> **The diagnosis, in one sentence.** You publish, and your posts get 80 views, two likes from your mom, and silence. Before deciding "the algorithm hates you," you need to **diagnose**. In 90% of cases the problem isn't one thing but several: no audience that engages early, a format the algo doesn't push, no signal in the first hour, an external link that tanks reach, or bad timing. We walk through the **5 causes** — backed by our first-party LinkHub data and dated third-party studies — then look at the fastest way to warm up your visibility: **commenting**. A caveat: reach and engagement are correlated, not mechanically causal — a good post can still flop, and vice versa.

## Key takeaways

- **Cause #1: you have no audience that engages early.** The algo tests your post on a small fraction of your network; without fast first comments, it never amplifies it. *(see [LinkedIn algorithm](/en/blog/algorithme-linkedin-2026))*
- **Cause #2: your format is weak.** Across **57,809 posts** (LinkPost), video averages **146 likes / 28 comments**, image **127 / 35**, carousel **100 / 29**, and plain text **65 / 16**. Raw text starts with a handicap. *(see [post format](/en/blog/format-post-linkedin-engagement))*
- **Cause #3: no signal in the first hour.** The "golden hour" decides everything — without early engagement, the post dies. *(third-party, [van der Blom 2025](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895))*
- **Cause #4: you add an external link.** A post with an external link loses ~60% of reach ([AuthoredUp 2025](https://authoredup.com/blog/linkedin-algorithm)), and "link in the comments" is now detected too. *(third-party, dated)*
- **Cause #5: you post at the wrong time.** Outside activity windows (Tue–Thu, mornings), your first sample is too small to trigger amplification.
- **The fastest fix: comment.** Until your audience engages, you borrow other people's — a well-placed comment is seen **3 to 40×** more depending on the host. *(LinkHub — see [grow without posting](/en/blog/grossir-linkedin-sans-poster))*

## 1. You don't (yet) have an audience that engages early

This is the most common cause, and the most misunderstood. LinkedIn's algorithm doesn't broadcast your post to everyone: it **tests** it first on a small fraction of your network, watches engagement in the **first few minutes**, then decides to amplify — or not (see [how the algorithm works](/en/blog/algorithme-linkedin-2026)).

The catch: if your network is small, cold, or full of people who never read you, that first sample produces **no signal**. No fast comment, no thread → the algo concludes nobody cares and cuts distribution. You could have written the best post in the world: without early engagement, it never leaves the test stage.

It's mechanical, not personal. That's also why already-big accounts break through more easily: they trigger comments fast, so they pass the test stage more often. The share of visibility held by "Top Creators" rose from **15% (2022) to 31% (2025)** ([van der Blom, 1.8M posts](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895)). When you start small, the system works against you.

## 2. Your format works against the algorithm

For the same audience, **format changes everything**. Our format data (LinkPost, n = 57,809) shows a clear gap:

| Format | Avg. likes | Avg. comments |
|---|---|---|
| **Video** | 146 | 28 |
| **Image** | 127 | 35 |
| **Carousel** | 100 | 29 |
| **Text only** | 65 | 16 |

A **plain-text** post starts with roughly **half** the engagement of a video or image. If you only publish walls of text, you mechanically cap your reach. Format isn't cosmetic: it drives **dwell time** (reading time), a central 2026 algo signal. A carousel or a video holds attention longer than a paragraph skimmed in two seconds.

Honest note: these are **averages**. Virality stays rare — median of **27 likes**, and only **17.4%** of posts reach 100+ likes. Format improves your odds; it guarantees nothing. Details in [which post format drives the most engagement](/en/blog/format-post-linkedin-engagement) and [what makes a post go viral](/en/blog/post-viral-linkedin-statistiques).

## 3. You trigger no signal in the first hour

Even with an audience and a good format, a post can die if it generates **nothing** in its first hour. That's the famous "golden hour": engagement in the **first 60–90 minutes** decides whether the algo amplifies ([van der Blom 2025](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895)).

Concretely, if you publish then close the app, nobody reacts, and the window closes. The levers to feed it:

- **Reply fast** to the first comments — each reply creates a thread that extends distribution.
- **Trigger comments, not likes.** A comment is the algo's strongest signal, well ahead of a like *(third-party, contested — see [comments vs likes](/en/blog/commentaires-vs-likes-linkedin))*. Ask a real question, take a stance.
- **Be present.** A post's worst enemy is the author who publishes and vanishes.

## 4. You add an external link (or hide it badly)

This is the silent error that tanks the most posts. LinkedIn wants to **keep people on the platform**: a post that sends them elsewhere gets penalized. According to industry studies, a post with an external link loses **~60% of reach** versus the same post without one ([AuthoredUp 2025](https://authoredup.com/blog/linkedin-algorithm)). *(third-party, dated)*

And the old workaround — "link in the first comment" — barely works anymore: since early 2026 the algo detects "bridge behavior" (a post clearly designed to push toward a link in the comments) and applies a similar penalty ([dataslayer 2026](https://www.dataslayer.ai/blog/linkedin-algorithm-february-2026-whats-working-now)). The fix: keep the link out of the post, or add it **after** the post has gained traction, once distribution is already running.

