LinkedIn went through a bad period from roughly 2020 to 2023. The platform was overrun by formulaic posts: the broken-line “hook” format, the cry-for-attention origin stories, the rage-bait career advice posts, the AI-generated motivational quotes. Real practitioners largely retreated from the platform because the feed read as a parody of itself.
Something has shifted since. Not all the way back to professionalism, but enough that the platform is again a useful place for someone with real expertise to build a following. The algorithm has rebalanced. The audience has gotten more discriminating. The cost of looking like a content marketer has gone up. And the people doing substantive work have started returning.
This piece is for the practitioner, founder, or executive who wants to use LinkedIn now as a place where their actual ideas reach the right audience. It is not a guide to gaming the platform. It is a guide to using it for what it is actually useful for.
What changed in the algorithm
LinkedIn’s feed ranking went through several visible changes in 2024 and 2025. The platform reduced reach for posts with high engagement-to-substance ratios (the broetry posts, the rage-bait, the obvious clickbait), increased reach for posts that produced sustained reading time and substantive comments, and started weighting “dwell time” and saves more heavily than likes.
The practical effect: a 350-word post with a clear argument and a few thoughtful comments now outperforms a 50-word hook-and-payoff post that gets 200 superficial likes. The platform is rewarding what it should have always rewarded.
The other shift is the rise of newsletters. LinkedIn made newsletters discoverable in a way articles never were, and started cross-promoting subscriber notifications via email. A practitioner with a strong angle can build a 5,000-subscriber LinkedIn newsletter in six months without any external promotion, just by writing well and consistently.
What still does not work
Some patterns are dead even though some accounts still try them. Worth naming clearly so they do not creep back in.
The broetry hook. “I quit my job at 32. Three months later I made $1M. Here is how.” Single sentences with line breaks between them, building toward an unsubstantiated payoff. The audience has caught on. These get downranked algorithmically and ignored by readers.
The fake humility post. “I just got rejected by [prestigious company]. Here is what it taught me about resilience.” The pattern is so worn that authentic versions of this story now read as performative. If the rejection actually happened, it is fine to mention it once, briefly, but not as the structural anchor of a content piece.
The “engagement bait question” post. “What is the one piece of advice you would give your 22-year-old self?” These produce comments without producing substance and the algorithm has caught on. Reach has collapsed.
The AI-generated post that pretends to be human. The patterns are visible: certain transition words, certain rhythmic structures, certain banned words like “leverage” and “unlock” used unironically. Readers detect this in seconds, and the algorithm has signal for it too. AI-assisted writing where a human shapes the final prose is fine. AI-generated writing posted directly is a credibility tax.
The naked self-promotion post. “Excited to announce…” followed by a press release. These are tolerable once a quarter when there is real news. As regular content they kill engagement.
What works in 2026
Several patterns produce reliable engagement and reputation effects on LinkedIn now.
The specific work-based observation. A post that describes something concrete you noticed in your work this week, framed as a small insight, written in clear prose. Examples: “I have been A/B testing email subject lines for our SaaS product for three months. The pattern that surprises me is…” or “We rolled out a new pricing structure last quarter. Three things we got right and one we got wrong.” These posts work because they signal that the writer is actually doing the work, not just commenting on the work others are doing.
The contrarian-but-substantive take. A clearly stated view that goes against industry conventional wisdom, supported by real reasoning. Not “X is dead, here is why” rage bait, but actual disagreement based on real experience. “I keep seeing advice that founders should focus on retention before acquisition. In our category, that has not been right for us. Here is what changed our thinking…”
The case study post. A short walk through a specific situation: the problem, the options considered, the decision made, the outcome. These read as practitioner content even at 300 words and get saved heavily by people who want to come back to them.
The methodology critique. A clear argument against a popular framework or technique, with specific examples of where it falls short. This kind of post attracts the right audience because the people who engage with it are the ones who care about the substance, not the entertainment.
The honest reflection on something that did not work. Failure posts have become cliché but the cliché is the performative version. A real reflection on a specific decision that did not work out, what was learned, and what would be done differently, still works because it is rare to read and useful to other practitioners.
The synthesis post. Pulling together threads from your reading or thinking on a topic that has been circulating, with your own framing. The synthesis post works when the writer actually has a point of view, not when it is just summarizing without adding anything.
The structural patterns that actually work
A few structural moves produce reliable engagement without becoming formulaic.
Open with a specific observation, scenario, or claim. Not a question, not a hook line, not a cliffhanger. A real opening sentence that tells the reader what the post is about.
Make the post 200 to 400 words for most pieces. Long enough to develop a thought, short enough that LinkedIn readers will finish it. Longer pieces (700-1,500 words) work as newsletters or articles; they do not perform as well as feed posts.
End with something that invites continuation rather than performative engagement. A reflection, a question that signals you actually want to hear answers, a tease toward future writing. Not “what do you think?” because that is the engagement-bait pattern. Something like “I am curious whether others see this differently in their own work.”
Use line breaks to make the post readable on mobile. Three or four sentences per paragraph. The platform is mobile-first and walls of text do not get read.
Avoid hashtags or use one or two only. The hashtag layer on LinkedIn does almost nothing in 2026 and makes posts read as marketing. Save them for the specific cases where they help (a campaign, a tracked event).
What to write about, specifically
The posts that compound across a year of writing tend to fall into a few buckets.
The work you are actually doing right now. If you are leading a product launch, write about the decisions inside it. If you are running a hiring process, write about the patterns you are seeing. If you are doing customer interviews, write about what you are learning. The audience for “the work you are actually doing” is large because it is rare.
The frameworks you have built or reshaped. Most senior practitioners have a handful of mental models they actually use that are not stock business school content. Writing those up specifically, with the cases where they apply and where they break down, builds a body of work that becomes citable.
The people you have learned from and what specifically you took. Not a generic “five mentors who shaped me” post. A specific story about working with one person and the specific thing you took from them. These posts compound goodwill (the named person often resharing) and signal that you have a lineage in your work.
The questions you are sitting with. Posts that frame an open question and walk through your current thinking, rather than declaring an answer, often produce the highest-quality replies. Senior practitioners do not have to perform certainty all the time, and the posts that admit uncertainty often build more reputation than posts that perform confidence.
The newsletter strategy
A LinkedIn newsletter is the highest-leverage move on the platform now for someone serious about thought leadership. The mechanics: pick a topic narrow enough to own and broad enough to write 30+ issues on, publish on a fixed cadence (weekly or biweekly), use a consistent format, write substantive issues of 800 to 2,000 words.
The algorithm promotes newsletters into the feed of subscribers and adjacent audiences. New subscribers get notified by email. The compounding effect over six months is significantly faster than building the same audience through individual feed posts.
Newsletters that work tend to share features: a clear narrow topic, a recognizable voice, real frequency, and a habit of citing other people’s work generously. Newsletters that do not work tend to be too broad, too inconsistent, or too self-promotional.
The cadence and time investment
A working cadence for someone serious about LinkedIn thought leadership: three to four feed posts per week, one newsletter issue per week or biweekly, regular substantive engagement on other people’s posts (15 to 30 minutes a day), and occasional longer articles (one a quarter).
Time investment is roughly five to eight hours a week for someone doing this well. Less than that and the work does not compound. More than that usually starts producing diminishing returns because the writer runs out of substance to share without inventing it.
The reputation effects of doing this consistently for 12 to 18 months are unusually large. The senior practitioner with a real LinkedIn presence is rare enough that the position itself is differentiated. Doors open. Inbound interest builds. Speaking invitations arrive. Hiring decisions get easier on both sides of the table. Not because the platform itself is magical, but because the audience that pays attention to substance is concentrated there now in a way it was not five years ago.