How LinkedIn Really Works: the Algorithm, Reach and B2B Sales Without the BS

On the second day of Infoshare, right after my talk, Paweł Szymków caught me for a quick chat in the hallway. He asked the question I probably hear most often: isn’t LinkedIn just a „mutual admiration society”? I answered on the fly, a bit rushed, so I’ve tidied that conversation up here at a calmer pace. If you’d rather have the original, the whole thing is on video — and under specific threads I drop links to the moments where I talk about them.

This isn’t a piece „about how to hack the LinkedIn algorithm”. It’s a piece about why most of what people take for hacking stopped working long ago — and what actually does work.

Where the „mutual admiration society” came from

Let’s start at the source of that belief, because it didn’t come out of thin air. A few years ago there was a craze for what I call „secret mutual-support groups on LinkedIn”. The rule was simple: members agreed to like and comment on each other’s posts to „trick the algorithm”. The mechanism resembled an MLM pyramid: at the top sat the organizer, who dropped content into dozens of such groups at once. It was enough for one person from each group to react, and he’d have a hundred likes and a hundred comments under his post.

The problem was that it was easy to spot — the same people kept showing up under the posts, and the comments were hollow fillers like „great training!”, „brilliant expert!”. It made no difference whether the praise came from a car-wash owner or an optician — the signal was artificial and visible to the naked eye. The further down the hierarchy you went, the stranger it got: you’d run into situations where a coach for CEOs was getting likes from random people completely outside her world.

At first LinkedIn couldn’t handle it, especially since in English-speaking circles people quickly started creating so-called pods made up only of people connected to a given topic, so it looked plausible. The problem grew when, with the rise of AI, the creators of such groups took members’ logins and bots did the liking on people’s behalf.

LinkedIn, however, dealt with it elegantly, as it described on its blog in March: news.linkedin.com/2026/ImprovingTheFeed.

It didn’t block these groups. It said, more or less: you want to like each other’s posts? Be my guest. Only I’ll show that content exclusively within your group, not to the rest of the world. The result? People locked themselves inside their own bubbles, and their real reach dropped (I talk about it here). Mutual admiration societies still exist — it’s just that today they harm themselves, and the reach of posts by Grzegorz Turniak or Marcin Banaszkiewicz, who ran such groups (and maybe still do), has very clearly fallen. Maybe they should speak up for themselves?

The new algorithm thinks in „tribes”, but not the Heroes kind

In April LinkedIn announced major algorithm changes and introduced a concept it calls tribes. I translate it into Polish as „plemiona” („tribes”) and I think it’s a good translation (I explain the concept here).

The idea is simple: LinkedIn builds a profile of your behaviour. It watches what you click, what you search for, and — most importantly — whether what you read on LinkedIn is consistent with what you publish. The algorithm rewards consistency and penalizes hopping from topic to topic.

The best example: you’re a marketing person and you suddenly post a funny meme about „Heroes of Might and Magic”.

Two years ago a meme like that would have flown. Today it’ll be rated lower, because it doesn’t fit the bubble you’ve fallen into — the algorithm sees a „sideways jump” and treats it as noise. Of course a good meme will still travel further than a good, substantive article, but no longer the way it used to.

And here comes the sentence that caused the biggest stir during the presentation: it isn’t you who decides the reach of your post on LinkedIn. That decision was made before you clicked „Publish”, because it was made on the basis of how, over many weeks, the algorithm learned you. The post is just the trigger. The rest is the consequence of your earlier consistency. Look at the slide from my Infoshare talk — do you see why focusing on the post itself, instead of on building „your bubble”, is a mistake?

There’s also a feature copied from Instagram and TikTok — on LinkedIn it’s called „Suggested posts”. Every now and then, between content from your network, a post appears with a big „Suggested post” label. If you click it, you’ll see a description of why LinkedIn decided it suits you. And what’s written there is the most important part — because it’s a literal description of your „tribe”.

How to check whether you’re in the right bubble

This isn’t theory. I got burned by it myself (I tell that story here).

When these changes were rolling out, I noticed my reach suddenly collapsed and odd people started engaging with my content. I checked why — and it turned out the problem was in my profile. For years I ran e-commerce and digital at large corporations, but the one thing that kept repeating in my professional experience was lecturing: on MBA programmes, at universities, on postgraduate courses. I listed every university where I lectured, thinking it built trust in me. For the algorithm, that was the strongest, most repeatable signal. So it reclassified me from „digital marketer and consultant” to… „academic lecturer”. I removed those mentions from my profile and started getting back to where I want to be.

You have two ways to check your „tribe”:

1. Look into your suggested posts and ask yourself: is this content that you yourself would want to publish? Is this your world and „your bubble”? If you see something totally „out of left field” there, it’s a sign the algorithm has filed you in the wrong place.
2. Run a Social Selling Index analysis. I’ll be honest: as a measure of LinkedIn success, SSI is a vanity metric to me. It’s a tool LinkedIn mainly uses to sell Sales Navigator — and LinkedIn itself admitted as much by hiding the analysis. But it has one real upside: at the very bottom of the analysis you’ll find a line about which industry you’ve been assigned to. It says outright: „we think you’re in such-and-such an industry”. Check it and compare it with where you actually want to be: linkedin.com/sales/ssi.

