LinkedIn's 360Brew: What Actually Changed (And What It Means for You)

LinkedIn's algorithm just went through its biggest change in history.

Everyone noticed something was off. Reach dropping by half. Content that used to work suddenly hitting a wall. The panicked posts asking "Is anyone else's LinkedIn broken?"

What they didn't know was why.

Here's what happened: LinkedIn replaced thousands of specialized ranking models with one massive AI system called 360Brew. It's a 150-billion parameter language model, the same technology that powers ChatGPT, but trained specifically on professional network data.

This isn't a tweak. It's a complete rewrite of how LinkedIn decides what you see.

AuthoredUp analyzed over 3 million posts to figure out what changed. Median reach dropped 47% year-over-year. Video fell off a cliff (down 72%). Even reliable text posts took a 34% hit.

But here's the interesting part: some creators are still growing. Fast.

The difference isn't luck. It's understanding how the new system actually works.

What 360Brew Actually Is (Without the Jargon)

Think of LinkedIn's old algorithm like a factory with 1,000 specialized machines. Each one had a specific job: one ranked job posts, another handled connection suggestions, another decided what showed up in your feed. Engineers spent years hand-tuning each machine.

360Brew is different. It's one adaptive system that handles everything, including feed, jobs, search, people recommendations, and ads, by reading and understanding text like a human would.

The technical term is "decoder-only transformer." LinkedIn's research team published the details in January 2025. But here's what matters for you:

The old system tracked clicks and connections. If a user interacts with posts on a specific topic, show them more content on that same topic.

The new system reads meaning. It understands that "Gong's revenue intelligence platform" and "Salesforce CRM integration" are related concepts, even if you never used those exact terms. It can infer your professional interests from the language in your profile, not just from what you clicked last week.

This changes everything.

How Your Feed Works Now

Your feed still uses a two-stage process, but what happens in stage two is fundamentally different.

Stage 1 narrows millions of posts down to a few hundred candidates based on your network, interests, and past behavior. This part hasn't changed much.

Stage 2 is where 360Brew takes over. It reads each candidate post, actually reads the text, and scores how relevant it is to you specifically.

Here's what it's analyzing:

Then it does something clever: it creates a temporary "personalized version" of the model just for you by analyzing your last 2-3 months of activity. The research calls this "many-shot in-context learning", but what it means is: the algorithm learns what YOU specifically consider valuable, then applies that understanding.

Example: You regularly engage with posts about B2B SaaS metrics, revenue operations, and CRM architecture. 360Brew notices this pattern and starts surfacing more content about ARR growth, RevOps workflows, and Salesforce integrations, even from people you've never heard of, because it understands these topics are semantically related to your interests.

The old system couldn't make those connections without manual feature engineering. This one just... reads.

What Signals Actually Matter Now

AuthoredUp's research analyzing 3+ million posts found something striking: one save drives 5x more reach than a like, and 2x more than a comment.

Think about what a save actually signals. Someone found your content valuable enough that they want to come back to it. That's a much stronger signal than a quick double-tap.

Here's what moves the needle:

Saves are the strongest signal by far. If people are bookmarking your content, LinkedIn interprets that as "this has lasting value."

In-depth comments matter way more than quick reactions. A paragraph-long response that adds perspective or asks a substantive question tells the algorithm this content sparked real thinking.

Comment threads are gold. When people start replying to each other in your comments, not just to you, that's indirect engagement, and it's valuable. It means you created a discussion, not just a broadcast.

Reposts signal that someone found your content valuable enough to share with their entire network. That's a strong endorsement.

Dwell time matters. Are people actually reading, or scrolling past after two seconds?

Delayed engagement is a quality signal. Posts that get saves and meaningful comments 24-72 hours after publishing perform 4-6× better in suggested feeds, according to AuthoredUp data.

