How LinkedIn Decides Who Sees Your Post: The…

One day your LinkedIn post gets 200 impressions; the next one hits 8,000

Same topic, same format, same timing, totally different reach and it feels random. In 2026, it isn"t random: LinkedIn runs your post through a fast series of distribution tests, scoring relevance, early engagement quality, retention signals, and trust. If you pass the first filter, reach expands; if not, distribution stalls quietly. We analyzed 10,222 LinkedIn posts from 494 creators to show what consistently gets amplified (and what gets throttled), so you can build repeatable reach instead of chasing spikes.

Phase 1: The Test (First 15-60 Minutes)

Every post starts life in the same place: a small test.

When you hit publish, LinkedIn does not show your post to all your connections and followers. It shows it to a small subset, typically 5-10% of your network. This is your test audience. They're the jury. Your post's entire future depends on how they respond.

The test audience isn't randomly selected. LinkedIn picks the connections and followers most likely to engage with your content based on several factors:

Relationship strength. People who regularly interact with your content (like, comment, view your profile) are prioritized. If someone consistently engages with your posts, they're almost guaranteed to be in your test audience.

Topic affinity. If your post is about marketing and a connection frequently engages with marketing content (from anyone, not just you), they're more likely to see your test post. LinkedIn reads your content and matches it against the interest profile of potential audience members.

Recent activity. People who are actively on LinkedIn at the time you publish are weighted higher. Someone who checks LinkedIn twice a day is more likely to be in your test audience than someone who logs in once a week.

Connection type. First-degree connections who accepted your request (or vice versa) are weighted higher than followers who simply clicked "Follow." The mutual connection signal is stronger than a one-way follow.

This selection matters enormously because the test audience determines everything. A great post shown to the wrong test audience will flop. A good post shown to a highly relevant test audience will succeed. The quality of your network directly affects the quality of your test audience, which directly affects your distribution ceiling.

Pro tip: This is why accepting every connection request from strangers can backfire. Those connections become part of your potential test audience pool. If they have no interest in your content topics, they'll scroll past your posts when they appear in the test, generating negative dwell time signals. Be strategic about your network. The people you connect with aren't just contacts. They're your distribution infrastructure.

What Happens During the Test

The algorithm watches several signals during the test phase:

Dwell time. How long test audience members spend looking at your post. This is the primary quality signal in 2026. If people stop scrolling and read, the post is working. If they scroll past in under 2 seconds, it isn't.

"See more" click rate. What percentage of people who see the truncated preview choose to expand the post. A high click rate tells LinkedIn the hook is working. A low click rate tells LinkedIn the preview wasn't compelling enough.

Engagement velocity. How quickly the first likes and comments arrive. A post that gets 5 likes in 10 minutes is performing better than one that gets 5 likes in 60 minutes. Speed matters because it signals genuine interest rather than courtesy engagement that trickles in over time.

Engagement quality. Not all engagement is equal. A thoughtful comment carries roughly 8x the weight of a like. A share carries roughly 4x. A save (bookmark) is increasingly weighted as well. The algorithm isn't just counting interactions. It's evaluating their depth.

Negative signals. The algorithm also watches for "hide post," "report" and "unfollow" actions. Even one "hide" from the test audience is a strong negative signal. These actions are rare (most people simply scroll past rather than actively hiding content), so when they happen, the algorithm takes them seriously.

The test phase lasts roughly 15-60 minutes depending on the size of your network and the velocity of initial engagement. For smaller accounts (under 5,000 connections), the window is closer to 30-60 minutes. For larger accounts, the algorithm can make distribution decisions faster because it has more data points.

Pro tip: The first hour after publishing is the most important hour for your post. This is not the time to publish and disappear. Stay online. Reply to every comment. The replies serve double duty: each one counts as additional engagement on the post AND it keeps the conversation going, which attracts more participants. A creator who replies to 10 comments in the first hour has 10 additional engagement signals that a creator who's offline doesn't have.

Phase 2: Quality Scoring (Hours 1-8)

If your test audience responded well, the post moves to Phase 2: quality scoring. This is where LinkedIn decides how wide to distribute.

The algorithm doesn't just check "did people engage?" It evaluates the quality of that engagement against benchmarks. Your post is compared to:

Your own historical performance. If your posts typically get 0.7% engagement rate and this post hit 1.2% in the test phase, the algorithm sees it as above your baseline. It gets wider distribution. If it hit 0.4%, it's below baseline and distribution expansion slows.

