How Does The LinkedIn Algorithm Work?
How Does The LinkedIn Algorithm Work?
Have you noticed a decline in your reach on LinkedIn lately? In the good old days you used to have lots of likes, comments and shares on your posts, but in recent months this volume of interaction has been significantly reduced. Why is that?
After reading the post from the LinkedIn Engineer Dept called “Strategies for Keeping the LinkedIn Feed Relevant”, I thought my audience would benefit from a quick summary of how the algorithm actually works (in layman’s terms).
The LinkedIn algorithm is not the same as Edgerank (Facebook algorithm), but there are some similarities. What surprised me when reading the blog was how labour intensive the process was. Yes, they actually use real people as part of their process to see who will see your content in the newsfeed.
There are four main parts of the algorithm, which has been supported by a nice looking flow chart (see below). Algorithms are introduced to social media sites as a way of reducing spam on the networks and making the content relevant for the consumers.
Step #1 – Initial Filter
LinkedIn’s automation system will classify every single image, text, long form post and also video as one of three options.
I was unable to source any content online to find out how they decide what category your post will be segmented into. I assume it is to do with the language that you are using in the content. For example, if you are using words such as “offer”, “promotion”, “discount”, etc. then it will be scored under the “spam” category. I am not sure of the definitions are for “Low-quality” or “clear”. (If you know, or have an idea, then please comment below)
Step #2 – Test Audience Process
Once your post has been categorised by the robots, it is then sent to a selection of your audience to see how popular the content is. It is NOT sent out to everyone that you are connected too, just a small number of your friends.
Step #3 – Scoring System
Once placed in front of a selection of your friends, different actions have different weights in the algorithm. For example, a “like” may only have one point, where as a comment takes more effort, therefore, it has two points. A “share” shows that the content is popular so will be awarded three points. (This is my version of the scoring to explain how it works). The score of the post will decide its future or its fate. It will either:-
a) Be demoted because it’s low-quality
b) Be shown to more people because it’s high-quality
Step #4 – Real People Assessment
If you start to get engagement on your post then it is passed along to “real people at LinkedIn” who then read each and every post and make a decision as to whether your content should be sent out to more people, or whether this is the end of the line. If the content is working really well then the Editors will also send the content to people outside of your network. It will start to appear in “trending content” on the app, or topical content in the newsfeed etc.
Gone are the days when you can simply add broadcasting messages to LinkedIn. If you are a recruiter and all you do on LinkedIn is post your jobs, then you will certainly fall into the “spam” or “Low-quality” category and no one will be fed your posts.
The same as any other social media site, if you produce content that is entertaining (which is why you are seeing more Facebook style posts on LinkedIn), interesting, educating or topical then you have more of a chance of your content being shared and seen by more people.
What do you think? Do you agree with the way in which the algorithm is implemented or are you fed up with seeing funny and amusing posts on LinkedIn? Let me know in the comments below.