Artificial clicking is a foot-gun these days

Someone will say: if reactions help, let’s build ourselves a reaction machine. Well, no. It’s exactly the same mechanism as white text on a white background in old-school SEO — a stretch that the algorithm now catches with ease.

I recently came across someone who set up 300–400 fake accounts and used them to click on their own posts. From the outside it looked decent — a hundred likes is impressive. But such an artificial, simultaneous spike (a hundred people coming in within the same minute from a link dropped on WhatsApp) is trivially easy for the algorithm to detect and is more likely to hurt than help (I talk about it here).

That doesn’t mean you should go it alone. I think everyone should have their own genuine support group — just handled with care. In my LinkedIn CRM (I’m an Excel person, so I missed being able to tag contacts — and now it does much, much more) I keep tags like „supporter” or „don’t reach out about this”. If I have a group of 35 people I message once every six months with „we’ve got an important campaign, please support it” — that works. But if you start treating that group like a daily newsletter and writing „I posted, go like it!”, you’ll burn it out. The difference is between a relationship and spam.

„How are you?” — automation that isn’t spam

Since the algorithm rewards real interactions, it’s worth keeping them up — but in a human way. I worked a lot in the Arab world, where there’s a beautiful, deeply cultural „kif halak” — meaning „how are you”. Except there it’s a ritual: if I know you have a wife, kids, and the kids have a dog, then I should start with „how’s the dog, I heard it got hurt recently?”, then „how’s the younger child?”, and only at the very end „how are you?”. If I barged in with „we’ve got business to do”, I’d be treated as a boor. If I care about my relationship with someone and I remember their kids’ names, it’s worth building lasting relationships on that — it shows I care about the other person.

I’ve carried that over to myself. Every so often I send selected people in my network a simple message: „How are you?”. I’m not trying to sell anything, I’m not trying to buy anything — I’m simply interested. I do it systematically (I have a rule I try to keep: I reach out to every contact at least once every six months) and it works great for two reasons. First, it builds the relationship. Second, since we’re corresponding, the algorithm treats it as a signal and more often shows each of us the other’s content. The key is to automate the reminder to reach out, not the content of the relationship.

Why be on LinkedIn at all? Four concrete reasons

I know plenty of entrepreneurs who have a profile only because „you have to in business these days”, and either don’t run it or once in a while have ChatGPT churn something out. For them I have four arguments I use in my trainings (I start on this here).

1. You win in Google. When did you last google someone? Every day we check salespeople, candidates and partners this way. And ever since LinkedIn data has been well indexed in Google (not just Bing), profiles rank superbly. I have a common surname — I compete with a tall-ship captain and a member of parliament — and yet my LinkedIn account ranks higher than my own domain. And very often a profile is the only place where you have full control over what the world reads about you.
2. You stay in the AI answers. LinkedIn has never hidden that it trains AI models on platform data — users gave consent (in the fine print) until February this year. Since LinkedIn belongs to Microsoft, and Microsoft is closely tied to OpenAI, the content you publish there has a real chance of showing up in ChatGPT or Gemini answers. I’ve started doing it like this: first I publish something on my own site, wait for it to get indexed (to avoid duplicate content), and only then post it to LinkedIn. That way I win two Google links for the same thing (different domains) and gain extra citations.
3. You know who you’re doing business with. On LinkedIn people as a rule use their real name and company. Fake accounts exist, but not on the scale you see on Instagram or Facebook. That genuinely changes the quality of contact.
4. You have access to hard B2B data. Here’s a question that surprises everyone in trainings: how many types of accounts are there on LinkedIn? Most people guess „a few”. The answer is 27 (I talk about it here). Five are ordinary accounts (All-in-one, Business, Job Seeker, etc.), but the rest are a whole ecosystem of sales and recruitment tools. There are accounts whose sole purpose is analysing the sales potential of your product, dividing territories among salespeople (everyone knows the problem that whoever got the capital region has it easier than whoever got the rural one), finding candidates or assessing market potential.

A concrete example: we ran a project opening service centres for a Swedish company that wanted to enter Poland. The natural choice was the Tri-City — close, the ferry comes in, job done. But the tool showed that in the Tri-City they wouldn’t recruit staff who speak Swedish, whereas there’s a university in Wrocław that trains people in that direction. As a result, Lower Silesia turned out to be the better choice. That data genuinely changed a business decision.

And here’s an important thing to understand: the fact that LinkedIn encourages you to fill in your profile thoroughly isn’t altruism. It’s a carrot. The better, verified data you enter, the better the „merchandise” they have in tools like Sales Navigator or Recruiter.

