LinkedIn Impressions Explained, and What to Do When Yours Are Low

Almost every LinkedIn conversation I have with a client starts the same way. Someone shares a screenshot of a post, points at the impression count, and asks whether that number is good. It is a fair question and it has no honest short answer, because impressions are the most watched and least understood metric on the platform.

I have spent years running LinkedIn programmes for recruitment firms, search businesses and B2B consultancies, and I have watched people make genuinely bad decisions off the back of that one number. They kill content formats that were working. They chase reach that never had any chance of turning into a conversation. They panic about a decline that is happening to everyone.

So this is the piece I wish I could hand people at the start. What an impression actually counts, how it differs from the metric sitting next to it, what a reasonable number looks like, why yours have almost certainly fallen, and the changes that genuinely move it. I have used LinkedIn's own documentation where it exists and named the studies where it does not, so you can check anything you want to check.

What are impressions on LinkedIn?

An impression is counted each time your post appears on somebody's screen. That is the whole definition. It is a measure of distribution, not of readership.

LinkedIn's post analytics documentation describes impressions plainly as the number of times your post was shown on LinkedIn, and its supporting help pages put a technical threshold on it: the post has to be visible for at least 300 milliseconds with at least half of it in view, on a signed-in member's device screen or browser window.

Three things follow from that, and all three matter.

It counts appearances, not people. If the same person scrolls past your post three times over two days, that is three impressions. The metric is deliberately non-unique.

300 milliseconds is not reading. It is roughly the time it takes to scroll past something. An impression tells you your post was in front of someone's eyes for about a third of a second. It says nothing about whether they took in a single word.

It includes other people's reposts. Since April 2024, the count on your post rolls in impressions earned when someone else reposts your content. That is a fairer picture of total distribution, but it does mean a single repost by a well-followed person can inflate a number in a way that has very little to do with the post itself.

None of this makes impressions useless. It makes them the top of a funnel rather than a result. Distribution is the thing that has to happen before anything else can, which is exactly why it is worth understanding properly rather than staring at.

Impressions and members reached are not the same thing

Sitting next to impressions in your analytics is a second number, members reached, and the gap between the two is one of the more useful signals LinkedIn gives you for free.

LinkedIn defines members reached as the number of distinct members and Pages that saw your post, excluding repeat views, and is careful to say the figure is an estimate. So if your post was shown to one person five times, you have five impressions and one member reached.

Divide impressions by members reached and you get an average frequency. In practice you will usually see something between 1.1 and 1.6. That range is normal and healthy. When it climbs well above that, your post is circulating hard inside a small pool of people who keep seeing it, which usually means the algorithm is not finding new audiences for it. When it sits very close to 1.0, almost everyone saw it once and moved on.

I look at members reached first and impressions second, for a simple reason. Members reached is the closest thing LinkedIn gives you to a count of humans, and humans are what you are actually trying to influence.

What impressions do not tell you

This is the part people skip, and it is where most of the bad decisions come from.

Impressions do not tell you whether anyone read the post. They do not tell you whether the people who saw it were the people you wanted to see it. They do not distinguish between a hiring manager at a target account and a jobseeker in another country who follows you by accident. And they do not tell you anything about whether the post moved anyone closer to picking up the phone.

LinkedIn does give you a partial answer to the targeting question. Under each post there is a viewer demographics panel breaking down who saw it by job title, company, industry, seniority and location. Almost nobody uses it, and it is worth ten times more than the headline number. A post with 2,000 impressions where a third of viewers are directors at firms you sell to is a far better post than one with 12,000 impressions from a general audience.

There is also a third-party check worth doing occasionally: the in-network and out-of-network split. High out-of-network percentages mean LinkedIn is pushing your content beyond your immediate connections, which is what growth looks like. High in-network percentages mean you are mostly talking to people who already know you, which has its own value for a professional services firm but will not build anything new.

What is a good number of impressions on a LinkedIn post?

The honest answer is that the raw number is meaningless without your follower count next to it, and comparing yourself to screenshots you see in your feed is a fast route to feeling terrible for no reason.

Benchmark data gives a sense of scale. An analysis of 1.3 million posts from more than 16,000 company pages found pages in the 10,000 to 50,000 follower band averaging roughly 1,850 impressions on multi-image posts, around 1,420 on video and about 1,150 on native document posts. Those are company pages rather than personal profiles, and personal profiles generally do better, but it anchors expectations somewhere sensible.

The test I actually use is a ratio. Take impressions and divide by your follower count.

• Below 0.5 suggests the post did not travel even within your own audience, which usually points at the opening lines or the format rather than the idea.
• Around 1 to 3 is a normal, working post. LinkedIn showed it to your audience and some way beyond.
• Above 3 means the post found people who do not follow you, which is what you want from anything designed to grow reach.

