Management Science · 2018
To Be or Not to Be Linked: Online Social Networks and Job Search by Unemployed Workforce
On LinkedIn and similar networks, job seekers' strong ties produce leads, interviews, and offers. The much larger pool of weak ties mostly doesn't, and can slightly hurt.
Key findings
- Unemployed job seekers now spend more of their search time on online social networks than on many traditional channels.
- Strong ties (people who know you well) are the connections that turn into job leads, interviews, and offers.
- Weak ties are ineffective on average, and marginally negative in some cases. A bigger network is not a better network.
- The finding comes from a structural economic model of search costs and benefits, estimated on survey data from 424 job seekers.
In plain English
The classic sociology finding, Granovetter’s “strength of weak ties,” says acquaintances matter in a job hunt because they carry information your close friends don’t have. That research was done when keeping up an acquaintance took effort, so networks were small.
Online networks changed the economics. Adding a connection now costs nothing, so the average job seeker carries hundreds of weak ties. This study asks whether those connections still pay off.
They mostly don’t. Across the stages of a search, from first lead to interview to offer, strong ties carry the weight. The large outer ring of online connections adds little, and in some cases it slightly lowers the odds of a good outcome.
Why it matters for leaders
- For hiring: Referral programs work through people who can vouch for a candidate, not through network size. Design referral incentives around depth of relationship.
- For platforms: Connection counts are a vanity metric for the outcome users actually want. Features that strengthen relationships beat features that only grow the graph.
- For workforce and outplacement programs: Coaching people to activate a few strong relationships is likely to beat “grow your network” advice.
Abstract
Prior research suggests that social connections, including acquaintances, friends, and family, are valuable in a job search process. In these studies, the size of an average job seeker's network was much smaller and limited by the available modes of communication and the costs associated with maintaining social connections. However, the recent growth of online social networks has enabled job seekers to stay connected with many connections, weak or strong. Thus, the number of online connections—especially weak—has increased significantly. In this paper, we first examine whether an individual's social network plays a role in driving job search behavior, taking into account online social networking sites (e.g., LinkedIn) and other job search modes. Second, we examine how ties in online social networks (both weak and strong) affect job search outcomes (modeled sequentially as job leads, interviews, and offers), and we compare the findings to job outcomes from traditional job search modes (e.g., career fairs, newspaper, Internet postings, and friends and family). To do so, we first construct an economic model of search behavior incorporating cost and benefit functions; we then estimate the model to recover structural parameters using the survey data of 424 users. Our findings show that users are spending more time searching for jobs on social networking sites. In addition, users' strong ties play a significant role in job search and are especially helpful in generating job leads, interviews, and offers; the weak ties, on average, are ineffective in generating positive outcomes and marginally negative in some cases.
Cite this paper
Rajiv Garg, Rahul Telang (2018). To Be or Not to Be Linked: Online Social Networks and Job Search by Unemployed Workforce. Management Science, 64(8), 3926–3941. https://doi.org/10.1287/mnsc.2017.2784
@article{garg2018linked,
title = {To Be or Not to Be Linked: Online Social Networks and Job Search by Unemployed Workforce},
author = {Garg, Rajiv and Telang, Rahul},
journal = {Management Science},
year = {2018},
volume = {64},
number = {8},
pages = {3926--3941},
doi = {10.1287/mnsc.2017.2784}
}