Peer performance. Your post is compared to other posts about similar topics published at similar times. If your marketing post generated better test-phase metrics than the average marketing post published that morning, it gets a boost. If it performed worse, it gets scaled back.

Content-type norms. A carousel is compared to other carousels. A text post to other text posts. This normalization prevents format bias. A text post doesn't need to generate carousel-level dwell time to score well. It needs to generate strong dwell time for a text post.

Based on this scoring, LinkedIn assigns your post a distribution tier:

Low tier: Connection-only distribution. Your post stays visible primarily to first-degree connections who are active on the platform. Reach is typically 200-1,000 impressions. This is where most posts end up. It's not a penalty. It's the default.

Mid tier: Extended network distribution. Your post starts appearing in the feeds of second-degree connections (friends of your connections) who share relevant interest profiles. Reach typically expands to 1,000-10,000 impressions. About 20-30% of posts reach this tier.

High tier: Topic feed and viral distribution. Your post enters LinkedIn's topic-based feeds (similar to hashtag feeds but algorithmically curated) and can appear in the feeds of people with no connection to you at all. This is where posts reach 10,000-100,000+ impressions. Only about 5-8% of posts reach this tier.

Viral tier: Platform-wide distribution. The top 2.16% of posts in our dataset. These break through all normal distribution limits and can reach hundreds of thousands of impressions. At this tier, the content essentially markets itself through shares and comments that create cascading distribution.

Pro tip: You don't control which tier your post reaches. But you control the inputs that determine it. Specifically: hook quality (drives "see more" clicks), content quality (drives dwell time), opinion strength (drives comments) and format choice (drives baseline dwell time expectations). Get three out of four right and you're consistently in the mid tier. Get all four right and you're knocking on the door of high tier.

Phase 3: Sustained Distribution (Hours 8-72)

Most people think a post's life is over after the first few hours. In reality, Phase 3 is where the best posts separate from the merely good ones.

After the initial quality scoring, LinkedIn continues to evaluate your post's performance in real time. If the post keeps generating engagement (new comments, shares, reactions) beyond the first 8 hours, the algorithm extends its distribution window. If engagement tapers off, distribution slows and eventually stops.

This is why some posts seem to "come back to life" after a quiet period. A post that gets moderate engagement in the first 8 hours, then receives a burst of comments on day 2 (perhaps because someone influential shared it), can trigger a new round of distribution expansion. LinkedIn's algorithm is not a one-time decision. It's a continuous evaluation.

The sustained distribution phase favors certain content characteristics:

Evergreen topics. Posts about timeless topics (career advice, industry fundamentals, professional development) can generate engagement for days because the content remains relevant as new people discover it. A hot take about a news event might burn bright in Phase 1 but die in Phase 3 because the news cycle moved on.

High comment-to-like ratio. Posts with active comment sections (where the creator and commenters are having real conversations) sustain distribution longer because each new comment is a fresh engagement signal. This is why your own replies to comments matter so much. They're not just polite. They're keeping the algorithmic engine running.

Share-driven distribution. When someone shares your post, it enters their network's feed as a new piece of content. That shared version goes through its own mini test-audience phase. If that test audience engages, the share generates its own distribution wave. This cascading effect is what turns high-tier posts into viral-tier posts. You can't force shares, but you can create content that people want to associate with by sharing.

In our data, posts that reached the high tier averaged 2.3 days of active distribution. Posts that went viral averaged 4.1 days. The sustained-distribution window is directly correlated with ongoing engagement velocity.

Pro tip: Check your best-performing posts at the 24 and 48-hour marks. If comments are still coming in, reply to every single one. You're extending the post's active distribution window with each reply. Some creators have added 30-50% more total reach by actively engaging in comments during Phase 3 rather than moving on to their next post.

The Interest Graph: Why Your Content Reaches People You've Never Met

The interest graph is LinkedIn's internal model of what every user is interested in, built from their engagement behavior. It's the engine behind "recommended" content in your feed and it's becoming more important every year.

How it works:

Every action you take on LinkedIn contributes to your interest profile. The posts you like, the articles you read, the profiles you visit, the topics you comment on, the newsletters you subscribe to, the groups you join, the hashtags you've followed (even though hashtag following is less relevant now). All of this data builds a multi-dimensional map of your professional interests.

When the algorithm distributes a post beyond first-degree connections, it uses the interest graph to find the right recipients. It doesn't push your marketing post to random LinkedIn users. It pushes it to users whose interest profiles indicate affinity for marketing content, your specific subtopic within marketing and related topics.

