September 27, 2022
Between 2015 and 2019, LinkedIn conducted experiments on more than 20 million users to improve the functioning of the platform for members and potentially affect livelihoods.
LinkedIn experimented by spontaneously modifying the proportion of weak and strong contacts suggested by its People You May Know algorithm.
One of the experiments compared the effects of encouraging the formation of strong ties (recommendation to add close friends) versus weak ties (recommending acquaintances and friends of friends). Subsequently, the researchers followed the users who took part in the “A/B testing” to see if the difference had an impact on their employment outcomes.
LinkedIn users were then randomly assigned to one of three treatment groups. Users in each group were shown slightly different new contact recommendations, leading some groups to form stronger bonds and others to form weaker bonds.
The team then counted how many jobs each group’s users applied for and how many job transmissions occurred. A job transmission implies that the new contact assisted in landing the job. Results showed that the recommender engine significantly influenced link formation.
The experiment shows that moderately weak ties are more than twice as effective as strong ties when it comes to helping job seekers join a new employer. Third, the strength of weak ties varied by industry. Strong ties increased job mobility in fewer digital industries, while weak ties increased job mobility in more digital industries.
The experiment concluded that your close friends (on LinkedIn) are not the best bet for finding jobs. Instead, you should turn to acquaintances with whom you have no personal connection.
The sources for this piece include an article in NYTimes.
