Feed behaviour

Why your work feed is full of holiday photos: how feed algorithms work

From:
, founder
Published:
Updated:

The short answer: most feeds, including professional ones, are ranked by predicted engagement. The algorithm shows you what it thinks you will react to, and personal, emotional or provocative posts get faster reactions than plain, useful ones. So they travel further, people notice, and they post more of them. Your feed fills with holiday photos not because anyone decided it should, but because the scoring system rewards them.

How does a feed decide what to show you?

Every time you open the app there are far more possible posts than you could read. The feed has to rank them. Most do it roughly like this:

  1. Gather candidates: posts from people you follow, people they interact with, and topics you have shown interest in.
  2. Predict your reaction: for each post, estimate how likely you are to like, comment, share or stop scrolling.
  3. Boost early winners: a post that gets lots of reactions quickly from its first viewers gets shown to more people.
  4. Apply rules: demote spam, bait or anything against policy, and mix in variety.

Steps 2 and 3 do most of the work, and they reward one thing: reactions.

Why do personal posts beat useful ones?

Because emotion is fast. A post about a hard week or a proud moment gets an instant "congrats" or a heart. A careful note on how your team cut invoice errors gets read, saved and acted on, but often without a reaction in the first hour. To a ranking system, the first looks more popular.

People are not daft. Once they see the personal post get twenty times the reach, they write more of them. Some go further: life lessons from a child's football match, polls with obvious answers, "comment YES" hooks. That is the route from "professional network" to the feed you have now. We wrote more about the result in the problem with modern networking sites.

Is LinkedIn changing its algorithm?

It says so, repeatedly. LinkedIn has announced that it shows fewer polls and less engagement bait, that it values substantive comments over quick reactions, and that it reduces the reach of generic, repetitive content that looks AI-generated. It has also targeted engagement pods, groups who agree to react to each other's posts.

These are sensible changes. But the underlying system still ranks by predicted engagement, because engagement is what a free, ad-funded network sells. Tweaking what counts as good engagement helps at the edges. It does not change what the feed is for.

How can you make your current feed less awful?

A few things genuinely help:

  • Unfollow, don't disconnect. You can stay connected to someone and stop seeing their posts.
  • Use "not interested" on the posts you dislike. It does feed back into ranking.
  • Follow specific people. Seek out those who post numbers, methods and mistakes from their own work.
  • Switch to most recent if your app offers it, so you see posts in time order instead of ranked.
  • Hide the feed. Some people use browser extensions that hide the feed and leave search, messages and jobs. It is blunt, but it works.

What would a feed built for work look like?

We think the answer is to change the question the feed asks. Instead of "what will this person react to?", ask "is this about work, and is it useful?"

On Office Worker:

  • Every post is checked before it reaches a feed, first by an automatic check and then by an AI review of the whole post.
  • Off-topic posts are caught early. Holiday photos and life updates are sent back before they reach anyone. The author is told why, and a personal post never leads to a strike.
  • Bait earns nothing. Posts that look like engagement bait are held for a person to review instead of going straight out.
  • No reaction contest. Posts are not ranked by how fast they collect likes, so there is nothing to gain from farming them.

It is a less exciting feed in the slot-machine sense. That is the idea. You can read the details in what a work-only network looks like.

Frequently asked questions

How does the LinkedIn algorithm work?
In broad terms, it predicts which posts you are most likely to engage with, based on who posted, how similar people have reacted, and how quickly a post gathers early reactions and comments. LinkedIn says it now weighs meaningful comments more than quick reactions and reduces the reach of engagement bait and generic AI-looking posts.
Why does my work feed show so many personal posts?
Personal stories tend to get quick emotional reactions, and feeds ranked by engagement reward quick reactions. So personal and emotional posts travel further than dry, useful ones, and people learn to post more of them.
Can I make my LinkedIn feed better?
Partly. Unfollow accounts that post bait, use 'not interested' on posts you dislike, follow people who post specifics in your field, and switch the feed to most recent if your app offers it. Some people use browser extensions that hide the feed entirely.
How does the Office Worker feed decide what to show?
Every post is checked for whether it is about work before it reaches a feed. Personal posts such as holiday, pet and kid photos are sent back, engagement bait is held for a person to review, and spam, harassment and misinformation are sent back. The feed is not a contest for reactions.

Sources

  1. Search Engine Land: LinkedIn changing feed, will show less low-quality content and polls
  2. Entrepreneur: LinkedIn is fighting back against AI slop
  3. Sezer, Gino and Norton (2018), humblebragging