What matters less than you think:

• Total like count alone
• Speed to 100 reactions
• One-word comments
• Engagement pods (the algorithm can spot these patterns now)

The Profile-Content Alignment Requirement

Here's something most creators miss: LinkedIn now checks whether your content matches your stated expertise before distributing it.

The research confirms this. The model reads your profile (your headline, About section, experience, skills) and uses that to understand what you're an expert in. Then it checks whether your posts actually align with that positioning.

If your profile says "Enterprise SaaS Sales VP" but you're posting about crypto trading, meditation, travel photography, and AI art, LinkedIn can't classify what you're actually expert in. So your content doesn't get matched to relevant audiences.

This isn't about being boring or one-dimensional. It's about being clear.

What works:

Example of good alignment:

Headline: "RevOps Director | B2B SaaS Go-to-Market Strategy | HubSpot + Salesforce Integration"
Posts about: Sales operations workflows, CRM data architecture, B2B funnel optimization, SaaS metrics
Occasional posts about: Team building, leadership lessons learned, industry events

Example of poor alignment:

Same headline
Posts about: Meditation practices, travel photography, political hot takes, fitness routines, crypto trading, occasional sales content
The algorithm literally reads your profile and posts as text. Make the connection obvious.

‍

It takes about 90 days of consistent, aligned posting for the system to fully categorize you and start optimizing distribution. Be patient.

💡 Pro tips

In AuthoredUp in both the Drafts and Posts tabs, you can tag every draft by topic or campaign. That simple habit keeps your content strategy focused even when you’re managing several personal or company profiles.

What Actually Happened to Reach (The Reality)

Let's talk numbers, because they're brutal.

AuthoredUp's analysis tracked median impressions from June 2024 to May 2025:

June 2024: 1,211 median impressions per post
May 2025: 636 median impressions per post

That's a 47% drop in one year.

Every format got hit:

• Video: -72%
• Documents: -43%
• Images: -45%
• Text: -34%

This isn't seasonal. It's structural. More creators are publishing, LinkedIn is stricter about relevance, and the bar for "good enough" just got a lot higher.

If your reach is down, you're not broken. You're experiencing the same reality as everyone else.

The creators who are still growing adapted faster. They figured out the new signals and adjusted.

Is 360Brew Actually Live?

Short answer: we don't know exactly, and it doesn't matter.

LinkedIn published the research in January 2025. The paper describes a "pre-production model V1.0" developed over 9 months. They haven't announced deployment dates, percentages, or which surfaces are using it.

But creators started reporting visibility changes in summer 2024. That's LinkedIn's typical pattern: quiet, gradual A/B testing without announcement. As analyst Petya Savova notes, "LinkedIn rarely announces algorithmic changes in advance and almost always rolls them out progressively through testing waves."

Current educated guess: 40-100% deployed across different surfaces.

Why this uncertainty doesn't matter:

Even at 40% deployment, nearly half your audience sees content ranked this way. 

And optimizing for 360Brew principles as semantic clarity, topical consistency and quality engagement, improves performance under both the old and new systems.

The direction is clear. This is LinkedIn's future. Adapt now or keep sliding.

What to Actually Do About It

Forget growth hacks. Forget optimal posting times. Forget magic formats. Those worked on the old ID-based system. This one reads meaning.

Here's what matters:

1. Fix Your Profile First

Your headline needs to clearly state your niche. Not clever wordplay. Not inspirational fluff. Just: what you're expert in and how you help.

Bad: "Marketing Maven | Thought Leader | Coffee Addict ☕"‍

Good: "B2B SaaS Marketing VP | ABM Strategy & Revenue Operations | Scaling $5M→$50M ARR"

Your About section should name your 2-3 core topics in the first paragraph. Make it obvious what you'll be posting about.

Your top skills should match those topics.

This isn't optional. The algorithm reads this to understand what to classify you as.

2. Lead With Specificity

Research on LLM position bias shows AI models prioritize information at the start. Your first 1-2 sentences get 3-5x more processing attention than the end.

