How the LinkedIn Algorithm Works in 2026: Depth Score, Rankings and Best Practices

The LinkedIn algorithm in 2026 ranks content primarily on Depth Score: how long and how meaningfully people engage with a post (dwell time, comment quality, saves), rather than how many quick reactions it gets. This is the biggest shift in how LinkedIn distributes content since it moved from chronological to algorithmic feeds and it changes what "good" LinkedIn content looks like.

LinkedIn's feed has moved from a Social Graph model (show people content from their connections) to an Interest Graph model, where the system evaluates each post on topic relevance and depth of consumption, then decides how far to push it based on how that engagement behaves.

Below: what LinkedIn has confirmed about how this works, what's still observed rather than official, how different content formats actually perform, and what to change about how you post. 

What is the LinkedIn algorithm?

The LinkedIn algorithm is the system that decides which posts appear in a member's feed, in what order, and how far each post travels beyond the poster's immediate connections. Like any major platform's feed algorithm, it isn't one rule but a layered system: a content-quality filter, a small-scale distribution test, then an expansion decision based on how that test group engages.

What's changed since LinkedIn's earlier filtering-focused approach is the shift toward an interest-graph model: instead of ranking mainly by your network, the system evaluates each post's topic relevance across a much wider pool of members, then decides distribution based on engagement depth with that topic — powered by 360Brew, a 150-billion-parameter model LinkedIn confirmed deploying to its live feed in March 2026.

For how this compares across platforms, see our guide to how social media algorithms work.

How does the LinkedIn algorithm work? A step-by-step breakdown

Every post moves through roughly four stages:

1. Content classification (0–60 minutes). LinkedIn's systems scan for spam signals, low-quality patterns and formatting or link red flags before the post is shown beyond a small pool.
2. Small-audience test. The post goes first to your most-engaged connections and a wider pool the system predicts is interested in the topic.
3. The golden hour. The first roughly 60–90 minutes are decisive — strong early engagement, especially comments, signals the algorithm to expand distribution.
4. Extended distribution. Posts that keep performing continue spreading well past the initial test group; since a mid-2025 change, strong posts can now circulate for two to three weeks rather than fading after a day.

What has LinkedIn confirmed about the 2026 Depth Score algorithm?

LinkedIn has directly confirmed three things:

1. Dwell time as a ranking input. LinkedIn's own engineering team documented this back in a 2020 post, "Understanding dwell time to improve LinkedIn feed ranking" — this isn't new for 2026, but it's the foundation the current system builds on.
2. The "Long Dwell" classifier. LinkedIn's published research, the LiRank paper and its successor LiGR, describes a binary classifier predicting whether a post's dwell time will exceed a context-dependent percentile threshold, feeding that prediction directly into ranking.
3. 360Brew, deployed live. LinkedIn confirmed via its Engineering Blog that this 150-billion-parameter model went live in the feed on 12 March 2026.

What's still unconfirmed about the Depth Score algorithm?

Most of the specific numbers circulating in 2026 come from third-party tool vendors and creator-analytics platforms, not LinkedIn. Useful directionally, not as exact figures:

• The 61-second benchmark. The claim that posts crossing 61+ seconds of dwell time see a 15.6% engagement rate (vs. 1.2% under 3 seconds) circulates widely but isn't traceable to a LinkedIn-published source (DigitalApplied's 2026 guide is one example repeating it).
• Signal weighting. Estimates vary widely: comments are reported anywhere from roughly 2x a like to the commonly repeated "~15x" figure. Saves are commonly cited around 5x a like. The spread itself signals these are estimates, not a published formula.
• The external-link penalty. Widely reported at roughly 50–60% reduced reach, consistent across independent sources but not confirmed by LinkedIn itself.
• Engagement-pod detection. Multiple sources report LinkedIn's systems now detect and suppress reciprocal engagement patterns without notifying the account, tightened further by a reported March 2026 "Authenticity Update."

Which LinkedIn engagement signals matter most in 2026?

Not all engagement counts equally:

• Likes and reactions are passive and low-weight — they keep a post visible but barely move distribution alone.
• Comments are the dominant active signal, especially substantive back-and-forth threads.
• Saves are a strong, comparatively new signal (added to creator analytics in late 2025), reportedly weighted more heavily than a like or comment individually.
• Shares/reposts signal content worth wider distribution and can extend reach over following days.
• Dwell time underlies all of the above — it's why formats that naturally hold attention tend to outperform.

What type of LinkedIn content performs best in 2026?

• Document/PDF carousels lead on engagement, commonly cited around a 6.6% engagement rate, largely because swiping through slides generates sustained dwell time.
• Native video builds connection fastest and is reported to drive roughly 5x more "conversations started" than static posts, though average reach still tends to trail carousels and text.
• Text posts remain the reliable workhorse — quick to produce, highest reach ceiling when the hook lands, though typical engagement rates struggle to break 2%.
• Single-image posts are currently weak, reportedly underperforming text-only posts by roughly 30% — a reversal from 2024–2025.
• Polls have largely collapsed, with engagement rates reported as low as 0.07% following the referenced 2026 update.
• Posts with external links see reduced reach — putting the link in the first comment is the common workaround.

