THE NEW LINKEDIN ALGORITHM: HOW TO INCREASE YOUR VISIBILITY?
Many LinkedIn users, including the most experienced experts, periodically wonder how does the LinkedIn algorithm really work? What’s in there under the hood? Why does post A enjoys great visibility and traction while post B is lagging behind?
A week ago, a very interesting article was published by the LinkedIn engineering team. It gave us a unique opportunity to take a sneak peek into the world’s biggest professional social media platform.
Exciting news for you! — LinkedIn is updating the algorithm it uses to rank content so that it now factors in the time users dwell on each post (“Dwell Time”).
In this article I will describe how this update affects us as LinkedIn users and those who work in marketing and sales.
A LITTLE BACKGROUND
Every time we sign in to LinkedIn, there are tens of thousands of posts that can potentially appear in our feed. These posts are initially filtered by the algorithm, which identifies the top posts that are relevant to us (for example, based on the content preferences we set in our profile or an analysis of our fields of interests). Next, the algorithm performs a second filtering process, whereby it ranks those top posts based on the “viral actions” they accumulated: comments, shares, and likes.
These viral actions can have a “downstream” and/or “upstream” effect. These terms should be clearly defined:
Suppose that a user called Mike is exposed in his feed to a post made by a user called Jane. LinkedIn assumes that if Jane’s article is interesting and relevant to Mike, he will respond to the post through one or more of the viral actions mentioned earlier: comment, share, and/or like.
If Mike decides to share it, this will create a downstream effect, whereby Mike’s connections will see the post he shared on their feed as well. Conversely, if Mike responds to Jane’s post (that is, posts a comment), this will create an upstream effect resulting in Jane’s post appearing more in her connections’ feeds.
For each of the tens of thousands of posts that are “candidates” for appearing in our feed, the LinkedIn algorithm calculates the likelihood of user engagement and its potential impact on the downstream and upstream effects.
SO HOW DOES THE DWELL TIME AFFECT THE LEVEL OF EXPOSURE TO OUR CONTENT?
LinkedIn realized that there are a few drawbacks to an algorithm that tries to predict the number of clicks and viral actions for each post…
For example, many users are defined as passive content consumers who rarely click on posts. Another limitation is the binary nature of the 3 viral actions, that is, they are either performed or not.
In other words, the LinkedIn algorithm measures whether a viral action was performed or not, but it doesn’t consider the time users spent on a piece of content after performing the action. After all, they may have clicked on the post but immediately returned to their feed. Shouldn’t that affect the post’s ranking?
To make up for these drawbacks, LinkedIn decided to factor in Dwell Time in its algorithm, since it offers several advantages over weighting the viral actions only: Dwell Time is always measurable, it provides a more true and reliable indication of how much interest the content generates and it’s readily available (as opposed to the viral actions, which are less common on LinkedIn compared to Facebook, for example).
WHAT IS THE DEFINITION OF DWELL TIME?
LinkedIn distinguishes between two types of Dwell Time:
1. Dwell Time “on a post” — the time counted from the moment when at least half of the post is visible on the user’s feed.
2. Dwell Time “after clicking” — this metric refers to the amount of time the user spends on the piece of content after clicking on it in the feed.
LinkedIn’s engineers have determined, through a series of tests, that Dwell Time is a reliable indicator of the user’s decision — whether or not to click on the post. Unsurprisingly, they have recognized that users tend to spend more time on posts which they are considering responding to through one of the viral actions. In addition, LinkedIn has identified a category of posts that yield brief Dwell Times from users without any viral action taken. They call this category a “Skipped Update”. The Dwell Time threshold that distinguishes between a viral post and a skipped post is expressed in a rather complex formula outlined in the body of the official LinkedIn article (link attached at the bottom of my article).
Another interesting point — LinkedIn found that the Dwell Time parameter applied to all types of content (text, image, video, etc.). Although users reacted differently to each type, the decision whether or not to click or skip did not depend on the type of content (format) but rather the content itself.
By the way, LinkedIn is not the first social network to use the Dwell Time parameter in its algorithm. Facebook did it before.
THE BOTTOM LINE — HOW WILL THE NEW ALGORITHM AFFECT US?
As mentioned earlier, LinkedIn users tend to spend more time watching updates to which they decide to respond. LinkedIn factored in the Dwell Time parameter in its algorithm, which uses Machine Learning technology. Posts defined as a skipped update get a lower score and less exposure, and vice versa.
LinkedIn believes (and rightly so in my view) that the quality of time the user spends on pieces of content is more important than the amount of time spent in the feed. The most high-quality posts that get higher ranking will not necessarily get more likes, comments and shares…
The impact on us, marketing managers, is clear — It will be less effective to try “cracking” LinkedIn algorithm and generate creative clickbaits. We need to go back to basics, understand our target audience’s needs and create posts that won’t just attract users’ attention but also captivate them for a longer time.
By updating the algorithm, LinkedIn is better assured that its users will be exposed to more relevant posts and gain the most value from their time.
Source: https://engineering.linkedin.com/blog/2020/understanding-feed-dwell-time
Something about myself: I’m a Marketing manager for global technology corporations, marketing mentor for early stage startups, public speaker for Marketing, Digital, Social selling, Partnership, Personal Branding and Recruitment Marketing.
>> Follow me on LinkedIn: www.linkedin.com/in/tomerzuker
