In this article, I explain how the LinkedIn algorithm works and show you how to create better content the algorithm loves to reach your dream audience on LinkedIn.
This LinkedIn algorithm guide is especially useful for content creators, industry thought leaders, and entrepreneurs who are looking for new ways to attract dream clients for their businesses via content marketing on LinkedIn.
After reading this article, you will know how to create high-quality content signals that help you generate a ton of views, engagement, and, ultimately, business on LinkedIn.
Enjoy!
What is the LinkedIn Algorithm?
LinkedIn is the fastest-growing, business-focused social media network with over 1.1 billion members.
User retention has always been one of LinkedIn's biggest challenges.
In the past, LinkedIn was nothing more than a CV database, with users only accessing their accounts whenever they sought to switch jobs.
Consequently, LinkedIn's earnings potential was greatly reduced due to the limited number of ads it could display to each user.
LinkedIn realized that it could dramatically increase user retention and the number of daily active users by opening up its content publishing platform that was previously only accessible to a few industry thought-leaders.
User-generated content quickly became both a silver bullet and a curse.
Yes, users love to create great content and engage with each other.
But, spammers quickly learned how to exploit the platform, resulting in a diminished user experience for everyone.
The solution? A LinkedIn newsfeed algorithm that could eliminate 99% of spam posts, find people interested in the topic of each post, and then use engagement signals to boost high-quality posts and slow down low-quality ones.
How Does the LinkedIn Algorithm Work?
To understand how the LinkedIn newsfeed algorithm works, looking at LinkedIn's big-picture goals is always a good idea.
What does LinkedIn want?
They want their members to spend at least an hour a day browsing through the LinkedIn algorithm.
The more time they spend scrolling, the more ads LinkedIn can show them and the more money they can charge advertisers.
Ok, that's all great for LinkedIn, but what's in it for members? Why would they want to spend an hour scrolling on LinkedIn?
Because they enjoy reading, watching, and engaging with content from people they know and care for.
What would prevent them from spending less time on LinkedIn?
A horrible experience of repeatedly seeing uninteresting, irrelevant, or spammy content in their LinkedIn newsfeed until they rage quit.
The solution?
Giving LinkedIn users such a great user experience by eliminating poor-quality posts and boosting high-quality content that they are willing to tolerate the "few" ads LinkedIn sneaks into their newsfeed between the exciting posts.
What factors affect the LinkedIn algorithm?
The LinkedIn algorithm's operation can be broadly categorized into three domains: eliminating low-quality spam posts, matching the right post to the right person, and determining which posts have the potential to go viral. Additionally, there are also factors such as content experiments and promotional content.
Identifying and Eliminating Low-Quality Spam Posts
LinkedIn employs various strategies to identify and eliminate low-quality spam posts.
This process involves spam and keyword filters to automatically detect and discard spammy or irrelevant content.
There is also a content moderation team that manually reviews posts to ensure they adhere to LinkedIn's content policies.
Showing the Right Post to the Right Person
The LinkedIn algorithm also focuses on showing the right post to the right person.
This involves conducting a semantic analysis of each post, which includes examining the text, video subtitles, and LinkedIn hashtags, to understand the topics of all posts.
After understanding the posts, LinkedIn maps each one onto a broader content cluster map.
The platform then tracks user engagement and groups it based on these content clusters to provide personalized content to each user.
Deciding Which Post Is Boosted with the Potential to Go Viral
LinkedIn uses various user engagement signals to determine which posts have the potential to go viral.
These signals include dwell time, views, likes, comments, shares, etc.
The more engagement a post gets, the higher the likelihood that the LinkedIn algorithm will boost it for increased visibility.
Content Experiments
Sometimes, the LinkedIn algorithm runs experiments by prioritizing different content types over others.
This is done to understand user preferences and engagement patterns better, and it helps LinkedIn improve the user experience.
Promotional Content
LinkedIn also uses the algorithm to promote various features and services.
This includes LinkedIn Ads, LinkedIn Premium, and LinkedIn Learning, which offer paid advertising, premium account features, and online courses, respectively.
The algorithm also suggests connections in the "People you may know" section, recommends pages in the "Pages you might want to follow" section, and provides job suggestions in the "Jobs you might want to apply for" section.