Then compare each post against your own median rather than against your best ever post. Your median is the number that tells you whether your programme is improving. Your best post tells you almost nothing, because outliers on LinkedIn are frequently the product of one large repost rather than anything repeatable.

What is a good engagement rate on LinkedIn?

Engagement rate is a more stable health check than impressions, because it is not distorted by how many followers you happen to have.

The same study of 1.3 million posts puts the overall average at 5.20 per cent, up around 8 per cent year on year. Broken down by format, the pattern is consistent enough to be useful:

Two caveats before you redesign your content plan around that table. First, these are company pages, and personal profiles behave differently. Second, format is correlated with effort. Native documents perform well partly because almost nobody makes a bad one casually, whereas anyone can post a link in four seconds.

Still, the direction of travel is clear and it matches what I see in client accounts. Formats that keep people on LinkedIn beat formats that try to send them somewhere else.

Why your LinkedIn impressions have dropped

If your impressions are down over the last two years, you are in very large company, and the cause is mostly structural.

An independent annual study of 1.3 million posts from 50,000 creators reported that reach for active creators fell around 60 per cent over two years, while engagement fell only about 20 to 25 per cent. Read those two figures together and the picture is clear: fewer people are seeing each post, but a larger proportion of those who do see it are responding.

That happens for a straightforward reason. More people and more companies are posting than ever, and a great deal of what they post is now generated rather than written. The amount of attention available has not grown to match. So the same finite pool of feed time is divided between far more content, and every individual post gets a thinner slice.

There is a second cause worth knowing about, and it is more recent.

How the LinkedIn algorithm decides who sees your post

In March 2026 LinkedIn published a detailed account of rebuilding its feed, and it is the most consequential change to how content is distributed in years. If you only take one thing from this article, make it this section.

Retrieval and ranking

The old system used several separate specialised models to work out which posts you might want to see. LinkedIn replaced that with a single retrieval system built on a large language model, followed by a ranking model it calls a Generative Recommender.

The practical effect is that the feed now works on meaning rather than on matching signals. LinkedIn's own explanation is that "LLM-based retrieval understands these topics are semantically related because the underlying language model brings world knowledge", and the example the engineering team give is connecting someone's interest in electrical engineering to a post about small modular reactors, without either piece of content sharing a keyword or a hashtag.

The ranking model reads more than a thousand of your past interactions as a sequence rather than as a list of separate events, so it treats your behaviour as a trajectory of interests over time. LinkedIn also says posts now surface "within minutes, not hours", and that the new system is better at reducing engagement bait and recycled thought leadership.

Three things follow for anyone posting.

Topic consistency now compounds. If your posts consistently sit in one professional territory, the model has something coherent to match against. If you post about six unrelated things, you are diluting the signal that gets your content in front of the right feed.

Hashtags matter even less than they did. Semantic retrieval does not need you to label your content. Write clearly about a subject and the label is implicit.

The system is harder to game. The previous approach could be nudged by stacking behavioural signals. A model that reads meaning is considerably less impressed by tactics.

Dwell time

The other signal worth understanding is older and quietly more important than the reactions counter.

LinkedIn's engineering team have written openly about using dwell time to improve feed ranking. They measure two kinds: dwell time on the feed, which starts counting when at least half of a post is visible, and dwell time after a click.

Their reasoning for moving beyond clicks and reactions is worth quoting, because it explains a great deal about what does and does not work:

Click and viral actions can be rare, especially for passive consumers of the feed... Click and viral actions are primarily binary indicators of engagement... Clicks are noisy indicators of engagement. For example, a member may click on an article, but quickly close out, realizing it's not relevant.

They built a model predicting the probability that someone will skip a post, on the principle that "members value their time, and will spend it appropriately on feed content that they're interested in".

This is why a genuinely interesting post with modest reactions often outperforms a post engineered to collect likes. Time spent is the honest signal, and it is one you cannot fake with a network of people who agree to click.

How to get more impressions on LinkedIn

With all of that in place, here is what actually moves the number, roughly in order of how much difference it makes.

Post consistently, at a frequency you can genuinely sustain

Frequency is the single largest lever, and the data on it is unusually clear. An analysis of more than two million posts from 94,000 accounts compared accounts against their own baselines and found, relative to posting once a week:

• 2 to 5 posts a week: around 1,182 more impressions per post, and 0.23 percentage points more engagement
• 6 to 10 posts a week: around 5,001 more impressions per post
• 11 or more posts a week: around 16,946 more impressions per post, and roughly three times the engagements

Before anyone commits to eleven posts a week, two pieces of realism. That study controls for account-level differences but cannot fully separate cause from effect, since accounts that post a lot are often accounts with more to say. And the author of the independent creator study I cited earlier has actually lowered his own recommendation to about three posts a week, on the basis that quality now carries further than volume.

