> **Methodology in brief.** We separate two things. What is **measured**: our first-party LinkHub data (**657,722 real comments** with impressions) — raw reach, likes and replies per comment. What is **estimated**: the comment vs like "multiplier" in the algorithm, which no one can verify from the outside. We cite dated third-party figures, flag which ones are disputed, and never derive a first-party stat from an estimate.

## Key takeaways

- **The direction is consensus, not the magnitude.** Everyone agrees: the comment is the strongest engagement signal, the like (reaction) the weakest. *(van der Blom 2025, AuthoredUp 2025)*
- **The exact "multiplier" is an unverifiable myth.** Estimates range from **~2x** ([AuthoredUp](https://authoredup.com/blog/linkedin-algorithm)) to **~5–15x** depending on the source — a 1-to-7 spread that reveals no figure is proven. *(third-party, disputed)*
- **What is certain is the measured raw reach.** A comment generates on average **179 impressions** (35 median) for **0.90 like** → that is **~199 impressions per like**. *(LinkHub, n = 657,722)*
- **A comment earns almost as many replies (0.71) as likes (0.90)** → a real conversational effect, where a like is a dead signal.
- **A distinct article from [comments vs likes](/en/blog/commentaires-vs-likes-linkedin)**: there we measured a comment's *underestimated value*; here we settle the *algorithmic weight* question — honestly.

## 1. What our first-party data says (the solid ground)

Before talking about a multiplier no one can prove, let's start from what is **measured**. Across **657,722 real comments** whose impressions we recorded:

| Metric per comment | Average | Median |
|---|---|---|
| **Impressions** | **179** | **35** |
| Likes received | 0.90 | — |
| Replies received | 0.71 | — |

**Reading.** The average comment is seen **179 times** but triggers only **0.90 like** — about **199 impressions for a single like**. This is *not* a measure of algorithmic weight (we don't see LinkedIn's internal scoring); it's a measure of **real reach**, and it's unambiguous: a comment exposes you to a wide audience, regardless of the like counter *(detailed distribution in [impressions per comment](/en/blog/impressions-par-commentaire-linkedin))*.

Note too the **0.71 reply per comment**, almost on par with likes. The reply isn't a detail: it's a conversational back-and-forth, and it's precisely this kind of interaction that every source agrees to weigh as "heavier" than a passive reaction. For the full breakdown of that underestimated value, see the sister study [comments vs likes](/en/blog/commentaires-vs-likes-linkedin). To track your own impressions per comment, browse our other [LinkedIn data studies](/en/blog) and the [2026 LinkedIn statistics roundup](/en/blog/statistiques-linkedin-2026).

## 2. The consensus: comment = strongest signal, like = weakest

On the LinkedIn algorithm there is **one** point of near-unanimity, and it's this: **the comment weighs more than the like.** Not the other way around, never the other way around.

- The comment signals **active** engagement: the reader stopped, thought, wrote. The like is **passive** — a click.
- Reply threads (back-and-forth conversation) trigger far more aggressive reach expansion than reactions ([AuthoredUp, 2025](https://authoredup.com/blog/linkedin-algorithm)).
- The [van der Blom report (1.8M posts, 2025)](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895) confirms the algorithm prioritizes meaningful comments over vanity reactions.
- 2026 playbooks (e.g. [LinkPost](https://www.linkpost.gg/fr/playbooks/linkedin-algorithm-playbook-2026)) rank the substantive comment above the like in the signal hierarchy.

This **direction** is solid. It's the moment you try to **put a number on it** that everything falls apart — which is the subject of the next section.

## 3. The multiplier debate: from 2x to 15x

Once we accept that "comment > like", the question keeps coming back: **how many times more?** And here, sources diverge wildly.

| Source (dated) | Estimated multiplier | Status |
|---|---|---|
| AuthoredUp (2025) | **~2x** (with quality scoring) | estimate |
| LinkedCraft, meet-lea & various guides (2026) | **~5x** | estimate |
| Common industry estimates | **~5–15x** | **disputed** |
| Widely shared "15x" figure | **~15x** | **unverified** |

The problem is glaring: **a 1-to-7 spread between estimates.** When a "constant" varies by a factor of 7 depending on who's talking, it isn't a constant — it's a rounded rumor. None of these sources has access to LinkedIn's scoring code; they're inferences from partial tests, correlations, or the repetition of a figure that became folklore.

