LinkedIn patent reveals the algorithm's secrets

Have you ever wondered why certain content appears in your LinkedIn news feed? Well, and analysis of the LinkedIn patent application to the US Patent Office published in early June 2016 turns out to give quite a few answers.

To summarise a 121-page document written by three very tech-savvy drones, I mean people, this algorithm determines which content, like Pulse articles, status updates and comments, you will see in your LinkedIn news feed. One of the ways the algorthim does this is by determining that you have interacted on a topic. Essentially you have liked, commented, uploaded or shared a topic. To quote from the patent application:

"When a user interacts with a particular item of web content, the social network service detects the interaction and determines future content."

However, it is important to realise that the algorithm is unable to determine if your comment is positive or negative. So if your comment is "why are we talking about Donald Trump's hair on LinkedIn?", you are actually telling the algorithm that you want to see MORE content about Donald Trump plus more content from the person who posted it. It is an implicit expression of interest, even if your intent was in fact the opposite.

These interactions on topics also help the algorithm determine which groups it will recommend to you and which companies you are likely to want to follow. For example, don't be surprised if LinkedIn suggests you follow a Republican Party group or follow Trump companies on LinkedIn.

Key takeaways: if you don't like something, ignore it. If you do, engage. At the very least with a 'like'.

Other key factors used by the LinkedIn algorithm for content personalisation include:

• Your employment history and similar companies you've worked at
• The company pages you follow
• Relationships with employees at a particular company
• Topics listed as interests in your profile
• Keywords used in your profile
• Interactions with third party websites, for example opening links to blogs
• Your listed industry
• Your geographic location

Let me give you an example of a client I'm working with at the moment. He was steered in the wrong direction by a Sydney-based LinkedIn expert. This 'expert' directed him to connect with anyone who asks, even though this is the direct opposite of what LinkedIn official guidelines stipulate. The result is he has a LinkedIn profile with 17,000 first connections. His news feed is full of news he doesn't care about, from people he doesn't know. The suggested birthday work anniversaries are for people he doesn't know for jobs he doesn't care they work at. There is no pattern to their geographic location as he's connected with people from around the globe (even though his business is very Sydney based). His experience on LinkedIn is not valuable and he is considering deleting his entire account and starting again.

Key Takeaways: To make your LinkedIn experience more valuable, make your profile comprehensive and regularly engage with content that appeals to your interests.

A final piece of intelligence revealed by LinkedIn's patent application is the information that the company will be starting to enforce it's publicly stated 30,000 connection limit. This limit has been known of for some time but to date has not been enforced. I've seen several LinkedIn members claiming to have the most amount of connections, as much as 219,000.

The patent application reveals that the algorithm will be used to delete connections and this could be at random. To avoid losing relevant and useful connections, take the time to engage regularly with the content of key people in your network like referral partners. This revelation supports our belief at LinkedIn Ninja Down Under that you should closely protect your LinkedIn network and limit making connections to people you either directly know or have a valid reason for trying to establish a relationship with.

Key takeaways: Keep your connections relevant and engage regularly with the content of people that you want to stay connected to.

There you have it.

To summarise 121 pages of a document clearly writing by tech-heads; to get the most out of your LinkedIn experience moving forward, remain positive in your interactions. Only follow companies and people that you actually admire and want to learn from, or those you want to work with. If by chance you grew your network with anyone that asked and went beyond 30,000 connections, you better start to sort it out or LinkedIn will start to randomly purge them for you. Be focused on the topics that are of interest to you in a professional context and ignore those you don't. LinkedIn will reward you with a better LinkedIn user experience which will ultimately result in more business.

For more information about the algorithm and my findings whilst reading through this riveting document simply email me at [email protected]