LinkedIn's engineers explained the algorithm (with screenshots)

Hola,

You might have seen people say LinkedIn "published its algorithm" this week.

That's a bit much.

It's actually the same algorithm LinkedIn's engineers already explained back on March 12.

What we did get is a specific look at one piece of it: how they pick posts from people you don't follow.

It comes from a November 9 post by LinkedIn engineer Shiyong Lin on their Engineering Blog.

I read the whole thing. So every point below comes with a screenshot of the exact line where they say it, marked in red.

Step 1 finds posts you might like. Step 2 ranks them for you.

This part is from that March post, Engineering the next generation of LinkedIn's Feed.

The new November post is all about Step 1.

LinkedIn sums up your profile into what they call a "member prompt." Then they match posts to it.

What goes in: your headline, your About (they call it summary), skills, work history with role descriptions, and education.

What's left out: your name and gender.

On the post side, they read your text, the title, and even your video transcript.

So the words you say in a video count. So does the title.

To grade how well a post fits a reader, they score it on 7 things:

→ topic
→ location
→ job title
→ company
→ industry
→ seniority
→ education

They add the reader's country and the author's country right into the prompt. The country comes from the location on your profile.

In their words, people are now "more likely to see contents that reflect their local professional context."

Before this system shows your post to strangers, it needs one like, comment or share from someone outside your network.

Just one.

They built this partly because posts from outside your network were hard to find, "particularly for creators who are still building their audience."

In their tests, people hid 18.8% fewer posts.

And their line for creators: "compelling posts can reach relevant audiences beyond their immediate network."

A few weeks ago I shared what LinkedIn's creator team said: your profile has to match your posts.

This report shows the machinery behind that advice.

LinkedIn matches your posts to strangers whose profiles fit them. So your profile and your posts need to talk about the same thing, in the words your buyer uses.

What I'd do this week:

→ Reread your headline, About and skills. Do they say the same topic your posts talk about?
→ Put your topic and the job title you're writing for in your first couple of lines.
→ Give your videos and carousels a clear title, and say the topic out loud early in your videos.
→ If your clients are in one country or region, make sure your profile location and your posts say it.
→ Write so a stranger gets it on the first read. That first outside like comes from someone outside your network.
→ Leave real comments on smaller creators in your niche. Yours might be the outside one that gets their post seen.

Talk soon,

D.

P.S. Want help lining it all up? My free LinkedIn Algorithm Audit Kit has 8 rules for your profile and posts, plus links to the engineering posts I used here → get.diandraescobar.com/linkedin-algorithm