LinkedIn Replaced Its Entire Algorithm — and Most Creators Haven’t Adjusted

LinkedIn quietly replaced its entire content ranking infrastructure in late 2024 and early 2025. The old system used roughly a thousand separate models to score and distribute content. The replacement is a single 150-billion-parameter large language model called 360Brew, trained exclusively on LinkedIn’s professional data: job titles, skills, industry context, engagement patterns, and post history.

By Q1 2026, 360Brew handles the majority of content scoring decisions across the platform. The result is a fundamental shift in how content reaches people. Views are down 50 percent year over year. Follower growth has declined 59 percent. But average engagement is up 18 percent — meaning the people who do see your content are more likely to interact with it.

Most creators saw the view drop and assumed LinkedIn was broken. What actually happened is that the platform stopped distributing content broadly and started distributing it precisely. If your content matches the interest graph of a specific professional audience, you reach exactly those people. If it does not, you reach almost nobody. The margin for generic content collapsed.

The most important structural change is what LinkedIn calls the shift from a relationship graph to an interest graph. Under the old system, your content primarily reached people who already followed you or were in your first-degree network. Your connections were your distribution.

Under 360Brew, content can reach anyone whose professional interests match what you wrote, regardless of whether they follow you. Only 31 percent of a user’s feed now comes from first-degree connections. Another 25 percent comes from second and third-degree connections. And 10 percent comes from people the algorithm identifies as interested strangers — users with no network connection to you at all.

This is a structural advantage for creators with clear expertise and a structural disadvantage for creators who relied on a large network to guarantee reach. A startup founder with 800 connections writing about a specific technical problem can now reach the same audience as a thought leadership account with 50,000 followers — if the content is more relevant to that audience’s interest graph.

The format hierarchy on LinkedIn looks nothing like it did a year ago. Document posts — PDF carousels — are now the highest-performing content format on the platform by a significant margin.

The single most important engagement metric on LinkedIn in 2026 is the save. When someone bookmarks your post, it drives 5 times more reach than a like and 2 times more reach than a comment. Posts that receive saves and substantive comments within 24 to 72 hours perform 4 to 6 times better than posts that only collect likes.

This changes what kind of content creators should optimize for. Likes reward emotional resonance — content that makes someone feel seen or validated. Saves reward reference value — content that someone wants to come back to because it contains specific, useful information they cannot easily find elsewhere.

The practical implication is that content designed to be bookmarked outperforms content designed to be liked. Frameworks, step-by-step processes, benchmark data, decision checklists, and comparison tables all drive saves. Hot takes and motivational quotes do not.

LinkedIn’s new ranking signal, called Depth Score, measures how long people actually engage with your content — not just whether they clicked or scrolled past. A post that a user reads for 30 seconds outperforms a post that collects 50 quick likes from people who never stopped scrolling.

This is why document posts perform so well. A 10-slide carousel forces the reader to swipe through multiple pages, generating dwell time with every slide. A text post that someone skims in two seconds generates almost no dwell signal, regardless of whether they hit the like button.

The first 60 minutes after publishing are critical. Only 5 percent of posts that underperform in the first hour recover to broader distribution. This means your initial audience — the people who see your post first and spend time on it — determines whether 360Brew distributes the post more widely. If your first readers skim and scroll past, the post is functionally dead.

LinkedIn now assigns every creator a topic fingerprint — an internal credibility score for how strongly you are associated with a specific subject area. This Topic Authority score is built from three inputs: topic consistency across your posts, engagement quality from people in the relevant professional community, and what LinkedIn calls semantic clarity — how cleanly your content classifies into a recognizable expertise area.

The critical finding: it takes approximately 90 days of consistent, aligned posting for 360Brew to fully categorize your expertise. During that window, the algorithm is watching whether your content stays on-topic or scatters across unrelated subjects. Creators who post about one niche consistently for 90 days get substantially more distribution than creators who cover multiple topics in the same period.

360Brew also audits your profile against your content. Your headline, summary, and work history need to align with the topics you post about. If your profile says marketing consultant but your posts are about cryptocurrency, the algorithm registers a mismatch and deprioritizes distribution.

Several tactics that worked on LinkedIn as recently as early 2025 now actively hurt performance under 360Brew.

The creators who are growing on LinkedIn in 2026 are running a different playbook than the one that worked two years ago. Here is what the data supports.

LinkedIn’s 360Brew is not a tweak to the old system. It is a replacement. The platform moved from distributing content based on who you know to distributing content based on what you know. That is a fundamentally different proposition for creators.

The view count drop scared many creators off the platform or into posting more aggressively. Both responses are wrong. The correct response is to post less frequently, with more depth, in a tighter niche, using the formats that generate dwell time and saves. The algorithm is no longer rewarding the loudest voice in the room. It is rewarding the most useful one.

For creators who have genuine expertise in a specific professional domain, this is the most favorable version of LinkedIn that has ever existed. The interest graph does not care about your follower count. It cares about whether your content is the best answer to what a specific professional audience is looking for right now.

Launchvibes approaches platform strategy by identifying where a creator’s expertise and audience engagement intersect, then mapping effort to the platform mechanics that convert that expertise into growth. LinkedIn’s 360Brew shift makes this more precise than ever: the algorithm is now built to find and amplify exactly the kind of focused, authoritative content that most creators should be producing but are not.