These promotional aspects ensure users have access to the full range of features and opportunities available on LinkedIn.
If you want to learn more about the complicated technology and software behind these systems, check out these three articles on LinkedIn engineering, here, here, and here.
How does the LinkedIn algorithm eliminate low-quality spam posts?
The exact process LinkedIn uses to identify spam content remains proprietary, but we can deduce a lot from their engineering blog.
The spam detection process comprises four significant areas: external link checks, content analysis, engagement analysis, and user account analysis.
Viral Spam Detection
LinkedIn monitors several features from both the post and member interactions to spot viral spam content.
These features include content type, content polarity, spamminess of the content, member network features, activity features, and engagement features.
To tackle the rare occasions when spam content bypasses LinkedIn's defense mechanisms, the platform has developed AI models to predict and detect potential viral spam content.
These models are divided into proactive defenses and reactive defenses.
Proactive defenses predict potential spam content as soon as it appears on the LinkedIn feed, with classifiers focused on specific spam categories like hate speech and on particular content types such as videos or articles.
These defenses run on the platform every few hours to either filter or flag content for human review.
Reactive defenses, on the other hand, provide an additional layer of protection, monitoring activities around posted content and predicting its potential to be shared widely.
They use a mix of predictive machine learning models and heuristics, using member behavior, content features, and interaction patterns to predict if viral content might be spam.
The implementation of both proactive and reactive models has resulted in a significant reduction in the number of unique viewers that encounter spam content.
It has also led to a 7.3% decrease in views on spam content and a 12% decrease in views on policy-violating content.
These efforts toward viral spam detection signify LinkedIn's commitment to maintaining a safe, trusted, and professional user environment.
You can learn more about this approach in LinkedIn Engineering's article.
External Link Checks
LinkedIn employs several measures to catch posts with external links leading to harmful entities such as viruses, malicious code, phishing sites, scams, get-rich-quick schemes, pyramid schemes, MLMs, and ICOs.
They maintain a series of blacklists encompassing domains, URLs, and IP addresses.
LinkedIn also utilizes a malware scanner to spot malicious code on shared URLs.
Behavioral patterns are also analyzed, such as excessive redirects, cloaking, and user reactions to a clicked link, like an immediate return to the LinkedIn platform.
Content Analysis
LinkedIn extensively analyzes content using keyword and phrase matching combined with machine learning.
This approach helps to understand text, recognize patterns, and pinpoint forbidden topics, keywords, and hashtags.
They scrutinize suspicious post metadata, like the number of words and hashtags per post, the ratio between post text and hashtags, and post frequency.
They keep an eye out for duplicate content, be it from the same user, across multiple users (indicative of a botnet), or if it's stolen content.
They also evaluate the post-engagement momentum.
Content creators with consistently low-quality scores receive a negative score, implying that any new content they publish will likely be of low quality, resulting in it being shown to fewer people.
Moreover, if a member marks a post that has slipped through LinkedIn's spam filter as spam, it is reviewed by a moderator.
When the moderator agrees with the spam marking, the post is used to further train LinkedIn's AI anti-spam algorithm, continually improving its efficacy.
Engagement Analysis
Engagement analysis is a critical part of LinkedIn's spam identification process.
They use recursive and graph-based algorithms to identify spam within the tree-like engagement structure.
Each engagement node receives a quality score, and if a top node is tagged as spam, all its subsequent subnodes are also suppressed.
LinkedIn considers several factors, such as whether people like, comment, and share the post, the originality and significance of the comments, and the reaction of new people to the content.
They also examine if the post engagement deviates from the baseline engagement rate to infer the quality of a post.
User Account Analysis
Finally, LinkedIn conducts user account analysis.
They look out for multiple accounts operating under the same IP address and also consider the account age.
They assess the amount of effort someone has put into their LinkedIn profile, like whether the profile is complete and whether a profile photo has been uploaded.
They evaluate the profile for uniqueness or any duplicate content, such as impersonating other users or copying someone else's bio.
Which posts are shown to which users by the LinkedIn algorithm?
LinkedIn uses an internal content cluster map to identify which posts to show to which users.
You can imagine LinkedIn's content cluster map as a huge library with dedicated sections for specific topics.
