What we learned analyzing 516,000 LinkedIn posts

We analyzed 516,144 LinkedIn posts across 9.6 million snapshots and 30,843 profiles. Here are the findings that rewrite what actually works — numbers first.

In short

→ 1 in 3 posts flops: 34.8% get fewer than 10 likes. Virality follows a brutal power law. → Video and carousels drive 2× more median likes than text-only posts. → Long posts (1,800+ characters) hit a median of 25 likes versus 12 for short ones. → 1 hashtag is optimal: going to 4+ drops engagement by 40%. → An external link in the body costs you −36% in median impressions.

How we built this study

Two corpora, two complementary angles:

• Fresh corpus: 516,144 posts, 9.6 million metric snapshots, 30,843 profiles (2020–2026, 62% French-language content). This is the basis for all measurements on this page.
• Playbook: 438,413 posts, 5,291,997 comments, 24,006 creators — dissected to extract the patterns of the top 1% creators.

These are observational results. Correlation does not imply causation. Sampling biases exist (active creators, public data). The findings are trends, not individual guarantees.

Virality distribution is brutal

The median is 18 likes per post. The average jumps to 85, pulled up by viral outliers. The p99 reaches 1,226 likes — 68× the median.

34.8% of posts flop at fewer than 10 likes. Only 14.4% cross 100. The distribution follows a power law: a tiny minority of posts captures most of the attention. Full breakdown in the real distribution of LinkedIn virality.

Format decides a factor of 2

| Format | Median likes | |--------|-------------| | Video | 26 | | Carousel | 24 | | Image | 20 | | Text-only | 11 |

Text-only posts get half the likes of video. Carousels also deliver 2.3× more median impressions than text — they're the format most amplified by the algorithm. Full detail in the LinkedIn format performance analysis.

Length works in your favor

| Length (characters) | Median likes | |--------------------|-------------| | Under 300 | 12 | | 300–700 | 17 | | 700–1,200 | 19 | | 1,200–1,800 | 21 | | 1,800 and above | 25 |

Median engagement doubles between the shortest and longest posts. Long-form content (1,500+ characters) also generates +49% engagement in the playbook. It's not that long content is inherently better — it increases dwell time, the algorithm's primary signal. See the ideal LinkedIn post length for the full breakdown.

Timing: less competition beats "the perfect slot"

Weekend posts show a median of 20 likes versus 17 on weekdays, with 3× less volume published. Less competition means more reach per post.

On hourly slots, the French morning window (7–10am CET) has the lowest median at 16 likes — it's the most crowded time of day. Noon CET climbs to 25 median likes. The best time to post on LinkedIn and the weekend effect each have dedicated analyses.

Hashtags: 1 is optimal, 4+ penalize

| Number of hashtags | Median likes | |-------------------|-------------| | 0 | 19 | | 1 | 23 (+21%) | | 2–3 | 17 | | 4–6 | 11 (−40%) | | 7 or more | 11 |

A single hashtag adds +21% versus none. Going to 4+ drops to 11 median likes — a −40% hit compared to posting without any. Full analysis: how many hashtags on a LinkedIn post.

An external link costs −36% impressions

Posts with an external link (n=74,891) have a median of 448 impressions. Without a link (n=284,130): 705. That's −36% in median impressions. Likes are slightly higher with a link — probably a selection bias, since creators sharing links tend to post more targeted content. Correlation, not causation, but the signal is consistent. See do external links kill your reach.

Writing tactics that make the difference

The playbook isolated the patterns of top 1% creators:

→ 80% of highly viral posts contain a strong hook in the first 200 characters. → 61% include a data point as proof in the hook. → Posts with high polarization (above 0.7) generate 2.75× more likes — when backed by data. → Viral posts stack 4 to 6 tactics. Single-tactic posts rarely cross the threshold.

Full detail in the writing tactics analysis across 350,000 posts.

English vs. French posts: +67% in median likes

English posts show a median of 40 likes (n=92,184) versus 24 for French (n=294,849). That's not because English is structurally better — English-language creators typically have a larger global network, which boosts distribution signals. Language choice remains a strategic decision based on your target market.

What this means for you, concretely

→ Write long, anchored on a number, with a hook in the first 200 characters. → Favor video or carousels when you're targeting reach. → 1 hashtag, not 4. → Avoid links in the body when you want impressions. → Post on weekends or at noon: less competition, better initial signal. → Stack tactics — hook + data proof + visual format + length.

Before you publish, run your post through the free LinkPost analyzer: it scores your content on 33 criteria drawn from this data and shows you exactly where you can improve.

For a complete LinkedIn acquisition strategy built on these findings, LinkEarn offers 1:1 B2B coaching that applies the full dataset.

Observational study by LinkPost (2020 to April 2026, 62% French-language content). Findings are observed correlations, not established causal relationships. Sampling biases apply: active creators, public metrics only. Methodology and limitations detailed in the full playbook.