How the new LinkedIn™ Algorithm Works (and why you see posts twice)

Hey, new algorithm — why do you act so strangely? Let’s talk about 14‑day déjà vu

Two weeks. In the age of AI‑driven overload, two weeks feels like centuries.
And yet… I see it happen. I commented on something. Then, exactly 14 days later, the same post shows again — right above or below something fresh.
Do I need a memory rewind? No. But why does the algorithm insist on this rerun?
Let’s unpack how social media curation works, how LinkedIn™’s algorithm has evolved, and what it means when your feed keeps recycling content.

How social media “algorithms” decide what you see

What is algorithmic curation?

Social platforms don’t show everything in chronological order. Instead, they use algorithmic curation — systems that pick which posts surface, and when — tailored to your interests. (Wikipedia)
Their goal: keep you engaged, scrolling, clicking. The smarter the algorithm, the more time you spend on the platform.

• Who you interact with most
• What kinds of content you engage with (text, image, video)
• Whether content is “fresh” or “relevant”
• Whether posts are flagged as “spam” or “low quality”

The “filters, test, ranking” pattern

1. Initial filtering — detect spam, remove clearly low-quality posts
2. Engagement test / sampling — show to a small audience, see how they react
3. Ranking & broad distribution — boost content that passes the test; suppress or stop distribution of underperformers

Why you see repeats

When your feed seems to loop, it’s because the algorithm is balancing “freshness” vs “relevance.”
If a post remains relevant (lots of engagement beyond its first push), it may resurface later. If you haven’t seen it or engaged with it, the system may “retry” showing it.
Also: sometimes the algorithm reuses content when it detects a lull of fresh content, or to boost engagement by reminding you of something you almost missed.

LinkedIn™’s algorithm: evolution, quirks, shifts

Early days: chronological ➝ relevance

In its early phase, LinkedIn™’s feed was more chronological or connection‑based. You’d see what your network posted in order. But as LinkedIn™ grew, volume exploded. It became impossible to show everything to everyone.
So LinkedIn™ had to adopt a recommendation system — not chronological, but relevance‑based.

The three-step model (as of 2025)

Hootsuite and other analyses break it down nicely: (Hootsuite)

1. Quality filtering
   The system evaluates whether your post is spam, low quality, or high quality. This step weeds out content with too many links, obvious engagement bait, questionable grammar, or violations of guidelines.
2. Engagement test (“golden hour”)
   Shown to a small subset of followers. If engagement (comments, shares, dwell time) is high, the post earns broader distribution.
3. Ranking and expansion
   LinkedIn™ pushes it further based on identity (your profile), content signals, and member activity. Posting consistently on a topic boosts your reach as a perceived expert.

Key shifts & trends

• Meaningful engagement — comments matter more than likes
• Native content > external links — put URLs in comments
• Experts gain traction — consistency builds trust
• Longer post lifespan — content can resurface weeks later
• Constant testing — reach volatility is expected

When déjà vu hits your feed — what’s going on?

This can happen because:

• The post maintained engagement over time, so the algorithm resurfaces it
• You may not have “completed” the conversation
• The algorithm is testing whether another exposure triggers action
• Content scarcity in your feed means strong older posts are recycled

Tips to “train” the algorithm (so it doesn’t play tricks on you)

• Engage early: Comment and start discussions in the first hour
• Encourage real conversation: Ask open-ended questions
• Avoid engagement bait: Don’t use “like this if…”
• Use native formats: Avoid outbound links in post body
• Stay focused on your niche: Build content consistency
• Balance content: Don’t rely too heavily on repeats

Closing thoughts

Algorithms are not perfect, but they aren’t random either. They’re constantly testing, filtering, and optimizing to surface content they believe will keep you engaged.

LinkedIn™’s algorithm has matured: it values expertise, conversation, and native content. But the downside is that you sometimes feel stuck in a content echo chamber, seeing what you already saw.
When the feed gives you déjà vu, it’s not personal. It’s algorithmic logic running behind scenes. But now that you know how it works, you can play to its strengths — and maybe avoid repeated posts hovering just when you thought you’d moved on.