LinkedIn's Feed Now Favors Strangers Over Your Network, and Most B2B Founders Still Write Like It's 2019

Your impressions are down. Your follower count is up. You've noticed the gap and you've probably blamed the algorithm without understanding what actually changed.

Here's what changed: LinkedIn stopped organizing its feed around your network and started organizing it around topics. That's not a minor ranking tweak. It's a different distribution model, and it means the post you wrote for your 4,000 followers is being evaluated against people who have never heard of you.

The Shift, Specifically

According to Saleshigher, the pre-2024 system ran on the social graph: content spread through who you know, your connections' engagement determined your reach, and relationships mattered more than relevance. More connections meant more reach, full stop.

The 2025-2026 system runs on the interest graph instead. LinkedIn now groups people by what they talk about, not who they're connected to, and relevance to a topic can outweigh whether you've ever interacted with that person before.

This lines up with how Tinuiti describes LinkedIn's actual ranking mechanics. Once a post clears the initial quality bar, LinkedIn decides how far to expand it using three signals: personal connection, interest relevance, and engagement probability. Personal connection is only one of three factors, and it's no longer the gatekeeper it used to be. A post can score low on "does this person know you" and still get pushed hard if it scores high on "does this person care about this topic" and "will this person engage."

Translate that into practice: LinkedIn is actively routing your content to strangers who share an interest, not just to the people who already trust you. If your writing only makes sense to someone who already knows your company, your offer, your inside jokes, you're optimizing for the smaller, shrinking half of your potential audience.

The Numbers Back This Up, and They're Uncomfortable

Scott Leese posted his own year-over-year data with a 118k-follower account, and the pattern is instructive. In 2025, across 66 posts, his average impressions were 19.1k, down 38% year over year. But average reactions were up 10%, average comments up 29%, and engagement rate up 14%. Reposts, the most network-dependent action, fell 21%.

Read that carefully. Reach is collapsing while engagement quality is rising. That's not a sign the algorithm hates you. It's a sign the algorithm is being more selective about who it shows your content to, and rewarding you when it picks well. Fewer, more qualified eyeballs, then a decision about whether to expand further based on what those eyeballs do.

Entrepreneur.com has been covering the broader organic reach decline directly, framing it as something business leaders need an active response to, not something to wait out. The reach isn't coming back to 2019 levels because the system isn't the 2019 system anymore.

The Objection Worth Taking Seriously

The obvious pushback: "My existing network is my buyer base. Why would I optimize for strangers who'll never buy from me?"

Fair, if you're a niche B2B seller with a genuinely small addressable market and your current connections already are that market. In that specific case, the interest graph does you fewer favors and personal connection signal still matters more.

But for most B2B founders and sellers, your current 1st and 2nd degree network is a fraction of your actual buyer pool. The interest graph is LinkedIn handing you access to buyers you'd never have reached through your own network alone, on the condition that your content earns it on topic relevance and engagement probability rather than relationship history. Treating that as a threat instead of a channel is the actual mistake.

What Changes on Monday

Stop writing posts that only parse for insiders. Practically:

• Open with the problem or claim, not the backstory. A stranger has zero patience for three paragraphs of context before you get to the point.
• Name the topic explicitly in the first two lines. If LinkedIn is matching on interest relevance, give it something unambiguous to match on.
• Write comments like they're posts. Per the advice circulating from creators like bookmarklee, comments are often what make people remember you, and the trick is writing them for everyone reading the thread, not just the original poster. Comments are also cheap, repeated proof to LinkedIn that you're engaging in a topic space, which feeds the same interest signal your posts rely on.
• Be genuinely active in others' posts, not just your own feed. LinkedIn itself recommends commenting on other people's content as baseline activity, according to advice shared by Luke Coburn, and it's a two-way signal: it tells the algorithm what you're interested in, which is exactly the graph it's now using to route your own content.
• Expect lower impressions and don't panic. If your reactions, comments, and engagement rate are climbing while impressions fall, as Scott Leese's data shows is possible, you're not losing. You're being filtered toward people who actually care, and that audience converts better than a bigger, colder one ever did.

The founders who keep writing for their existing network in 2024 and beyond aren't wrong that their network matters. They're wrong that it's still the whole game.