My own view, formed watching this across a lot of client accounts: two to three good posts a week from the people who matter, sustained for a year, beats daily posting that collapses after six weeks. Consistency is the point. The feed rewards a steady presence, and a gap of a month resets more than you would expect.

Choose the formats the feed is currently rewarding

Native documents, multi-image posts and video all outperform text alone, and all of them substantially outperform posts whose main purpose is to move someone off LinkedIn.

That is not a mystery. LinkedIn is a business that sells attention on LinkedIn. Formats that hold people in the feed serve that business, and formats that export people do not.

For professional services firms, native documents are the most underused of the three. A five to eight page PDF that answers one real question properly gives people a reason to stay on the post for thirty seconds instead of one, and dwell time is the signal that decides how far it travels.

Win the first hour

Early engagement decides how far a post goes. The commonly cited window is the first 60 to 90 minutes, and it functions as a test: if the people shown your post early respond to it, LinkedIn keeps showing it to more people.

The most practical thing you can do with that is be available. The independent algorithm study found that replying to comments within the first 30 minutes was associated with around 64 per cent more total comments and roughly 2.3 times the views. That matches what I see. Posting and then disappearing into a meeting for four hours wastes the window that mattered.

So publish when you can actually be at your desk for the following hour. That single habit is worth more than any amount of agonising over the perfect posting time.

Write for dwell time rather than for the like

If time spent is the signal, then the job of your writing is to hold someone for a few seconds longer than the post above it.

In practice that means the first two lines carry almost all the weight, because that is what shows before the "see more" cut. It means one idea per post rather than five, because a post trying to cover everything gives no reason to stay. It means short paragraphs and real line breaks, because a wall of text is scrolled past. And it means the payoff has to justify the click on "see more", because a post that opens with a tease and delivers nothing trains people to skip you.

I would also say something unpopular here. The formulaic openings that were effective three years ago, the one-line hook and the artificial curiosity gap, have been used so heavily that they now signal low quality to a lot of readers. A plain, specific first line about a real situation outperforms them in every account I have watched recently.

Treat comments as the metric that matters

Reactions are cheap and comments are not, and the platform's economics reflect that. Buffer's analysis puts the weight of a comment at roughly fifteen times that of a like, and found that posts where the author replies to comments perform around 30 per cent better across their lifecycle, a pattern that held for 83 per cent of the accounts they looked at.

There is a second-order effect that people miss. Commenting thoughtfully on other people's posts is one of the most reliable ways to grow your own reach, because it puts you in front of their audience with something useful to say. For a recruitment or search firm, twenty minutes of genuine commenting a day on the posts of people you want to work with is more valuable than an extra post of your own.

The link question, answered honestly

This one comes up constantly and the evidence genuinely conflicts, so I will give you both sides.

LinkedIn's position, stated publicly by a senior product director, is that posts containing links are not deliberately penalised, and that a post carrying value on its own will not lose reach for including one.

Independent analysis says otherwise. The most cited creator study found that a single external link in the body of a post reduced median reach by around 18.8 per cent, and some practitioner reports put the figure considerably higher.

Both can be partly true. It is entirely plausible that there is no link detector applying a penalty, and that link posts still underperform because a great many of them are thin, offer nothing on their own, and get skipped. That would show up in the data as a link penalty without one existing.

What I do, and what I would advise: write the post so that it stands alone and is worth reading even if nobody clicks anything, then put the link in the first comment and mention that it is there. If the penalty is a myth you have lost nothing. If it is real you have avoided it.

Personal profiles carry the reach, the company page carries the credibility

Personal profiles consistently outperform company pages on LinkedIn, and the difference is not small. People engage with people, and the feed reflects that.

This does not make your LinkedIn company page pointless, it makes its job different. The page is where a prospect goes after a person has caught their attention, to check that the firm behind them is real, current and credible. Left dormant, it actively undermines the individuals doing the work.

So the split I recommend for professional services firms is straightforward. Individuals carry the daily visibility. The page carries proof: recent work, the team, what the firm actually does. The two reinforce each other, and the humans lead.

Optimise your profile, because the ranking model now reads it

This used to be advice about looking professional to visitors. It is now advice about distribution.

Under the new ranking approach, your posts are evaluated alongside your professional profile, because the model is trying to work out whether you are a credible source on the subject you are posting about. A headline that says nothing specific, an about section left half-finished, and a work history with no detail all weaken that assessment.