## 5. Is timing really the culprit?

Timing matters, but less than people think — and it mostly acts through cause #1. If you publish on a Sunday at 11pm, your first network sample is **too small and too inactive** to generate the signals that trigger amplification. Activity windows give you a baseline:

- **Days**: Tuesday, Wednesday, Thursday; weekends underperform.
- **Hours**: mornings (6am–1pm UTC) are the activity peak in our data; sharp drop in the evening.

But beware the **correlation trap**: posting "at the right time" won't save a post with no audience or format. Timing maximizes your starting sample size; it doesn't create engagement. It's a multiplier, not an engine. For detailed engagement windows, see [the best hours to comment](/en/blog/meilleures-heures-pour-commenter-linkedin).

## 6. The fix that works right now: warm up by commenting

All these causes share one root: **you don't yet have an audience that engages**. Building that audience by posting takes months — and the system already favors the big accounts. The shortcut is to **borrow other people's audiences by commenting**.

When you comment early under a creator in your niche, you expose yourself to **their post's audience peak**, not yours. Our data quantifies it: across **4,861 creators**, the median comment earns **36 impressions**, a top 10% **124 (~3.4×)**, a top 1% **357 (~10×)**, and the observed max **1,522 (~42×)** — same comment, different host (see [who to comment on](/en/blog/sur-qui-commenter-linkedin)). And timing matters fully here: **391 impressions** at 0–30 min vs **104** after 24h, i.e. **~3.8×** (n = 261,137).

Commenting makes you **visible to the exact audience you're targeting**, sparks connections, and — as a side effect — warms up your own network: people who discover you in comments come engage on your posts later. That's the strategy detailed in [grow on LinkedIn without posting](/en/blog/grossir-linkedin-sans-poster) and [is commenting better than posting](/en/blog/commentaire-vs-post-linkedin).

To run it without burning your whole day: [personalized feeds](/en/features/feeds-personnalises) gather your target creators so you see their posts the moment they go live, the [AI profile recommendation](/en/features/ia-recommandation-profils) infers your [ideal customer](/en/blog/icp-commentaire-linkedin) and finds the best hosts across 100,000+ profiles, and [personalized AI comments](/en/features/ia-commentaires-personnalises) — always validated by you — quickly draft a comment that triggers the right signals.

## FAQ

**Why did my LinkedIn post only get 50 views?**
Because the algo tested it on a small fraction of your network and saw **no engagement signal** in the first hour. Without a fast comment or reply, it doesn't amplify. Also check the format (plain text underperforms) and that there's no external link.

**Do external links really kill reach?**
Yes — it's one of the most documented factors: ~60% less reach per industry studies ([AuthoredUp 2025](https://authoredup.com/blog/linkedin-algorithm)). "Link in the first comment" is now detected too. Keep the link out of the post or add it after takeoff.

**Which format should I post for more reach?**
Video, image, or carousel over plain text: our data (n = 57,809) shows up to ~2× more engagement. But format is a multiplier, not a guarantee — the median stays at 27 likes. See [post format](/en/blog/format-post-linkedin-engagement).

**Do I need to post at a specific time?**
Timing maximizes your starting sample size (Tue–Thu, mornings), but it doesn't create engagement. A good slot won't save a post with no audience or format: it's a multiplier, not the engine.

**How do I get reach when starting from zero?**
Comment before you post. You borrow established creators' audiences: a well-placed comment is seen 3 to 40× more depending on the host, and it warms up your network for your future posts. See [grow without posting](/en/blog/grossir-linkedin-sans-poster).

## Sources & methodology

- **First-party LinkHub data** — post formats (LinkPost, n = 57,809: video 146/28, image 127/35, carousel 100/29, text 65/16; median 27 likes, 17.4% ≥100), reach per comment (4,861 creators: p50 = 36, p90 = 124, p99 = 357, max = 1,522), timing (261,137 comments: 391 impressions at 0–30 min vs 104 at 24h+). See [post format](/en/blog/format-post-linkedin-engagement), [who to comment on](/en/blog/sur-qui-commenter-linkedin), [when to comment](/en/blog/quand-commenter-sur-linkedin).
- [van der Blom — Algorithm InSights Report 2025](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895) · [AuthoredUp (2025)](https://authoredup.com/blog/linkedin-algorithm) · [Botdog (2025)](https://www.botdog.co/blog-posts/linkedin-algorithm-2025) · [dataslayer (2026)](https://www.dataslayer.ai/blog/linkedin-algorithm-february-2026-whats-working-now)
- **Methodology honesty**: reach and engagement are correlated, not mechanically causal. A good post can still flop, and a good format is only a multiplier of odds. For a full data overview, see the [LinkedIn statistics 2026 roundup](/en/blog/statistiques-linkedin-2026).
- Find all our studies on the [LinkHub blog](/en/blog).