Open with the specific topic and concrete entities:

Weak: "I've been thinking about sales tools lately..."‍

Strong: "Gong's conversation intelligence helped us identify why 67% of enterprise deals stalled in legal review. The pattern surprised our entire RevOps team."

Name specific companies, products, frameworks, people. Use industry-standard terminology. Make your meaning clear.

3. Optimize for Saves, Not Just Likes

Ask yourself before publishing: would someone want to come back to this next week?

Content that earns saves:

• Step-by-step frameworks with real examples
• Data breakdowns with non-obvious insights
• Annotated templates or checklists
• Contrarian perspectives backed by evidence

One save is worth five likes in terms of reach. Structure your content accordingly.

4. Encourage Real Discussion

You don't just want comments. You want:

• Paragraph-length responses that add perspective
• Questions that show someone actually read it
• Discussions between commenters (not just replies to you)

End with specific, answerable discussion prompts. Not "What do you think?" but "Which part of your CRM integration causes the most friction?"

When people start discussing with each other in your comments, let it flow. That indirect engagement is valuable.

5. Stay in Your Lane for 90 Days

Pick your 2-3 core topics. Post 80% of your content within those areas for three months straight.

This isn't about being one-dimensional. It's about being classifiable. The algorithm needs to understand what you're expert in before it can match you with relevant audiences.

After 90 days, you'll see:

• Engagement from outside your immediate network
• Higher save-to-like ratios
• More followers gained from posts
• Profile views correlating with your topics

The Metrics That Actually Matter

LinkedIn recently added advanced analytics showing saves, followers gained per post, and profile views driven by posts. These are exactly the signals 360Brew prioritizes.

The catch is that LinkedIn only shows these for posts up to one year old.

This is where AuthoredUp becomes essential. The platform tracks these metrics historically, beyond LinkedIn's one-year limit. You can see which topics consistently drive saves, which formats generate quality followers, what content builds lasting authority.

AuthoredUp participated in Richard van der Blom's research analyzing 600,000+ posts, so the platform is built specifically for algorithm optimization.

Track these monthly:

• Saves per post (trending up?)
• Followers from posts (discovery working?)
• Profile views from posts (interest matching successful?)
• Comment depth (substantive or surface?)
• Discussion threads (indirect engagement happening?)

Impressions still matter for context, but they're a lagging indicator. The quality signals predict future performance.

💡 Pro tip 

Inside AuthoredUp’s Posts tab, you can explore every post you’ve ever published and its advanced metrics, such as saves, reposts, and comment count. Filter, compare results, and uncover which topics consistently drive real intent signals.

Having your full historical data in one place helps you spot patterns, so you can focus on creating content that keeps performing over time.

Common Myths to Stop Believing

"Video is dead because of 360Brew"

Video reach is down 72%, but that's user behavior (low completion rates), not algorithmic penalty. The research explicitly states the model is "format-agnostic." Video underperforms when people don't watch it through.

"You have to post daily"

Frequency isn't a primary signal. Topic consistency over 90 days matters more than posting schedule. One excellent post per week beats seven mediocre ones.

"Put links in the first comment to avoid penalties"

This is backwards. First comments with links get deboosted. If you need a link, put it in the main post, remove the preview card, and accept 15-20% lower reach. Even better is to minimize external links entirely.

"Write everything in the text when posting video"

No. People consume EITHER text OR video, not both. If your copy says everything, why would they watch? Keep video post copy to context/setup only.

What This Really Means

LinkedIn's 360Brew killed the era of algorithm hacks.

You can't game a system that reads and understands meaning. The tactics that worked on ID-based ranking such as optimal post times, magic formats, hashtag stuffing, don't translate.

What wins now:

• Clarity over cleverness
• Depth over frequency
• Specificity over generalization
• Consistency over virality
• Real expertise over borrowed credibility

For the first time, LinkedIn can actually understand what you're expert in and connect you with relevant audiences, if you communicate clearly.

AuthoredUp's data analyzing 3M+ posts confirms it: quality and clarity now beat quantity and manipulation.

