How the LinkedIn feed works

LinkedIn's feed behaves unlike the other social platforms, and the difference explains most of what confuses people about it.

Test, then expand

A post is shown first to a small slice of your network. What that slice does decides whether it goes further. Perform well and it expands to more of your network, then to second-degree connections, then to people who follow the topic.

This is why LinkedIn reach is so bimodal: posts either stall at a few hundred impressions or run to tens of thousands, with little in between. You are not being ranked on a curve; you are passing or failing a threshold.

The first hour matters disproportionately

Because expansion depends on early performance, engagement in the first hour has outsized influence. This is true, and it is also the origin of a great deal of unpleasant behaviour — engagement pods, "commenting for reach", posting at prescribed times.

The honest version: post when your audience is actually online, and reply to every comment quickly, because replies are engagement and they extend the window.

Dwell time is the underrated signal

LinkedIn weighs how long people stop on a post, not just whether they clicked something. This is why long text posts with a "see more" break perform — the expand is a dwell event — and why a post that is instantly comprehensible can underperform one that requires reading.

Used cynically this produces the formatting tricks the platform is known for. Used well it means: give people something worth stopping for.

Topic classification

LinkedIn classifies what your post is about and who cares about that topic. Posting consistently on one subject makes this classification sharper and expansion more reliable. Posting about everything makes you hard to place, and hard-to-place posts stop at the first threshold.

Your network shapes your ceiling

The seed audience is your connections. A network full of people who do not care about your subject caps you before quality is even assessed. Whose connection requests you accept is a distribution decision.

What trips people up

• Blaming the algorithm for a post nobody wanted to read.
• Copying formatting from viral posts in unrelated categories.
• Optimising posting time to the minute, which matters far less than the first hour of replies.

You have got this when

You can explain why your posts cluster into "a few hundred" and "many thousands", and what decides which one happens.