LinkedIn Now Understands What You Write. Three Consequences for Publishers.

For ten years, the LinkedIn feed was governed by dozens of specialised ranking systems, each with its own job. One decided the order of posts in the feed. Another filtered job recommendations. Another managed notifications. Each looked at the world through its own narrow window: a handful of signals, quick calculations, isolated decisions. It worked — but with one clear limitation: none of those systems actually read what you wrote. They counted reactions, measured clicks, mapped keywords. They didn’t understand.

Since early 2025, LinkedIn has changed its approach. It has publicly unveiled 360Brew, a single reading infrastructure that replaces much of those specialised systems. This isn’t an internal footnote: it’s the algorithm that decides what your network sees of your profile, and in what order.

What 360Brew Does, in Brief

360Brew treats every distribution problem as a natural-language question: “Given this member, given this context, how likely is it that this post will interest them?” To answer that, it reads the content. It doesn’t extract numerical features to feed into a separate system — it processes the text as text, with all the meaning embedded in it.

What changes operationally for anyone publishing on LinkedIn? Three things.

1. Specificity Beats Volume

When the reading infrastructure was an aggregation of signals — number of reactions, speed of interactions, recognised keywords — publishing frequently and broadly worked. Five posts a week, all on popular topics, captured enough traffic to guarantee visibility.

360Brew rewards the opposite. A specific post — a precise observation, a well-sourced data point, a defensible argument — is interpreted as relevant to the right people. A generic post is read as generic, and distributed accordingly: to everyone in theory, to no one in practice.

The practical upshot: one post a week that actually says something specific is worth more than three posts that circle a topic without landing on it.

2. Editorial Consistency Beats Variety

A reading system that understands content can also connect an author’s posts over time. If you publish on two or three consistent themes for six months, 360Brew builds a representation of you as “a voice on those topics.” When you publish a fourth post on the same thread, the system already knows who to show it to.

If, on the other hand, you publish everything — a post on leadership, one on football, one on sustainable agriculture, one on emerging markets — that representation stays blurry. The system struggles to categorise you, and therefore struggles to find you an audience. It isn’t a punishment: it’s the consequence of an editorial identity the system simply can’t read.

For professionals, this means one straightforward thing: pick one or two topics and own them for months. Become “the person who talks about X” — not for the sake of ego, but because that’s how people find you.

3. Engagement Tricks Lose Their Power

“Comment hooks” like “What do you think? 👇” or “Tag someone who needs to see this” used to work because the old ranking mechanism treated comments as a strong signal. More comments meant more visibility, even if they were one-word replies.

A system that reads content sees through that mechanism. A comment saying “interesting!” adds no information — neither for the reader nor for the distributor. A post that generated ten empty responses is read for what it is: a text that asks for attention without giving any.

That doesn’t mean comments stop mattering. It means the ones that matter are the relevant, substantive ones — the comments that actually add something. And those, as a rule, come when you’ve written something that merits a considered response.

What to Do in Practice

The implicit invitation from 360Brew is to stop treating LinkedIn as a blind distribution channel and start treating it for what it is: a distributed newsroom where each author builds a recognisable voice over time.

Three concrete actions:

• Declare an editorial lens. Decide on two or three topics you want to own, and make sure the vast majority of your posts fall within them.
• Prioritise your own sources. Publish from what you read yourself — papers, articles, data — rather than from what’s already circulating in the feed. The algorithm rewards fresh angles, not rehashes of familiar ones.
• Write the way you speak. A recognisable voice, week after week, is just as readable to the system as it is to a human reader. A generic voice is invisible in both directions.

LinkedIn, in other words, has stopped counting and started reading. The way you manage your presence should follow the same trajectory.

It is also why tools built to “produce more” age badly on a platform that reads: the iPeople vs Taplio comparison lines up the two approaches — multiplying posts versus curating them.