This is why a post about "how to run A/B tests on cold email subject lines" might reach someone you've never met who works in a completely different industry. If that person regularly engages with A/B testing content and cold email content, the interest graph matches them to your post. The connection isn't social. It's topical.

The interest graph also explains why some topics seem to "perform better" than others. In our data, Software Engineering posts have a 2.57% engagement rate, the highest of any category. Social Media Marketing hits 1.34% with 210 average comments. These categories perform well not because the algorithm likes them but because the interest graph accurately matches this content to highly engaged niche audiences. The people who receive these posts through interest-graph matching are predisposed to engage because the topic aligns with their professional identity.

Contrast this with Personal Development posts: 1,222 average likes (the highest) but only 0.39% engagement rate. The interest graph serves these posts to a broad audience because many people's interest profiles include "personal development." But the engagement per impression is low because the audience is less targeted. Personal development content appeals broadly and shallowly. Technical content appeals narrowly and deeply. The algorithm rewards depth.

Pro tip: You can influence your interest-graph positioning by being consistent about topics. If you post about B2B sales consistently for three months, LinkedIn builds a strong interest-graph profile for your account. People interested in B2B sales start seeing your content more reliably. If you jump between sales, leadership, AI, personal stories and industry news with no pattern, the interest graph can't categorize you. The algorithm doesn't know who to show your posts to because your signal is scattered. Pick 2-3 topics and commit.

The Signals That Matter Most (Ranked)

Based on our analysis of 10,222 posts and observable algorithm behavior, here's our best estimate of signal importance in the 2026 algorithm, ranked:

1. Dwell time. The primary quality signal. How long people spend reading your post drives more algorithmic decisions than any other single factor. Estimated at 35-40% of the overall quality score.

2. Comment quality and quantity. Comments are weighted approximately 8x more than likes. Longer, substantive comments carry more weight than generic one-liners. Estimated at 25-30% of the quality score.

3. "See more" click rate. The percentage of people who expand your truncated post. This is effectively a hook quality score. Estimated at 10-15% of the quality score.

4. Shares. Each share creates a new distribution wave. Shares are rare (most posts get zero shares) but extremely powerful when they happen. Estimated at 8-10% of the quality score.

5. Likes/reactions. Still count but carry less individual weight than any other positive signal. Their main value is in volume: a large number of likes contributes to overall engagement velocity during the test phase. Estimated at 5-8% of the quality score.

6. Saves/bookmarks. LinkedIn has been increasing the weight of saves over the past year. A save signals that someone found the content valuable enough to return to later. This is a strong quality indicator. Estimated at 3-5% and growing.

7. Negative signals. Hides, reports and unfollows after seeing a post. These are rare but heavily penalized. One "hide" might offset several likes in algorithmic terms. Estimated at negative 5-10% impact when triggered.

Pro tip: The practical implication of this ranking: spend 80% of your optimization effort on dwell time and comments. Those two signals account for the majority of the quality score. A post that generates 30 seconds of average dwell time and 15 substantive comments will outperform a post with 200 likes but 5-second average dwell time and 3 generic comments. The like count looks better. The algorithmic score is worse.

Common Misconceptions About the Algorithm

"The algorithm is suppressing my reach to sell ads." No. LinkedIn's business model depends on a high-quality feed that keeps professionals engaged. Suppressing good organic content would degrade the feed, reduce time-on-platform and make the ad inventory less valuable. The algorithm is a quality filter, not a revenue tool. Posts with low engagement get less distribution because they're low quality, not because LinkedIn wants your ad spend.

"Posting more frequently beats the algorithm." Publishing more posts doesn't guarantee more total reach. In fact, our data shows that posting more than once per day can cannibalize your own distribution. Your test audiences overlap between posts, and if post #2 goes up while post #1 is still in Phase 2, they compete for the same attention. Quality over quantity wins every time.

"The algorithm penalizes certain topics." The algorithm doesn't have a topic blacklist. Some topics generate lower engagement because the audience isn't engaged by them, not because the algorithm is suppressing them. A post about supply chain logistics might get less engagement than a post about career advice, but that's because fewer people care about logistics, not because LinkedIn is anti-logistics.

The exception is content that triggers LinkedIn's automated moderation system (political content, certain sensitive topics). These can see reduced distribution. But this is moderation, not algorithmic preference. There's an important distinction.

"LinkedIn shows your post to more people if you use their new features." There's a persistent belief that LinkedIn boosts content from creators who use new features (LinkedIn Live, newsletters, collaborative articles) as an incentive for feature adoption. While this may have been true during initial feature rollouts, we see no evidence of systematic feature-usage bonuses in 2026. The algorithm evaluates content quality, not feature adoption.