LinkedIn profile optimisation, in this context, means something narrower and more useful than it usually does. Make your headline state what you actually do and for whom. Make your about section cover the same territory your posts cover. Keep your current role accurate and described. You are giving the model evidence that you are a reasonable person to show on this topic.

Best time to post on LinkedIn

I have put this near the end deliberately, because it is the smallest lever on this page and it gets the most attention.

The consensus across the major published studies, for a UK B2B audience, lands on Tuesday to Thursday, mid-morning, roughly 10am to midday. LinkedIn's own guidance points at weekdays with mid-morning and lunchtime performing best on Tuesdays, Wednesdays and Thursdays.

That said, the studies disagree with each other more than their confident headlines suggest. One large analysis of 4.8 million posts puts peak engagement in the late afternoon and evening, between 3pm and 8pm, which is close to the opposite of the mid-morning consensus. When rigorous studies of enormous datasets reach different conclusions, the sensible reading is that timing is a weak effect swamped by everything else.

The useful version of this advice is much simpler. Post on a weekday, at a time when you can be present for the next hour to reply to comments, and check your own analytics after a couple of months to see when your particular audience is actually online. Your audience is not the average of 4.8 million posts.

What not to do: pods, bait and bought engagement

There is a whole economy built around inflating this number artificially, and it has become a genuinely bad bet.

LinkedIn's Professional Community Policies are explicit: do not do things to artificially increase engagement with your content, and do not agree with others in advance to like or reshare each other's content. That covers engagement pods squarely.

In November 2025, LinkedIn's VP of Product Management set out how they intend to enforce it:

Our goal is to make engagement pods entirely ineffective. We are increasing the number of ways we detect these pods and the suspicious behavior that happens in these pods.

They also said they would limit the reach of content flagged this way, and crack down on third-party browser extensions and plug-ins automating the manipulation. Flagged content typically stays visible to existing followers but loses the recommended distribution that reaches anyone new, which is precisely the distribution the pod was supposed to buy.

Set the rules aside for a moment, though, because there is a more practical argument. A ranking system built on dwell time and semantic relevance is not measuring the same thing a pod produces. Twenty people clicking like within four minutes generates a burst of shallow activity and almost no time spent. Under the old signal-stacking model that was worth something. Under the current one it looks exactly like what it is.

The same logic applies to AI-generated comments, which the independent creator study found made up around 80 per cent of the comments arriving in the first five minutes on the author's own posts. They add nothing a human values and they are increasingly obvious to readers, which is a reputational cost on top of everything else.

Turning impressions into pipeline

Here is the part that matters most for anyone running a business rather than building an audience.

Impressions are worth chasing only if they land on the right people, and even then most of those people will not be ready to buy anything. Research conducted for LinkedIn's B2B Institute by Professor John Dawes at the Ehrenberg-Bass Institute produced the figure that reframed this for me: businesses change providers of services like banking, legal advice or software roughly every five years, which means only about 20 per cent of a category's buyers are in the market in any given year, and around 5 per cent in any given quarter.

That has two consequences and they pull in opposite directions, which is why people get it wrong.

The first is that reach to people who are not currently buying is not wasted. It is the entire mechanism by which you become the firm they think of when the buying window finally opens. A post that generates no enquiries this month is not a failed post.

The second is that this only works if the reach is landing on the right 95 per cent. Impressions from an audience that will never hire you are not a long-term investment, they are a vanity number. This is why the viewer demographics panel matters more than the headline figure, and why I would rather have 1,500 impressions from directors at target firms than 15,000 from a general audience.

How I would report on a LinkedIn programme instead

If I were setting up monthly reporting for a firm today, impressions would be on the list but nowhere near the top of it. This is roughly what I would track:

The last row is the one people leave off, and it is the only one your finance director cares about. Ask every inbound enquiry where they first came across you. Written down over twelve months, that single question tells you more about whether LinkedIn is working than every analytics panel combined.

Where this fits

Impressions are a diagnostic, not a goal. They tell you whether your content is being distributed, and when they fall you now know how to work out whether the cause is the platform, the format, the timing or the writing. What they will never tell you is whether the programme is working, and treating them as though they can is the most common mistake I see.

If you want the layer underneath this, the decisions about what to post and who should post it, I have written separately on building a LinkedIn content strategy and on running a LinkedIn newsletter, which remains one of the few formats on the platform with a genuine distribution advantage.

And if you would rather someone else ran it, our LinkedIn marketing service covers profile management, content, outreach and company page programmes for recruitment and search firms, as one channel inside a wider marketing function rather than as a standalone activity. Either way, the advice above holds. Judge the work on conversations, not on the number under the post.