobody comments.
9. "Your age, gender or race affects your reach." Contradicted.
LinkedIn says its ranking model relies on "professional signals and engagement patterns, never demographic attributes" (LinkedIn Engineering). Jurka says the system "does not factor in any of the author's demographic information, such as age, race, or gender" (LinkedIn), and the Profile Relevance prompt excludes "personal attributes such as the member's name or gender" (LinkedIn Engineering). LinkedIn also says the model is "regularly audited for equitable treatment" (LinkedIn Engineering). The stated design excludes demographics.
10. "Posts with external links get buried." Not documented.
This may be the most repeated claim of all, and we couldn't find it in any of the LinkedIn sources above. That doesn't prove it's false. It means it's an observation or an outside study, not a LinkedIn rule. If links matter to your business, test it on your own account: alternate posts with the link in the body and without, on the same kind of topic, and compare them over several weeks.
The verdicts at a glance
- Documented: meaning-based retrieval, a ranking model reading 1,000+ past interactions, dwell and skip predictions, reach beyond your network, profile information read alongside posts, reduced reach for AI slop, bait, pods and automated comments.
- Contradicted: connections-only reach, keyword matching, profile irrelevance, pods and bait as growth tactics, a blanket AI penalty, demographic ranking.
- Not documented: a golden hour, hashtag counts, a link penalty. Treat these as hypotheses.
A worked example: one post, rewritten with the documented signals
Here's a hypothetical illustration, not a client. A founder runs a payroll consultancy for multi-country businesses. Their headline says "Helping businesses grow". Their draft post reads, in summary: five generic tips on "scaling your team", eight hashtags, and a closing line asking people to comment PAYROLL to receive a checklist.
Here's how we'd rework it, one documented signal at a time.
- Profile first. Change the headline to say who they help and with what, for example payroll for companies hiring across several countries. LinkedIn reads the headline alongside every post.
- One clear topic. Replace five generic tips with one specific buyer problem, such as what goes wrong in the first month of running payroll in a new country. Meaning-based retrieval works best when a post is plainly about one thing.
- Earn the dwell. Walk through a real sequence of decisions, with the trade-offs, so a finance lead actually reads to the end. Skips and long dwells are both modelled.
- Drop the bait. Remove "comment PAYROLL". Offer the checklist with a plain line instead, or ask a genuine question a finance lead would want to answer.
- Keep it human. Use the founder's own example and opinion. LinkedIn's quality test is whether a post has a clear point of view, not which tool helped write it.
None of these changes is a trick. Each one lines up with something LinkedIn has documented, which is exactly why they're worth doing.
What this means for founders who want sales calls, not just reach
Our view, not LinkedIn's: the algorithm in 2026 rewards the same things buyers reward. Write for one buyer problem at a time. Make your profile explain who you help. Say something only you could say. Never manufacture engagement. And do what Jurka suggests, which is to "look at their posting patterns by topic, format, and timing" to see how each post finds its audience (LinkedIn).
Then judge your posts on buyer signals rather than reach. Our LinkedIn content ROI scorecard shows what to track, and our seven-step system for turning LinkedIn posts into sales calls covers the conversations that follow. If you're wondering about cadence, we looked at what LinkedIn and a large study say in how often a founder should post on LinkedIn, and for the wider numbers see LinkedIn statistics for B2B founders.
How Influx works with the feed, not against it
Influx is built around the documented signals, not the folklore. We capture your expertise in one hour of onboarding, our team writes in your voice, nothing goes live without your approval, and each post is distributed to the buyers you want to reach with no ad spend and no pods (The Influx Authority Engine).
We measure ourselves on outcomes. We guarantee 30 qualified sales calls in 90 days. If we don't hit it, we keep working at no cost until we do. We only guarantee what we control: the right buyers on your calendar. Closing is yours. (The Influx Guarantee Explained)
About Influx
Alex Alleyne is the founder of Influx, a two-time LinkedIn Top Voice with over 60,000 followers on LinkedIn. Our founder, Alex Alleyne, was named a LinkedIn Top Voice in Sales in 2019 and 2020. Influx is a done-for-you LinkedIn authority system for founder-led B2B firms that turns a founder's expertise into qualified sales calls, with content written in their voice and distributed to their buyers. Influx, the LinkedIn authority firm, is operated by Fetched L.L.C-FZ in Dubai, UAE, and is not affiliated with other companies called Influx.
Want LinkedIn content that works with the feed and brings in qualified sales calls? Book a call.

