✍️ how to hack the LinkedIn algorithm

If your LinkedIn posting has felt like shouting into a void, you're not imagining it, and you're not alone. Reach has been atrocious across the board — one widely-cited report clocked median reach down 74% year-over-year.

Your CMO probably wants to know why. (We did too.)

Our resident LinkedIn-fluencer and senior copywriter, Sarah Suzuki Harvard, went down the rabbit hole to figure out what's actually happening.

TL;DR: Sometime in the last 6 months, LinkedIn subtly ripped out the machines that counted your likes and replaced it with one that reads your posts. The feed stopped tallying engagement and started interpreting meaning. Which means the playbook you've been running since 2022 is now working against you.

(Also, they’re declaring war on AI-generated posts. Also, they launched a new metric. Don’t worry, we’ll explain it all.) 

Let's get into it.

For years, LinkedIn’s feed worked like a factory.

One system tracked engagement, another tracked connections, another evaluated content quality.

“The old system was like an assembly line with hundreds of tiny workers each doing one job,” explains Sarah. “This one counts likes, that one tracks connections. But LinkedIn’s engineers fired all of them and replaced them with a foundational model that reads your posts A LOT closer to the way a human editor would.”

That means LinkedIn is getting better at understanding what a post is actually about, whom it’s useful for, and whether it demonstrates real expertise.

Which explains why a lot of old LinkedIn advice stopped working overnight. Including engagement pods — groups of users who systematically like, comment on, and share each other’s posts.

As Sarah puts it, “Pods are cooked.”

That means there are no shortcuts to building engagement. You’re gonna have to do it the old-fashioned way: creating stuff people want to read.

Rule #1: Your audience trains the algorithm

"The algorithm is basically doing casting now,” says Sarah. “It looks at who keeps stopping to read your stuff. If it's marketing directors and founders, great, it goes and finds more of them. If it's your mom and 3 college friends, congrats, you're getting served to more moms."

The algorithm wants to help you find more of the right people, but you have to tell it who those people are.

Another great reason to get super clear on your “Market of 1” and write directly to that person. 

Rule #2: The algorithm can smell BS

Compare these two openings:

• Option A: Content strategy is more important than ever in today's digital landscape.

• Option B: We spent 6 hours rebuilding our content briefing process and cut revision cycles by 42%.

One sounds professional (annnnd a little robotic). The other sounds real. As LinkedIn gets better at detecting expertise-driven content, generic AI slop has a much harder time breaking through.

The more your post sounds like it came from real experience, the better.

Rule #3: Your profile is a ranking signal

Before it evaluates a post, LinkedIn evaluates the person publishing it.

"If your profile says CMO but your feed is AI tools on Monday, marathon training on Wednesday, and a marketing take every third Tuesday, the algorithm genuinely cannot figure out what you are,” says Sarah. “It shrugs, deprioritizes everything, and moves on."

The strongest executive accounts tend to own a handful of clear lanes and stay remarkably consistent inside them.

When someone lands on the profile, the expertise is obvious.

When the algorithm lands on the profile, it is too.

Rule #4: Build for the second read

A like is different from a save, which is different from a send.

With a send, someone spent enough time with your post to decide it was worth revisiting or sharing with someone else.

That's the behavior the algorithm increasingly cares about, says Sarah.

“A like is the algorithm equivalent of someone nodding at you in a hallway. Polite but meaningless. What the new system actually cares about is whether someone saved your post — saves carry 5x the weight of a like now. Did they share it via DM? Or stop long enough to actually read it? Did they stay? Did they think? A post with 10 real comments beats 50 drive-by likes because the algorithm is trying to figure out if your content actually did something for someone. Applause is cheap. Attention is not."

Which helps explain why certain post formats seem to be everywhere right now.

• Frameworks

• Checklists

• Decision trees

• Benchmark tables

• Teardowns

They're the kinds of posts people save for later, send to a coworker, or come back to when the problem shows up again 6 weeks from now.

Remember the purpose of LinkedIn 

"We've built a real part of our business on LinkedIn and I will die on that hill. It works,” says Sarah. “But the algorithm just got a complete architectural overhaul. It will get overhauled again.”

6 months ago, marketers were optimizing for one version of LinkedIn.

Today they're optimizing for another.

A year from now, they'll probably be optimizing for a third.

The marketers we see winning long-term are tracking audiences they actually own: DMs from actual ICPs, profile visits from target accounts, and newsletter subscribers.

One more thing worth saying: LinkedIn will change this again. 

"LinkedIn always updates the algo, so don’t build your strategy around the current rules set right now. Build it around the goal: getting your right buyer to find you, trust you, and end up in your world,” she adds. 

“Sure, LinkedIn is where they discover you, but your newsletter is where you keep them.”

*Accurate as of June 2026. LinkedIn will absolutely change this again.

Campaigns that got us talking: “ChatGPT has never taken a pregnancy test 8 DPO.” 

Peanut's new "Ask Peanut" feature is an AI engine, but it's built from community discussions, lived experiences, and real data from real women. Not scraped internet content. When every AI tool is competing on intelligence, Peanut is competing on intelligence and where it comes from.

AI spotlight: Mirror, mirror on the wall, what does LinkedIn think I do at all?

A clever profile audit tip: Use LinkedIn’s AI headline generator. What it suggests is the platform’s best guess at who you are. If the answer makes you say, “Wait, what?” you’ve found a positioning problem worth fixing.

Stuff that makes us scroll back up: The Acrobat Cinematic Universe was not on our 2026 bingo card. Adobe Acrobat’s 5-episode workplace comedy, The Marketers, is betting that world-building and entertainment can change how people think about a product better than any feature list ever could.

Hope you didn’t hit delete on b’arb.

Last week she emailed a survey to our long-time readers. This week, she’s back with a gentle reminder.

We’re headed to Cannes soon, and your answers will help shape what we pay attention to, whom we talk to, and what insights make it back into The Standard.

Help us bring home the good stuff.

All we wanted was a little inbox detox. A reality check on what no longer served us. An unsubscribe here, a pause there. And then we’re hit with:

“We will miss you.” 

No, you won’t. And we know we’re not the only ones you’re saying that to. 

We’re not looking for reassurance. We’re trying to complete a task. Sometimes the best UX is letting us leave without making it weird.

Marketers have spent an ungodly amount of time trying to figure out what the algorithm wants.

The answer, as always, is specificity, expertise, and usefulness. The kinds of things good marketers have focused on all along.

In other words, the algorithm is starting to reward the same behaviors we’d recommend even if LinkedIn didn’t exist. 

And honestly, we’d much rather spend time getting better at marketing than gaming the algo.

See y’all next time. 

— the storyarb writers’ room 🫡

Oh! And another thing... 

The # symbol has a real name. And no, it’s not “pound sign” or “hashtag.”

It’s octothorpe.

Octo refers to its eight points. Nobody really knows where thorpe comes from.

Anyway, enjoy being insufferable at your next trivia night. #welcome