What's a good LinkedIn engagement rate in 2026 (data)

The median LinkedIn engagement rate is 2.99% in 2026. That's the real median — measured across 516,144 posts, not pulled from a vague marketing study.

In short

→ The median engagement rate is 2.99% (likes + comments / impressions). → The like rate alone is 2.16%; the comment rate is 0.62%. → Median impressions per post: 620. → The p90 sits at 26.8% — 10% of posts exceed that threshold. → Comparing yourself to the "average" means comparing yourself to viral outliers. Use the median.

How LinkedIn engagement rate is calculated

LinkedIn doesn't show you an "engagement rate" directly in its native analytics. The standard calculation — the one we use across our corpus:

Engagement rate = (likes + comments) / impressions × 100

Some tools add shares or clicks. Our benchmark sticks to likes and comments divided by impressions: it's the most stable and comparable definition across tools.

Real benchmarks — measured across 516,000 posts

Across our corpus of 516,144 posts covering 30,843 profiles, here are the raw numbers:

| Metric | Median | |--------|--------| | Impressions per post | 620 | | Like rate | 2.16% | | Comment rate | 0.62% | | Total engagement rate | 2.99% | | Engagement rate p90 | 26.8% |

What this tells you: half of all posts generate less than 2.99% engagement. And the top 10% exceed 26.8% — which shows just how skewed the distribution really is.

Median vs. average: the classic mistake

Most published benchmarks use the average — pulled upward by viral posts from top creators. In our analysis of 516,000 LinkedIn posts, the average can run 4 to 5× higher than the median. Targeting the "average" is effectively targeting the p70 or above, which completely distorts your self-assessment.

Look at three numbers instead:

→ Your personal median across your last 10–20 posts (your actual current level). → The trend of that median month over month (your trajectory). → The gap between your best post and your median (your untapped potential).

What counts as a "good" engagement rate — a simple grid

| Engagement rate | What it means | |-----------------|---------------| | Below 1% | Below median — needs work | | 1%–3% | In the normal range (around the median) | | 3%–10% | Solid performance | | Above 10% | Top 10% of posts (p90 = 26.8%) |

One important caveat: engagement rate varies significantly by format. Comparing a text post to a carousel on this metric alone doesn't make sense. The real distribution of LinkedIn posts shows that visual formats structurally generate more comments — and therefore a higher rate. Normalize by format before drawing conclusions.

Why your rate might be low — the real causes

Several factors drag engagement rate down:

→ No hook: if your first two lines don't stop the scroll, nobody reads — and nobody reacts. → Text-only format: the median like rate for text-only posts is roughly half that of video or carousel. → No question or CTA: without an explicit invitation to engage, readers scroll silently past. → Impressions without interaction: a post with an external link loses around 36% of median impressions. Less reach means fewer opportunities to earn engagement. → Single-lever approach: posts that stack 4–6 writing levers (data proof, polarizing angle, visual format, strong hook…) consistently outperform posts that rely on just one.

What to do right now

→ Calculate your median across your last 10–20 posts (likes + comments / impressions). → If you're below 1%: start with the hook and switch up your format. → If you're between 1% and 3%: stack levers — add a data point and a clear call to action. → If you're above 3%: focus on consistency and publish timing.

Before your next post goes live, check its engagement profile with the Post Analyzer — it shows you which factors are pulling your score down, for free. And to track your metrics post by post over time, take a look at LinkPost Analytics.

Engagement isn't declared — it's built by stacking the right signals, post after post.

Observational study by LinkPost across 516,144 posts (2020 to April 2026, 62% French-language content, 30,843 profiles). Figures presented are medians measured on the corpus. Correlation does not imply causation, and sampling bias (active creators on the platform) should be taken into account. Full methodology in the playbook.