LinkedIn Tweaks Its Algorithms to Skew Feeds to User Interests

LinkedIn on Tuesday disclosed algorithm changes designed to favor conversations catering to niche professional interests. Axios first reported the shift.
The top 1 percent of power users grabbed much of the attention on its platform, LinkedIn found through internal research in 2018.
The criteria for posts showing up in members’ feeds can be summed up as “people you know, talking about things you care about,” said LinkedIn Senior Director of Product Management Pete Davies.
LinkedIn prioritizes and ranks posts based on the following criteria:
Individuals a member has interacted with directly;
The member’s profile;
The member’s colleagues; and
Who would benefit from hearing from the member. It terms this “creator side optimization.”
“The actual algorithms are fairly straightforward,” said Stephen Lynch, senior communications manager at LinkedIn Engineering.
One selects likely content for a member’s feed, and the other finalizes what actually pops up in the feed, he told the E-Commerce Times.
One algorithm is called “Community-focused Feed Optimization” and the other “Communities AI: Building Communities Around Interests on LinkedIn.”
LinkedIn tested 100 variations of its feed models through online A/B testing over the last year, and implemented several targeted optimizations that make it easy for members to find content that interests them personally, Lynch said. “We are heavily invested in making sure that our members see things that are interesting to them.”