Saves vs likes: which metric really matters on LinkedIn

Creators have been hyping LinkedIn saves as the platform's ultimate metric. The problem: in our corpus of 516,000 posts analyzed, saves are nearly absent from measurable data. What actually decides your reach is much simpler.

In short

→ LinkedIn saves do not appear as a reliable signal in the data — too infrequent in our corpus to draw robust conclusions. → Likes measure an active reaction, with a median rate of 2.16% of impressions. → Comments are the most powerful signal: they generate dwell time and engagement velocity. → Impressions (median: 620 per post) are the true measure of your raw reach. → Steering your strategy on saves is navigating without a compass. Focus on what is measurable and actionable.

LinkedIn saves: a signal that disappears in the data

You often hear that "saves send a strong signal to the algorithm." That may be true in theory — LinkedIn can weight them in its ranking model. But in our analysis of 516,000 posts, saves surfaced at near-zero levels in accessible metrics. LinkedIn does not expose saves publicly through its API the same way it does likes or comments.

The practical consequence: you cannot steer your growth on a metric you cannot measure systematically. And more importantly, you have no way of knowing whether your "saved" posts had any impact on their distribution in the feed. It is a black box inside a black box.

Treat saves as an indicator of perceived quality — not as a reach lever.

What likes actually measure (and their limits)

The like is the most visible metric, the easiest to count. In our corpus, the median like rate is 2.16% of impressions, with a median of 18 likes per post. The average climbs to 85 because of viral outliers — which confirms the power law that governs LinkedIn.

A like signals that content triggered a reaction, but it is a passive signal: it generates no conversation and does not extend time spent on the post. LinkedIn factors it in, but with lower weighting than comments.

What likes do not measure: → Whether someone read your full post (dwell time) → Whether your post created intent (DM, profile click, link visit) → Whether you gained new followers from that post

For those, you need to look at other metrics.

Comments: the signal the algorithm values most

LinkedIn's algorithm weights comments significantly more than likes. The reason is mechanical: a comment proves someone read the post, thought about it, and took the time to write something. That is extended dwell time, a strong interest signal, and it is visible to the commenter's network.

In our corpus, the median comment rate is 0.62% of impressions. Low at first glance — but posts that cross this threshold see their distribution expand noticeably. That is not a coincidence.

Our study of 438,413 posts confirms that the tactics generating the most comments — polarizing with data, open questions, comment-gates — are also the ones that amplify reach most. The two are linked by the same mechanism.

Impressions: the only metric for raw reach

LinkedIn impressions count every time your post appears in the feed, whether or not it is followed by a reaction. It is the most honest measure of your actual reach. While likes can vary with topic or tactic, impressions directly reflect the distribution the algorithm grants.

| Metric | Median | p90 | |--------|--------|-----| | Impressions per post | 620 | 3,867 | | Like rate | 2.16% | n/a | | Total engagement rate | ~3% | 26.8% | | Comment rate | 0.62% | n/a |

Source: LinkPost corpus, ~359,000 posts with metrics over 180 days.

An engagement rate of ~3% — likes plus comments relative to impressions — corresponds to the median. Below that, your post struggles to escape the initial sample. Above it, it enters the amplification phase.

To understand how LinkedIn distinguishes impressions, views, and unique members reached — three concepts many people confuse — read the dedicated guide.

The real ranking of metrics to watch

| Metric | Measurable | Algorithm weight | Worth tracking | |--------|-----------|-----------------|----------------| | Saves | No (near-zero in our corpus) | Potential, unvalidated | No | | Impressions | Yes | Reach benchmark | Yes, first | | Comments | Yes | Strong | Yes | | Likes | Yes | Moderate | Yes | | Followers gained per post | Yes | Indirect | Yes |

Saves do not disappear from the equation permanently. Maybe LinkedIn will make them more accessible, or a future study will isolate them properly. But today, they are not part of what you can actually steer.

What to do, concretely

→ Track your impressions week over week — that is your real reach, not likes. → Optimize for comments: ask a genuine question, take a position with data, create conversation rather than validation. → Target an engagement rate above 3% as your minimum quality threshold. → Ignore saves in your strategy for as long as they remain unmeasurable in any systematic way. → Before you publish, check your post's score across 33 virality criteria with LinkPost's free analyzer.

The most seductive metric is not always the most useful one. On LinkedIn, what counts is what you can measure, steer, and improve over time.

Observational study by LinkPost (corpus: 516,144 posts, ~359,000 posts with metrics over 180 days). Since saves are not reliably surfaced in accessible metrics, conclusions on this point remain cautious. Correlation ≠ causation. Full methodology in the playbook.