LinkedIn has launched new features for Pages to enable virtual gatherings and conventions with Events & Live and has also explained factors considered while ranking a post on a feed.
LinkedIn Events & Live
LinkedIn Events features are launched to aid professional communities and help them adjust to the reforming state of in-person events and stay connected to your company and meet consumers.
The platform claims LinkedIn Live is seeing 23X more comments per post and 6X reactions per post than native video and is a great way to have real-time engagement with your followers and connections.
LinkedIn Live will give you the option to live stream to your Page followers or Event attendees. Hosts can use 3rd-party broadcasting partners, including Restream, Wirecast, Streamyard, and Socialive. The platform will be adding more partners in the coming months.
Users can send direct invitations to their first-degree profile connections and share the event with the Page followers. Drive buzz and engagement for the event or live broadcast by posting updates to the Page or Event feed and by recommending key posts for attendees.
Extend the shelf life of the LinkedIn Events and keep discussions around the topics or concerns raised with the Video tab — a dedicated hub for a Page’s organic video content.
LinkedIn Feed Ranking
In a recent blog post, the platform explained how the platform ranks posts on a user’s feed, and the dwell time and users’ actions to a post affect the feed ranking.
Essentially, in layman terms, a user’s reaction to a post plays the major role in the feed ranking of the post. ‘Viral Actions’ such as ‘Like’, ‘Celebrate’ have an ‘upstream’ or ‘downstream’ on a post.
For instance, if Tanya sees a post by Bob and finds it interesting, she would hit ‘Like’. If she re-shares the post, it will propagate the article downstream, as Tanya’s connections will see the article in their feed.
But, a comment on Bob’s post will have an upstream effect as “it provides valuable feedback to the creator(Bob) that may encourage him to post more often”.
Therefore Tanya’s(or your followers’) reactions with both considerations – her engagement to the post and upstream or downstream effects on her network, decide the feed ranking of a post.
Machine learning models are then trained to predict several quantities for each possible click and viral action (click, react, comment, share):
P(action) = Probability of Tanya taking this action on the update
E[downstream clicks/virals | action] = Expected downstream clicks/virals if Tanya takes this action
E[upstream value | action] = Expected upstream value to Bob if Tanya takes this action
The outputs of these models are then synthesized into a single score using a weighted linear combination, where the weights are tuned appropriately and the three components are balanced in order. This score is used to perform a point-wise ranking of all the candidate updates.
Shortcomings of this approach
Click and viral actions can be rare, specifically for passive consumers of the feed, who visit the feed frequently and find value in the updates but may shy away from taking viral actions.
Click and viral actions are primarily binary indicators of engagement—either you carry out the action or you don’t. For actions related to sharing, the text associated with a comment or re-share (if available) can provide a richer signal, although that signal can be more difficult to interpret.
Clicks are noisy indicators of engagement. For example, a member may click on an article, but quickly close out, realizing it’s not relevant, and return to the feed within a few seconds, called ‘click bounces’.
LinkedIn incorporates ‘dwell time’ into their machine learning models that rank posts in the feed to make up for the shortcomings related to viral actions.
Dwell time is fundamentally categorized into two segments: dwell time on the feed – when at least half of a feed update is visible while scrolling through the feed; & dwell time after the click – time spent on content after clicking on an update in the feed.
To learn more about the technicalities of the mathematical equation that the machine learning model uses to rank posts on a feed, head to this blog that explains it with an example.