"Engagement in the first 10 minutes decides everything." Close but not quite. The first hour matters. The first 10 minutes contribute, but the algorithm needs a sufficient sample size before making distribution decisions. A post that gets 3 likes in 10 minutes isn't necessarily outperforming one that gets 2 likes in 10 minutes. The sample is too small. The algorithm waits for statistically meaningful signals, which typically takes 15-60 minutes.

Pro tip: The biggest misconception of all is that the algorithm is something you "beat." You don't beat it. You work with it. The algorithm wants to surface content that keeps professionals on the platform. Your job is to create content that does exactly that. When your interests align with the algorithm's interests, you get distribution. It's a partnership, not an adversarial game.

How the Algorithm Treats Different Post Types

The algorithm evaluates content within its format category. Here's how each format interacts with the three-phase system:

Text posts. Evaluated primarily on dwell time, "see more" click rate and comment quality. The algorithm expects moderate dwell time (8-20 seconds for a medium-length post). Text posts that generate comments significantly above the text-post baseline get disproportionate distribution boosts because the algorithm interprets this as "the ideas are strong enough to overcome the less engaging format."

Image posts. Evaluated on scroll-stop rate, dwell time (including image viewing time) and engagement rate. The algorithm expects higher baseline engagement from image posts (0.93% in our data versus 0.50% for text). A mediocre image post that hits 0.60% is actually underperforming relative to its format, even though it would be above average for text.

Carousel/document posts. Evaluated on swipe-through rate (what percentage of viewers go beyond slide 1), total slides viewed, dwell time per slide and engagement after completion. The algorithm has the highest engagement expectations for carousels. A 10-slide carousel where most viewers only see 3 slides is underperforming. One where viewers average 7-8 slides is performing well.

Video posts. Evaluated on watch time, completion rate and post-video engagement. LinkedIn's algorithm treats video as a dwell time powerhouse but penalizes videos with high drop-off rates (people clicking play, watching 5 seconds, then scrolling away). Short videos (under 90 seconds) with high completion rates tend to perform best.

Polls. The algorithm has minimal engagement expectations for polls. In our data, polls average 0.07% engagement rate. Even a "good" poll barely registers in algorithmic terms. The format generates negligible dwell time (tap an option, move on) and minimal comment activity. The algorithm correctly categorizes this as low-value content.

Newsletters. Technically bypass the feed algorithm entirely for subscribers (delivered via notification). But the newsletter notification itself creates engagement that feeds back into the algorithm's model of your account quality. Active newsletter publishers tend to have stronger overall algorithmic standing because the newsletter engagement contributes to LinkedIn's assessment of the account.

Pro tip: Match your content strategy to format strengths. If your goal is comments and conversation, text posts with strong opinions work well because the format encourages reading and responding. If your goal is maximum reach, carousels with educational content generate the highest dwell time and widest distribution. If your goal is building an owned audience, newsletters bypass the algorithm entirely. There's no "best" format. There's the best format for your specific goal on a specific post.

What This All Means for Your Content Strategy

The algorithm is a distribution machine. It takes your content, tests it with a sample audience, evaluates the response and decides how many more people should see it. The entire system runs on reader behavior. Not your behavior. Not your format choices. Not your posting time. Reader behavior.

Everything you do as a creator is ultimately about influencing that reader behavior:

Your hook determines whether they stop scrolling (dwell time starts). Your first paragraph determines whether they click "see more" (distribution potential unlocks). Your body content determines how long they stay (dwell time accumulates). Your closing determines whether they comment (comment quality signal fires). Your replies determine whether the conversation continues (sustained distribution extends).

Each stage feeds the next. A weak hook kills everything downstream. A strong hook with weak body content generates a high "see more" click rate but low dwell time, which the algorithm reads as "misleading preview." You need the full chain to work.

In our data, the top 10% of posts by engagement all share three characteristics: a scroll-stopping hook, body content that holds attention for 25+ seconds and a closing that prompts substantive comments. The specific topic doesn't matter as much as people think. The format matters less than people think. The three-part attention chain is what separates the 8,000-impression post from the 200-impression one.

The algorithm isn't trying to stop you. It's trying to help its users find content worth reading. Give it something worth distributing and it will do the distribution for you.

That's the entire game. Everything else is details.

Data sourced from ViralBrain's analysis of 10,222 LinkedIn posts across 494 creators. ViralBrain turns algorithm understanding into actionable insights so you can create posts that work with the distribution system, not against it.