A Look Inside Of The LinkedIn Algorithm & “Record Engagement”

This is a quick post on how I estimate LinkedIn has changed, why posts aren’t seen or engaged with as frequently and why AI posts are working and yours aren’t…

There has been a lot made about LinkedIn and filtering their feed and your reach.

Quick Backstory:

  • LinkedIn has a record level of engagement. Which means more users are posting. 

  • Record number of posts means you serve LinkedIn goals and ensure the algorithmic feed serves you what it feels is most likely for you to engage or react. 

  • Use case to learn from: I am a CMO, I like Marketing content, I comment on x number of people’s posts I am connected to or follow and I watch many short clips - LinkedIn knows this and know what you like, what enrages you to engage you and serves a mix of popular and content from your connections. That’s why the same commenters appear most in your feed.

LinkedIn’s record engagement quarter

This is a screenshot from LinkedIn’s recent results broken out from Microsoft’s financial performance Q4 2024

The LinkedIn algorithm isn’t unique. However, it does look for the velocity of engagement by your closest connections and most engagement users and then shares out to outer circles of connections from there. Here’s how to think in a simple diagram

Think of how your shares reach your connections and then out of your first circle, when you post, what you post, how it’s engaged with & who engages and how quickly all helps dictates how your share spreads.

Some aspects that you will have to take into consideration with filtering, engagement and why your post(s) might not get as much reach as before.

  • Geo and location - check your analytics and see how geo-restricted your typical posts are (vs how they used to be) 

  • Time of posting - if you post at 7 am UK time you won’t be hitting the US and other large markets and other geo’s without engagement, tagging and general conversation 

  • Type of post (video, image, text only etc) - video often gains more attention, and more attention can = more useful minutes spent on the platform (an internal goal you’ve hit for them)

  • Following the same formula - AI edited or generated posts are engineered, and they have the same template its simple to see but they gain feedback and engagement - then everyone looks for the same tool others are using, feeding the machine content that generates engagement (and unlocks more ads)

  • Are you limiting your own reach by the content you’re sharing

  • Where does your engagement come from (DM, WhatsApp, feed, share/reshare)? - if your engagement comes from same people and doesn’t go out to extended networks it won’t show to more members (if it looks like an engagement pod or group of the same people it's a simple signal…)

  • Everything posted on LinkedIn could be an ad - a service you’ve created, a piece of feedback from a client, a win you’ve had etc etc. Just think if it could be an ad for LinkedIn there’s probably some throttling that could happen 

  • Most algorithms need to learn from you, learn from your previous shares and they expect a quality vs quantity of posts - consider how you have been tagged by LinkedIn into categories and subcategories 

  • Every engagement is a signal score - every comment and reaction you make is fed into an engagement signal scoring system, the more you engage and react the more the content you are going to be served  

  • Popular accounts show up most in your feed (think how TikTok reengineered the friend’s graph)- if you are active and sharing your expertise or opinions deemed helpful and engaged with the more your closest connections will show up on your “top feed”. The top feed is curated more to you - recent is ordered chronologically. 


Take hints from the internal leaders…Below are a few podcasts where the LinkedIn CPO talks about algorithmic filtering and how AI is in control of some of the feed

Previous
Previous

Anti SNARF? - Do We Need Another Social Network!

Next
Next

How To Own Your Success As A CMO