Why LinkedIn Reach is Down, and What to Do About It

Is it weird that I actually enjoy using LinkedIn?

It’s become the social platform I use most. I rely on it to find new clients. And it exposes me to strategic thinking from absolute strangers I might never see otherwise. (Which is how you keep a workaholic scrolling on a Saturday, BTW.) 

So it sucks that the platform feels like it’s all over the place this year.

And it’s not just me. Nearly everyone is suffering from algorithmic whiplash — one study of 400K posts shows median impressions are down a whopping 68% from their 2023 peak.

“More creators, more ads, more competition” doesn’t suffice to explain it.

But what does? There’s been no single source of truth — just various LinkedIn executives sporadically sharing tidbits of info. Which is how we wound up in a wilderness of conflicting stats, contradictory advice, and overconfident interpretations.

So about a month ago, I launched my own investigation.

I’ve combed over LinkedIn’s scientific papers and employee statements. Turned to a handful of smart people to help me better understand them. And took care that my sources were diligent about their sources — avoiding unverified claims from random gurus — while taking into consideration my own experience and that of my peers and clients.

Today I’m breaking down everything I’ve pieced together about how the algorithm works, and what it means for the marketers, like you and me, who use it. 

Hop aboard the magic school bus and into the black box we go.

Today’s LinkedIn feed is not a singular algorithm but actually 3 different ones working together. 

What it does: Rounds up all the most recent posts from your direct connections, plus creators and company pages you follow.

How it works: It has a strict 30-day memory. FishDB won’t surface anything older, no matter how brilliant it was, no matter how long it’s been since somebody logged in. 

Key takeaways: Go more than a month without posting, and you’re invisible in the feed. But also, just because FishDB pulls your post out of the water doesn’t mean anyone’s going to see it. We’ve still got two more layers of filtration to get through.

What it does: Analyzes your profile and the posts you regularly engage with to suggest similar posts from outside your network. 

How it works: It acts as a matchmaker, trying to figure out your exact professional vibe, then scouring the entire platform for the content you’re most likely to actually care about. 

Key takeaways: Every time you publish, Causal LLM checks the post against your profile. Does the topic match? Does the language match? If the answer is murky, it hedges — and a hedging algorithm means your post goes to fewer people overall (and those who do see it… may not be your intended audience). 

What it does: Makes the final call on what actually shows up at the top of your feed. The first two layers built a pool of 2,000 “candidate posts.” GR picks the winners.

How it works: It watches your recent behavior on the platform: what topics you’ve been sharing, which posts you’ve stopped to read and comment on, which ones you scrolled right past. Then it uses that pattern to predict what’ll keep you hooked.

Key takeaways: Not all engagement is equal. Shares and saves are weighed more heavily than likes. So are comments, because they take more than a second to produce. And a thoughtful comment beats a one-word “congrats!” — the algorithm prefers it when someone lingers on a post, crafting their response.

It also clocks who’s doing the commenting. If a healthcare CMO comments on your healthcare marketing post, the algo reads that as a strong signal and pushes the post out to more healthcare leaders. Which is great when it’s working, and a problem when it isn’t.

_

Here’s the big sea change all of this adds up to. The old algorithm rewarded engagement, the new one rewards coherence. 

That’s why you can write a brilliant post, but if it doesn’t align with your profile, your platform behaviors, or those of your intended audience — it gets buried at #400… and nobody is scrolling that far to find it.

Which raises the question: how do you make sure the algorithm can read you clearly?

Telling multipassionates to flatten themselves goes against everything I know about carving out a unique personal brand. (And I hate it.)  

Unfortunately, we’re no longer optimizing for people here. 

This new algorithm wants to file you into a specific corner of LinkedIn — healthcare comms, B2B SaaS, climate tech, whatever. And it needs to do that confidently, based on your profile. The more your profile points in one clear direction, the easier its job gets. (Apparently, human dimensionality confuses the hell out of the robots.)  

So it’s time to revisit your strategy. What are you actually using LinkedIn for? Who do you want to reach? What’s your brand’s goal? Anything that might dilute or complicate that identity has probably got to go.

For example, I’m removing “artist” from my headline (even though it’s a core part of my professional identity, and kills me) and rewriting my bio to include more language specific to social media marketing. But there’s a lot to test here, so I’ll report back.  

If we could ever afford the art of going off on a tangent, those days are over.

Regrettably, meandering is another very human trait that can make for a great story, but just confuses the algorithm. Remember, your every word is being measured for relevance — specific terms, concrete details, real expertise — and then checked against your profile.

To be fair, the occasional juicy anecdote might override these checks. But more often than not, I’d reserve more expressive posts for other platforms — LinkedIn doesn’t want them. 

And yet, it doesn’t want you to be robotic either. LinkedIn claims to be throttling AI-generated content — but as far as we know, their only means of catching it is through engagement signals (unless they’ve finally cracked the code on AI content detectors, which studies have shown to be unreliable). Which brings us to Step 3.

Now that you’re past the machine, you need early readers to not just consume your post — but interact with it.

I recently mentioned how LinkedIn is throttling “promotional” posts. It turns out they don’t just mean sales pitches, but anything that lands like a one-way broadcast, closing every loop and leaving the reader with nothing to do but nod and scroll.

So write like you’re dropping in on a conversation that’s already started. Pose a real question. Take a position someone might push back on. Share something half-formed and ask what others are seeing. Leave points unsaid so there’s room to contribute. 

And when the comments come in, respond — quickly, and with more than a two-word “thanks!” The more your post generates real back-and-forth, the further it travels. 

Everything else you might try to ‘game’ the algorithm — post format, links, hashtags — is moot until you’ve nailed this. Truly.

I’m putting all this to the test on my own profile and posts over the next month. (Are we connected yet?)

I aim to report back with more answers: What are the best ways to invite thoughtful comments, saves, or shares? And can you stay true to a multipassionate identity, or must we truly flatten ourselves to be seen?

Until then, let’s compare notes. What’s your team been seeing on LinkedIn? Hit reply and let me know. 

See you in two weeks,

P.S. Know your profile needs updating to keep up with these changes? I have 2 spots open this month to handle the audit and optimization for you. Reply LINKEDIN and I’ll send over the details.