- Romance
- Fantasy
- Crime
- IT
- Business
- etc.
On LinkedIn, one of these sections could be #marketing.
Within the #marketing section, you can find individual books about subtopics such as #socialmediamarketing, #linkedinmarketing, etc.
The #linkedinmarketing book contains links to all posts that have been created about this topic and all authors who have ever created posts about this topic.
Now, here comes the magic...
Similar to a real-world library, if you borrow a book, your name will be saved in the library database.
"Tim read the #linkedinmarketing book 12 times."
LinkedIn now has three intersections.
- List of people who created content about #linkedinmarketing.
- List of posts about #linkedinmarketing.
- List of people who engaged with #linkedinmarketing content.
This makes it super easy to recommend the right post to the right person.
Let's say you engaged with #linkedinmarketing content in the past.
You log into your LinkedIn account.
LinkedIn looks at your list of topic preferences and compares it to content that has recently been posted.
Your interests:
- #linkedinmarketing x 12 times
LinkedIn's list of recently published content in your network:
- #catmemes x 457 times
- #jobs x 312 times
- #politics x 271 times
- #linkedinmarketing x 3 times
And just like that, LinkedIn found the perfect match for your newsfeed!
Let's add the first #linkedinmarketing post at the top of your feed.
It then continues to measure your engagement.
Did you stop scrolling and read it? Let's increase your interest counter from 12 to 13.
Did you scroll by and stop at a cat meme? Let's add #catmemes x 1 time to your list of interests.
The more time you spend on LinkedIn, the better LinkedIn understands what interests you and the more relevant the posts (and ads) you see in your feed.
How does the LinkedIn algorithm decide which posts to boost?
The first two stages of the LinkedIn algorithm eliminated spam and matched the right content with the right people based on interests.
The third stage focuses on determining the quality of each post to decide which posts get seen by more people and which ones lose visibility.
This determination is made using LinkedIn's Activity Graph. Here LinkedIn looks at a graph combining user-generated content and user engagement, such as dwell time, views, likes, comments, shares, etc.
From LinkedIn's perspective, engagement can become "content" itself, for example:
- Tim liked post X.
- Jane commented on article Y.
Generally speaking, LinkedIn assumes that content with high engagement equals having high-quality content that results in a great user experience.
But not all engagement was created equal.
LinkedIn understands that it takes far less effort to hit the like button than to write a comment.
The more time it takes to engage, the higher the engagement quality and the bigger the impact on the LinkedIn algorithm.
In the past, LinkedIn prioritized countable actions such as the number of likes or comments, or engagement velocity.
This has dramatically changed over recent years. LinkedIn realized that the majority of users fall into the lurker category. Those are people who take the time to read a long article, but would never ever like or comment.
Today, LinkedIn focuses on measuring engagement time, for example, how long someone looks at a specific post. You can learn more about this in my video about LinkedIn Dwell Time.
Here is an example of how LinkedIn prioritizes engagement from most to least important.
- Someone reads your post for 2 minutes.
- Someone writes a meaningful comment for 1 minute 32 seconds.
- Someone shares your post and writes a summary for 42 seconds.
- Someone write the comment 🔥🔥🔥 for 3 seconds.
- Someone hits the like button for 1 second.
LinkedIn tracks every engagement to create an overall engagement score.
Posts with the highest engagement are pushed out to more people. Posts with average engagement are shown to your connections and followers. And Posts with below-average engagement are only shown to your hardcore fans and then quickly lose visibility over time.
LinkedIn Content Experiments
Every once in a while, LinkedIn optimizes its newsfeed algorithm by running content experiments. This could happen because of an updated, more efficient newsfeed algorithm or because LinkedIn is testing a new content type that everyone wants to use.
That's why it is always a great idea to experiment with different content types over time to see if LinkedIn currently gives a specific content type an unfair advantage.
- Write a text-only post
- Publish a video
- Share a photo
- Upload a document
- Write an article
- Create a poll
- Upload a carousel post
- Share a link to an article or video
- Reshare a post created by your LinkedIn page
LinkedIn's promotional content
This is content that LinkedIn hopes to get away with by getting you hooked on valuable content.