What are the best LinkedIn posting practices for 2026?

• Lead with a short, specific first line — not a question. Analysis of over a million posts found number-led openers outperformed vague ones, while question-style hooks actually underperformed and drew fewer comments. LinkedIn folds text after roughly 210 characters behind "see more," so this line carries real weight.
• Keep posts scannable — short paragraphs, generous line breaks, roughly 900–1,500 characters as a starting range.
• Post consistently, not constantly — two to five posts a week is the practical sweet spot; more than once a day can cannibalize your own reach. For ideas to sustain that cadence, see our LinkedIn post ideas.
• Time posts for Tuesday–Thursday, mid-morning — the consistent finding across independent studies, though the exact peak hour varies by source.
• Ask genuine, specific questions — not engagement bait. Prompts like "Comment YES if you agree" are increasingly detected and suppressed.
• Protect the golden hour — reply to comments in the first 60–90 minutes.
• If sharing a link, put it in the first comment, not the post body.
• Keep your profile itself optimised — the algorithm's topic-relevance matching starts with what your profile signals. See our LinkedIn profile best practices.

What Sendible's own data shows about Pages vs. Profiles

Most 2026 LinkedIn advice, including the algorithm mechanics above, points toward personal profiles as the only place worth investing — and for individual reach, that's borne out. But Sendible's 2026 Social Media Management Trend Report, based on an analysis of over 12 million posts scheduled through Sendible in 2025, found LinkedIn Company Page activity climbing, not shrinking: scheduled volume to Company Pages grew 6.1% year-over-year to a 9.2% share of all scheduled posts, even as personal profile scheduling grew more modestly (+0.4%, 4.5% share).

The likely explanation: as AI-generated content and individual "creator fatigue" increase noise on personal feeds, a well-maintained Company Page becomes a verified point of reference — particularly for agencies and multi-location brands that need brand-safe, scalable presence a single employee's profile can't provide. In practice, that argues for a hub-and-spoke approach: personal profiles drive reach and Depth Score-friendly engagement, while the Company Page serves as the credible destination that traffic converts against. See our LinkedIn scheduling tool if you're managing both in parallel.

Does this data hold up against real LinkedIn posts?

Sendible's Head of Content, Tamara Biljman, put this guide to the test, comparing her own LinkedIn posts since July 2026 against the research above.

What matched:

• PDF carousels performed best for her too, at roughly a 6.7% engagement rate — closely tracking the ~6.6% figure cited earlier in this guide.
• Her posts with a link preview reached about half as many people as her average posts, consistent with the external-link reach penalty.
• Statement-led openers reached roughly a third more people than question-led openers, matching the finding that question hooks underperform.
• Text-only posts stayed just under a 3% engagement rate, close to the ~2% ceiling cited above.

What didn't:

• Single-image posts didn't underperform for her. Real, unpolished photos — including one of her actual home-office setup — reached about 40% more people than her text posts, with nearly double the engagement rate. Branded graphics, meanwhile, were her second-weakest format.

Her read: it's less that "images underperform" and more that people respond to something that feels like a person rather than a brand asset, which tracks with a Depth Score algorithm that rewards genuine dwell time over polish.

It's a small sample from one person's audience, not a universal rule, but it's a useful real-world gut check on the aggregate data above and a reminder to test format performance against your own audience rather than take any guide, including this one, as gospel.

Frequently asked questions about the LinkedIn algorithm

What is the LinkedIn Depth Score?
Depth Score is the informal name for LinkedIn's 2026 ranking approach, which weighs how long and how meaningfully people engage with a post — dwell time, comment substance, saves — more heavily than raw reaction counts.

How does LinkedIn's algorithm actually work, step by step?
Posts move through content-quality classification, a small-audience test, a decisive "golden hour" of early engagement, and extended distribution if they keep performing — with dwell time and comment depth weighted throughout.

What type of LinkedIn content performs best in 2026?
Document/PDF carousels currently lead on engagement rate, followed by native video and text posts; single-image posts and polls are both underperforming relative to prior years.

What are the best times to post on LinkedIn in 2026?
Tuesday through Thursday, mid-morning, is the consistent finding across independent studies, though the exact peak hour varies by source.

Do external links hurt reach on LinkedIn in 2026?
Multiple independent sources report a meaningful reach reduction (commonly cited around 50–60%) for posts containing external links, though LinkedIn hasn't published an official figure.

Is the LinkedIn algorithm the same as Facebook's or Instagram's?
No — each platform ranks on different signals suited to its content type. See our overview of how social media algorithms work for a cross-platform comparison.