Worse: the weight depends on the **quality** of the comment. A two-word "Great post!" doesn't weigh like a 20–40 word comment that adds an angle ([AuthoredUp](https://authoredup.com/blog/linkedin-algorithm) describes a semantic scoring) — see [comment examples](/en/blog/exemples-commentaires-linkedin) that open a real conversation. Giving "one" single multiplier therefore ignores that the variable is continuous, not binary.

## 4. Does a comment really count 15x more than a like?

Honest answer: **nobody knows, and the "15x" figure is unverifiable.**

Let's cleanly separate the two levels:

- **The direction — reliable.** Yes, a relevant comment weighs clearly more than a like. On this, van der Blom, AuthoredUp and the 2026 playbooks converge. You can build your strategy on it without risk.
- **The magnitude — a myth.** "15x", "5x", "7x": these are seductive numbers because they're precise, but none is proven. LinkedIn doesn't publish its scoring; it evolves; and it depends on comment quality, the author, the timing. A fixed figure is, by construction, wrong.

What remains **true and measurable** is the **raw reach**: our 179 average impressions per comment for 0.90 like (~199 impressions/like). That's not the internal algorithmic weight — it's better: it's real exposure to the audience, observed across 657,722 cases. It's on that data, not on a mythical multiplier, that you should steer.

**The practical takeaway:** stop hunting for "how many likes is a comment worth". The right question isn't the ratio, it's the **reach** — and it depends mostly on **who** you comment under and **when**. A comment posted early under the right creator earns far more than a "well-weighted" comment lost under a dead post. To put your comments in front of the right audience at the right time, gather your targets in [personalized feeds](/en/features/feeds-personnalises) and let AI spot the best hosts via [AI profile recommendation](/en/features/ia-recommandation-profils).

## FAQ

**How many likes is a comment worth on LinkedIn?**
Nobody can say precisely. Estimates range from ~2x ([AuthoredUp](https://authoredup.com/blog/linkedin-algorithm)) to ~5–15x depending on the source — a 1-to-7 spread showing no figure is verified. What is certain: a comment weighs more than a like (direction), and it generates on average 179 impressions for 0.90 like (measured reach).

**Is the "15x" figure reliable?**
No. It's the highest and most shared estimate, but it's unverifiable from the outside. LinkedIn doesn't publish its scoring, which also depends on comment quality. Treat it as an order of magnitude, not a constant.

**Do comments count more than likes in the algorithm?**
Yes — that's the consensus point. The comment is the strongest engagement signal, the like (reaction) the weakest. The *direction* is solid; the *magnitude* is disputed.

**What's the difference with the "comments vs likes" article?**
The [sister study](/en/blog/commentaires-vs-likes-linkedin) measures a comment's *underestimated value* (its real reach vs its meager like counter). This article settles the *algorithmic weight* question — what's verifiable vs what's myth.

**What should I steer on, then, if the multiplier is wrong?**
On measured reach, not a ratio. Comment early and under the right creators: see [when to comment](/en/blog/quand-commenter-sur-linkedin) and the [LinkedIn algorithm 2026](/en/blog/algorithme-linkedin-2026).

## Sources & methodology

- **LinkHub dataset** — **657,722 real comments** with measured impressions: average 179 / median 35 impressions, 0.90 like and 0.71 reply per comment. Reach is measured; internal algorithmic weight is not, and is not claimed.
- [AuthoredUp — How the LinkedIn Algorithm Works (2025)](https://authoredup.com/blog/linkedin-algorithm) (~2x with quality scoring) · [van der Blom — Algorithm InSights Report 2025](https://www.linkedin.com/posts/richardvanderblom_chapter-1-algorithm-insights-report-2025-activity-7322514599126130688-Q895) (comments prioritized) · [LinkPost — Algorithm Playbook 2026](https://www.linkpost.gg/fr/playbooks/linkedin-algorithm-playbook-2026).
- ~5–15x estimates: **disputed and unverifiable**, cited as an order of magnitude only.
- Sister study: [comments vs likes](/en/blog/commentaires-vs-likes-linkedin).
- Find all our studies on the [LinkHub blog](/en/blog).