Wednesday, November 03, 2021

The Case of altered LinkedIn algorithm

Once upon a time, a person called Reid Hoffman co-founded a website that would allow professionals to share their knowledge and experience with each other. Unlike industry specific websites, this one was for all professionals to connect with each other. 

The platform turned profitable within 3 years of launch and has, since inception, remained the world's no. 1 professional platform. 

In 2016, it was acquired by Microsoft at a valuation of 26.2 billion dollars. The leadership remained the same. 

Some changes were made to the UI then. These changes were criticised in the short term but the overall usage of the platform did not suffer. 

In 2019, Linkedin decided to change its algorithm. The key driver of the change was "engagement" - that elusive Utopia of networks. 

In its blog post,  Linkedin spoke about the technicals, but this article helps break it down, to some degree.

What we can understand is that underrepresented content creators will be given some algo space on the feed so their reach can grow, and community engagement will be one of the key drivers of what succeeds, and what does not. In short, the algo will learn what works and what does not. 

The intention, perhaps..

was to create a situation in which exceptional content created by any content creator, would be highlighted. In time, the best content would get traction normally, and the algo would learn what works and what does not. 

The program worked perfectly, but... 

So, what happened was exactly what was intended - content creators were given feed space. The best performing feeds were picked up by the algo and similar feeds were amplified in the future. 

The result of which was that images of Gods, personal stories of pathos, achievements of children, random polls, political jokes, and other assorted paraphernalia increased. In 2021, it is hard to tell whether one is scrolling Facebook or Linked in. 

The algo learnt perfectly. The humans behaved ... well, not as intended. 

The success of the algo was based on the assumption that users would only engage with professional content on a professional platform. That didn't happen. 

So, dear readers, please do not blame Linkedin for the perceived deterioration of your LI feed. This is what other humans want to see and read.  

Users who still value only professional content on LI, are called LI purists (and that is not a positive term). 

No comments: