What Content Actually Performs Well on LinkedIn

Ask ten creators what works on LinkedIn and most will answer with a format. Carousels. Text posts. Personal stories. Job announcements, oddly enough.

The format answer feels useful because it is easy to act on. Post a carousel tomorrow, see what happens. But format turns out to be the wrong variable to test first. The same idea, told two different ways, can perform completely differently depending on something that has nothing to do with images, video, or word count.

The variable that actually matters

Content that gets ignored tends to share one trait: nobody could disagree with it. A tip like "post consistently" gets scrolled past because everyone already agrees, so there is nothing to react to. The same idea reframed as a mild challenge, something like "consistency is overrated, the people who actually grow post in bursts around real insight instead of on a schedule," creates friction. Friction is what gets comments.

This is why format tests are misleading. Two memes about the same industry can perform completely differently. One says something generic. The other captures something true and slightly uncomfortable that people recognize from their own work. The format was identical. The stance was not.

Specificity does similar work from a different angle. A post that names a real number, a real tool, a real cost, or a real mistake reads as lived experience rather than content marketing. Vague thought leadership, the kind that could have been written about any industry by anyone, tends to underperform even when it is polished. Several creators pointed to the same pattern: the more a post sounds like a marketing team signed off on it, the worse it does.

Specificity without a point of view just reads as a detailed observation. A point of view without specificity reads as a generic hot take. Put them together and the post reads like a real person talking directly to someone who understands their situation, which is the thing LinkedIn readers actually respond to.

Why job posts and memes are a trap

Job posts genuinely get disproportionate engagement on LinkedIn, whether or not the account is hiring for anything. It says something about the platform's incentive structure, not about content strategy. High reach on a job post does not translate into an audience that remembers you next month, and several creators noted that the engagement rarely holds up outside the format itself.

Memes work, but only under a narrow condition: they need to tie back to something real, a service, a lesson, a case study, a genuine point about the industry. A meme posted for its own sake gets likes and nothing else. A meme that lands a real observation about the audience's day-to-day work performs closer to a well-written post, because the joke is doing the same job specificity does elsewhere: it signals that a real person who understands the reader's situation made it.

Comment quality beats comment count

One pattern came up repeatedly and rarely gets mentioned in format guides: five comments from people who actually match the target audience matter more than fifty comments from people who will never buy, hire, or follow through. A post with modest total engagement but the right commenters is often doing more real work than a viral one full of passersby.

This reframes what "testing" a post should even measure. Reach and like count are the easiest numbers to check and the least useful ones. Who showed up in the replies, and whether those people look like the audience the account is trying to reach, is the number that actually predicts whether the content is working.

What this means for what to post next

None of this argues against structure. Story posts, tip lists, and case studies still give a post shape, and shape helps readability. But structure is scaffolding, not the reason a post performs. The scaffolding holds up a specific, opinionated idea grounded in something that actually happened. Remove the opinion and the specificity, and the best-built scaffolding in the world still reads as filler.

The practical takeaway is less about picking a content type and more about a filter to run before publishing anything: does this post say something a reasonable person could push back on, and is there a real detail in it that could not have been written about any other account in the niche. If the answer to either is no, the post needs another pass, regardless of what format it is in.

That filter is also, not coincidentally, the same thing an identity-driven writing process is built to protect. Generic AI content tends to fail this test by default, because it optimizes for sounding plausible rather than sounding like someone with an actual stance. Content that starts from a real voice, a real set of opinions, and real specifics tends to pass it without much extra effort.