Linkedin’s secret experiments

Linkedin Business Network research algorithm connect job search
The Linkedin network can completely change resumes. (Image: Souvik Banerjee / unsplash)

Business network Linkedin has conducted numerous experiments to improve its algorithms in the background. The results are astounding.

These days almost everyone has to be on Linkedin, the global business network, somehow. A school friend here, a work colleague there – and over time a useful professional network is created that can also be relied upon when looking for a job.

But now the media has discovered via a scientific publication that the social network is running certain tests in the background to improve its algorithms. This is not too surprising, because users agree to this when they first register.

However, the latest experiment has drawn some criticism, for example from  the “New York Times“, because it is likely to have influenced the lives of many people.

Actively recommends unknown contacts

The study is about the fact that Linkedin, over a period of five years, had analyzed the influence of the strength of interpersonal relationships on the change of a job. This included actively suggesting new acquaintances, it said in the journal “Science”.

The hypothesis was that job offers work better through people with whom there is a less-strong connection than just through good friendships.

Rare possibility

The study now confirms this, and scientists from MIT, Stanford and Harvard Business School emphasize that it is usually very difficult to test such a thesis with data, and therefore the experiment on the platform is important.

According to the results there is indeed a correlation that people’s jobs changed in exactly the direction where the business network had suggested loose acquaintances for them to connect with.

This phenomenon, known as the ‘paradox of weak ties,’ is perplexing – namely that people did not change jobs in response to firm friendships, but rather used less-strong connections to become more professionally mobile.

Two billion connections

Linkedin conducted the random selection-based experiments between 2015 and 2019. In the process the company sent messages tailored to members about people they might know. In doing so, the network wanted to test the extent to which weak ties increased professional mobility within the world’s largest business network.

According to the researchers they actually found evidence of such connections.

To do this the creators included more than 20 million users and analyzed around two billion new ties over the five-year period, which in turn led to 600,000 new jobs.

Three observations

But what they found specifically was that – firstly – the strength of the ties was non linear. Rather, the relationship looked like a U-shaped curve between the relationship and the job change. That is, in other words, relationship weakness increased job flexibility – only up to a point.

Secondly, weak ties assumed on the basis of other relationships within one’s network, for example, showed strongly-different effects. Thus, weak and extremely weak ties led to the greatest job mobility.

“People you may know”

And thirdly, it also depended on what industry the participants were in. In digital-savvy fields the correlation to weak ties on job mobility was clearly greater, they reported.

This now supposedly means for the future that if once again an email comes from Linkedin asking if you know this or that person, that you should definitely add them to your network. It could be your future employer – even if it is a completely unknown person!


Linkedin’s secret experiments

Leave a Reply

Your email address will not be published. Required fields are marked *