Content from this category is generally not very useful to most users.
It includes, among other things, LinkedIn Ads, promotions for LinkedIn Premium and LinkedIn Learning, dynamic promotions for people or pages you may know, and jobs you might want to apply for.
From time to time, LinkedIn will change the ratio between useful and promotional content with the effect or seeing fewer or more promo content in your newsfeed.
If that happens, the best thing you can do is to end your LinkedIn session early to teach the LinkedIn algorithm that they went too far and that they better show fewer ads in the future... 😉
How to create LinkedIn content that the algorithm loves?
Now that we got a clear picture of how the LinkedIn algorithm works, let's create some awesome content the algorithm loves, shall we?
Let's recap what the algorithm wants:
- It wants to keep members on LinkedIn for longer, so it can squeeze in a few more ads
- For that to happen, each member's newsfeed has to be 100% unique and personalized based on posts about their interests.
- The content also has to be of good quality and provide value.
- The more time someone spends interacting with your posts, the higher the perceived value.
Audience relevancy
The first thing we have to get right is targeting the right audience with our content.
There is no point in creating highly valuable content for people who can't relate to it. They won't engage with it, and the LinkedIn algorithm will wrongly assume that our content is of poor quality because of the wrong user signals.
Your LinkedIn content strategy must address this fundamental challenge at the core of audience targeting and positioning strategy.
Once you have absolute clarity on who your ideal target audience is, it is very easy to create content targeting their specific challenges and needs.
Content relevancy
As a successful LinkedIn content creator, you must know your ideal audience inside-out to create highly relevant content.
Putting yourself in their shoes allows you to understand what motivates them.
Use this information to identify themes and topics that are highly relevant to your dream audience and shortlist content ideas with a high emotional impact.
In most cases, these are topics that have a high level of urgency and pain, but very few solutions available online.
Posts with a high relevancy score have the highest chance of positively resonating with the people you are trying to reach.
Engineering engagement
Successful LinkedIn content creation begins with knowing your audience and creating bespoke content tailored to their needs.
However, this is only the beginning. To make this work, you need a third, very important success factor.
You have to be able to engineer user engagement on the post level.
Put another way. You need to keep their attention at 100% while they read or watch your content from start to finish.
This article cannot cover all the details of the exact process. But I will teach you some of the core principles that will allow you to achieve a 100% attention score with every post.
- Grab attention instantly. You must stop people from scrolling at all costs. Focus all of your energy on writing attention-grabbing headlines that trigger instant curiosity.
- Don't reveal everything all at once. Once you get their attention, it's important to keep them engaged. Don't reveal the secret to your question right away. Split the answer into multiple success factors and pace their revelations throughout your post. The longer people look at your post, the higher your post's dwell time will be, and the more the LinkedIn algorithm will favor you.
- Optimize micro-transitions. Every sentence should naturally transition from one to the next. As simple as this sounds, as complicated as this can be in practice. Your goal is to remove any and all possible brain glitches that could occur while consuming your content. One of the best ways to eliminate these obstacles is to read out your text aloud. If you notice any difficulties as you speak, it's probably better to rewrite your text for better readability.
For video content, the same principles apply.
- No intros. Start with your content instantly. Instead of writing a headline, convey the value proposition in the first five seconds of your video.
- Create and upload subtitles for each of your videos. On LinkedIn, videos are played muted by default. That means that even if you have the perfect value proposition at the start of your video, many people might never hear it. By adding subtitles, your spoken words will be overlayed on top of your video, dramatically increasing the odds of other members hitting the play button to hear what you have to offer. I recommend using Otter to create your subtitles.
- Remove silences and filler content. People are impatient, especially when it comes to video content. When you edit your videos, make sure to remove anything that isn't absolutely necessary, such as duplicate explanations, unnecessary pauses, etc. I recommend using TimeBolt to automatically remove all silences from your videos.
How to create a LinkedIn Content Strategy for your business?
If you want to attract high-quality leads by publishing content on LinkedIn, you need a good LinkedIn content strategy.
I recommend checking out my LinkedIn Leads Bootcamp. It's a 21-day program that walks you through the exact steps I use when developing LinkedIn content strategies for my VIP clients.
Make sure to check it out.
