Richard van der Blom has been studying the LinkedIn algorithm since before most recruiters considered it a serious business tool. His annual LinkedIn Algorithm Insights Report is now in its seventh edition, built from the analysis of 1.3 million posts published in the first half of 2026. It’s cited by Salesforce, PwC, and Nestlé, and has become one of the most widely referenced studies of how LinkedIn actually works.
His central finding for 2026: LinkedIn had what he calls “brain surgery.” The platform changed its core operating logic, moving from a relationship-based graph (which asked “who do you know?”) to an interest-based graph that asks “what do you want to be known for?” If LinkedIn can’t answer that question from your profile and your activity, it won’t put your content in front of the people you’re trying to reach. It won’t advocate for you at all.
To test this, Richard rebuilt his own profile. He stripped out everything no longer relevant to his expertise and services, deleted all his non-English language profiles, and rewrote his headline, about section, and work descriptions entirely around his core topic. Within three weeks, his reach was up 20 to 30 percent.
“If LinkedIn cannot classify you,” he said during this conversation, “it cannot advocate for your content. It cannot show your content to the right people.”
This episode covers the full picture: what topic fingerprinting is and how to build it, how to structure your content across four distinct pillars, which formats are outperforming right now, why the LinkedIn newsletter is one of the most overlooked tools for recruiters, and what happens to your reach when you rely too heavily on AI to write your posts.
Episode Outline and Highlights
- [01:06] How the LinkedIn Algorithm Insights Report grew from six pages to more than 300.
- [04:47] LinkedIn’s “brain surgery”: from relationship-based to interest-based graph.
- [06:24] Less reach, more engagement, and why that’s not necessarily a bad thing.
- [08:41] Topic fingerprinting: what it is and why it starts with your profile.
- [10:24] Richard redesigned his profile and got 20 to 30% more reach within three weeks.
- [11:17] Why every action you take on LinkedIn teaches the algorithm who you are.
- [12:14] How to cluster subtopics under your main expertise without confusing the algorithm.
- [19:19] The four content pillars: authority, affinity, proof, and demand.
- [27:03] Why LinkedIn is now actively hiding comments that contain links.
- [31:34] The new LinkedIn mobile video feed and what it signals for recruiters.
- [38:38] Format performance data: document posts, text-only, and the newsletter case.
- [43:27] The LinkedIn Loop Cycle: how to connect posts into a funnel.
- [48:05] AI detection on LinkedIn and the 30 to 40% reach penalty.
LinkedIn Changed the Game. Most Recruiters Are Still Playing the Old One.
For two years, reach has been declining for most LinkedIn creators. Richard’s data makes clear this isn’t random — it’s structural. LinkedIn moved from a relationship-based graph, which distributed content primarily to people who already knew you, to an interest-based graph, which distributes content based on topic classification. The question LinkedIn is now trying to answer, for every piece of content, is: does this person’s network and beyond actually care about this topic?
The concept Richard calls “topic fingerprinting” is how LinkedIn determines the answer. It looks back approximately 90 days, or around 1,000 actions, and classifies you based on what you post, what you engage with, what you comment on, and how your profile describes your expertise. If it can’t classify you clearly, it can’t show your content to the right people.
The practical consequence: posting on multiple unrelated topics dilutes your fingerprint. Commenting on content outside your main area does too. “It’s very important to understand that LinkedIn does make a difference between what they call relationship-based engagement,” he said. Congratulating a connection on a milestone is fine. Commenting substantively on content outside your topic area counts against you.
The profile is where to start. Your headline, about section, and work descriptions need to reflect not your career history, but your current expertise and who you serve. Richard’s own three-week result after rebuilding his profile is the clearest proof of how quickly the algorithm responds. Think of it less as a resume refresh and more as telling LinkedIn exactly what you want to be known for.
The Four Pillars and Why Most Recruiters Rely Too Heavily on One
Once LinkedIn understands your expertise, the next question is what you should actually publish. Richard’s data identified four types of content that account for 95% of high-performing posts on LinkedIn.
Authority content, including educational posts, frameworks, how-to content, infographics, and process explanations, should make up 40 to 50% of your total output. This is how LinkedIn now reaches outside your existing network. “Authority content on your topic is now sent to 50% outside of your network,” Richard said. It’s the pillar most likely to bring in people who have never heard of you.
Affinity content is personal storytelling — posts that let your audience connect with you as a person rather than a professional. Under the new algorithm, it reaches a smaller audience than authority content (LinkedIn limits it to connections and loyal followers, who are more likely to engage), but it builds the human layer of trust that makes everything else land harder.
Proof content, including case studies, client results, testimonials, and peer collaborations, sits in the middle of the funnel. It answers the question: Does this person actually understand my business? Aim for 10 to 15% of your posts here.
Demand content is where conversion happens. “One of the biggest myths,” Richard said, “is that if you optimize your profile and you are consistent in sharing content, leads will flow in from itself. That’s a myth.” Demand posts explicitly invite action: sign up, get in touch, book a call. Use this pillar every 15 to 20 posts. And when you do, put the link in the post itself. LinkedIn is now actively suppressing comments that contain links, in some cases hiding them entirely from view. What worked in 2024 no longer does.
Formats, the Newsletter, and the Loop Cycle
Text-only posts are performing better than last year, but they require genuine writing ability. LinkedIn has AI detection patterns that are becoming more sophisticated, and posts classified as largely AI-generated are being penalized by around 30 to 40% in reach. The approach Richard recommends: write your own draft, then use AI to suggest alternative hooks or improve the structure. Don’t outsource the thinking or the writing.
Document posts (carousels) are still outperforming at roughly 1.4 times the benchmark. They take longer to create, but the data supports the investment. Video is picking up again. LinkedIn’s mobile video feed is coming out of beta and rolling out to all users, which historically signals an algorithmic boost for video content. If you’re open to it, video builds trust faster than written content. If you’re not, don’t feel compelled — reluctant video rarely reads as confident.
The format most worth considering right now, for any recruiter with a consistent point of view: the LinkedIn newsletter. Newsletter articles average 25% reach relative to subscriber count, compared to under 6% for standard posts. Links inside newsletter articles aren’t penalized. And LinkedIn newsletters are now being indexed by AI tools including Claude and Perplexity. Richard has received multiple messages from people who found his content through AI responses, not through LinkedIn itself. That’s a distribution channel most recruiters haven’t thought about yet.
And finally, the LinkedIn Loop Cycle, a framework Richard introduced in the 2026 report for connecting posts into a deliberate sequence. A typical loop starts with a poll (gauging interest and building an engaged audience for what follows), moves to an authority post that goes deeper on the topic, then bridges to a resource or webinar invitation, and closes with a demand post. Each post in the loop builds on the engagement from the last. “It’s much more fun to write a book,” Richard said, “and then sell the book at the end.”