If you're a CMO managing your company's LinkedIn presence: focus on topical expertise, not vanity metrics. Establish clear positioning for your executives, create save-worthy thought leadership, and track quality engagement over raw impressions.

If you're a consultant building your personal brand: define your niche explicitly, stay in your lane for 90 days, optimize for saves and discussion, and be patient with the timeline. The algorithm rewards sustained expertise signals, not one-off viral hits.

The game changed. The new rules reward substance.

Adapt or keep sliding.

Wrapping Up

360 Brew changed how visibility works on LinkedIn, but it didn’t change what good content looks like. It simply rewards it better.

The system now values clarity, relevance, and long-term consistency over viral tricks.
If your content delivers substance, builds trust, and sparks genuine discussion, you’re already aligned with how the new model thinks.

That’s exactly where AuthoredUp comes in:

• With Posts, you can study your entire content history and see which topics earn saves, comments, and follower growth, the metrics that actually matter.
• In the Editor, you can format, preview, and refine every post for clarity, structure, and readability the qualities that makes you recognized as expertise.
• With Drafts, you can tag and organize ideas to stay consistent across multiple profiles or brand voices, so the algorithm sees a clear, trustworthy signal of what you stand for.

The algorithm now rewards meaning, so your edge isn’t posting more, it’s adding value to topic with each post and comment. 

That’s how you grow in the 360 Brew era.

FAQ

How can I tell if 360 Brew is already affecting my posts?

Watch for small but clear behavioral shifts: sharper topic clustering in your feed, a rise in save-based engagement, and delayed “Suggested for you” resurfaces. If your impressions drop but profile visits and saves climb, you’re already inside the new ranking logic.

Why do some of my posts resurface days later?

Because 360 Brew doesn’t treat a post as “done” after the first 24 hours.
The new system re-checks older posts whenever new engagement matches their topic cluster.
If a post keeps collecting saves, long comments, or diverse replies, it can be shown again in “Suggested for you” feeds even a week later.

Think of it as delayed reach: fewer impressions early, but stronger long-tail visibility for quality content.

Will hashtags regain importance with 360 Brew?

Unlikely, because 360 Brew uses natural-language understanding to detect topics. It reads your actual text and follows meaning, not hashtags.

That said, think of them as navigation aids for humans, not signals for ranking.

Does the algorithm detect engagement pods or recycled comment patterns?

Yes and with accuracy. Because 360 Brew is a language model, it measures lexical diversity (how similar comments sound) and cluster overlap (who engages with whom).

If 10 comments use the same phrases or come from the same small circle, the system marks them as low-value signals. In contrast, varied replies from different users, different styles, sentence lengths, and job titles, strengthen trust.

So pods, automation, and “great-post” chains are now more than ever marked as red flags. It’s safer to have 5 unique comments than 50 AI generic ones.

How does 360 Brew handle multilingual content?

The model is trained mainly on English, but it uses cross-lingual embeddings (think semantic translation). That means it can connect similar ideas across languages, yet it still performs best when your profile language matches your post language.

If you mix languages in one post, LinkedIn may split its topic classification, reducing clarity.
It’s fine to post in multiple languages overall, but keep each post internally consistent.

Can creators still go viral under 360 Brew?

Yes, but virality looks different now. Instead of one explosive hour of reactions, success comes from sustained engagement across clusters (different audiences, roles, and regions).

A post might start small, then gain momentum three days later when a new professional group engages. This feed compression doesn’t kill virality; it rewards relevance longevity instead of clickbait bursts.

Are weekends or posting times still important?

Timing still matters, but less than before. 360 Brew personalizes distribution by interest and availability patterns, so great content can pick up later if it aligns with active conversations.

Peak hours still help early testing, but quality signals sucha as saves, replies, and topic clarity determine long-term reach.

Think of timing as a nudge, not a lever. If your content performs well on relevance metrics, it will travel beyond its posting hour.