Where the „30% sales uplift” from Sales Navigator comes from

On my website there’s a promise that Sales Navigator increases sales by about 30%. The natural question is: isn’t that BS? Let me tell you where that number comes from (I explain it here).

I’ve worked with Navigator since its beta, roughly from 2015–2016, back in my corporate days when I was responsible for digital marketing in the CEE region at Philips. Even then you had to learn to report to the board whether such a tool made sense at all. And that’s how I learned how to calculate its real return.

The way to count it is simple and honest. Don’t roll Navigator out to all one hundred salespeople at once „in a frenzy”. Set up a small group of social sellers with Navigator and leave a control group of sceptics who work the old way. After a quarter or two, compare them on ordinary business KPIs. It turns out the Navigator group has more sales opportunities in the CRM — because they simply gained another channel for reaching people. And they close deals faster — because sometimes it’s easier to nudge a deal with a message („Mr Smith, I’ve already sent a few emails, maybe they’re not getting through?”) or, the other way round, use email to remind someone about a LinkedIn invitation. In practice that difference is usually somewhere between a dozen and over twenty percent more opportunities — and that’s where the order of magnitude „about 30%” comes from.

It’s not magic. It’s a plain A-versus-B comparison: you know the cost of the employee, you know the cost of the tool, you know whether the deal got done — so you can calculate the return. A good head of sales can always run that comparison.

By the way — the same long-term thinking is where my LinkedIn CRM under the B2B Marketing AI brand came from. I bought the domain a decade ago and I consider it one of the better decisions I’ve made, because I’ve recently had a few offers to sell it for genuinely nice money.

Where to really start: comments, not posts

The most common roadblock in trainings is: „I really want to start, but publishing is beyond me, I don’t know where to begin”. I have a concrete answer to that, and it isn’t „write posts every day”.

Start with observation. Pick people from your industry or the competitors you aspire to, and start watching what they do. If they don’t have their connections hidden (on LinkedIn that’s a single toggle — which is why it’s worth flipping it), you can peek at who they have in their network and who comments under them. That’s your map of the industry.

And then — and this is the crux — start with comments, not your own posts (I show it on data here). For two reasons. First, you gain visibility in front of someone else’s already-built audience. Second — and this surprises most people — comments on LinkedIn have disproportionately large reach relative to posts. With a good comment you can easily rack up over ten thousand views, the kind you’d have to sweat hard for under your own post. You don’t have to invent everything from scratch — you step in where the conversation is already happening.

One caveat though: not all comments are equal. Those old „admiration circles” relied on one-line spam like „great post!”, written with no knowledge of the content. That builds nothing.

The anecdote that explains everything: how I won with a comment

To finish, a true story that shows what this is really about (I tell it here).

We know that people read headlines, not the content. Paweł Tkaczyk once leaned into this deliberately. He wrote a very substantive blog article, but on LinkedIn he gave it a clickbait, deliberately controversial title — a thesis that was nonsense, but one the article itself debunked. He counted on people clicking in, reading, and figuring it out for themselves. Except nobody clicked. Everyone commented on the title alone, along the lines of „Paweł, you must have got it wrong”. It travelled, but not the way he wanted.

I had a rule back then: fewer comments a day, say three to five, but in-depth. So I actually went into that article, read it to the end, and wrote a comment starting from its last paragraph: „Paweł, I see you wanted to make clickbait, because you wrote this — but in the text you yourself explain that it’s different, so my view is this and that”. The result? Paweł started showing my comment in his trainings as „the one person who actually read it”. It stuck. Although Paweł in turn has joked a few times about various pieces of my LinkedIn advice — e.g. when I said it’s worth writing your hook so people want to click „See more”.

And that’s the whole philosophy in a nutshell — no magic, no hacker tips. If you keep repeating revealed truths that everyone already knows, you won’t build a position as an expert in any field. But if you run knowledge through your own experience — show where you fell over yourself and how you dealt with it — people will appreciate it. Because they see that you don’t just know, but that you actually do it.

What’s in it for you and what you can do next

If you’re to walk away from this long article with an action plan, here it is:

• Check which „tribe” the algorithm has assigned you to — via suggested posts and via the industry at the bottom of your SSI analysis. If it’s not your world, fix your profile.
• Drop the artificial clicking and closed circles — these days they hurt more than they help.
• Build yourself a list of people in your industry to follow, and peek at their network.
• Start with in-depth comments, not posts — they have greater reach and a lower barrier to entry.
• Keep relationships up in a human way („how are you?”), rather than treating contacts like a mailing list.
• If you sell in B2B, test Sales Navigator on a small group and calculate the return honestly, factoring in HR, training and sales KPIs. I can help with that.

It all starts with one decision: stop treating LinkedIn like an unpleasant chore, and start treating it like a tool you simply have to understand.

I also encourage you to book a LinkedIn consultation or LinkedIn / Sales Navigator training.

This article is based on my conversation with Paweł Szymków recorded at Infoshare. Watch the full recording here.