Optimization

How to build an effective professional network on LinkedIn: Some data-driven insights

Co-authors: YinYin Yu, Guillaume Saint Jacques, and Paul Matsiras

Introduction

In 2019, LinkedIn announced an initiative to close the network gap as part of our commitment to equity, and ensuring that everyone has  equal access to opportunity. We asked people to take the Plus One Pledge and help someone outside of their network who may not have access to the same resources, and pledged systematic changes toward equity on LinkedIn. This work is important because networks can translate into tangible, significant professional opportunities. Yet as we’ve seen, there are many factors, including socioeconomic status, gender, and race that contribute to inequality in network strength among people. Amid a global pandemic that has upended the livelihoods of people around the world, the importance of having a strong professional network is more important than ever.

So how can we, as a company, help close the network gap? As a first step, we can make sure the products we build do not unintentionally reinforce systemic inequality or implicit bias. This is one of the purposes of Project Every Member. We also wanted to provide a concise and actionable list of things a LinkedIn member can do to expand their network. In particular, what can job seekers do to improve their professional network?

How to build your network: Insights from social science

Scientists from LinkedIn’s Applied Research-Computational Social Sciences team leveraged the career transition and network data of our nearly 740 million members across the world to arrive at the following insights into the relationship between your professional network structure and your career mobility (how quickly a worker moves onto their next job). Compared to those with similar demographics, educational backgrounds, and professional experiences, 

  • Members with at least 13 connections from companies other than their current employer are 22.9% faster in transitioning to their next job than those who do not meet this bar.

  • Members who are a part of at least one group on LinkedIn are 8.6% faster in transitioning to their next job than those who are not a member of any group.

  • Members who follow at least 5 organizations on LinkedIn are 7.1% faster in transitioning to their next job than those who do not.

These new insights build on our own experimental results that network expansions that help members reach into more professional “clusters” or establish more diverse networks increases career mobility.

Methodology

Your career mobility depends on a myriad of factors, including demographics, education, professional experience, and network structure. Our goal is to identify the top network features that contribute to members’ career mobility and to quantify the contributions of those features, while controlling for other factors that are not directly related to their networks. Our approach can be broken down into the following steps:

  1. Estimate members’ expected job transition time based on their demographics, education, and professional experience, among other non-network features.

  2. Select top network features that are the most predictive of the residual variation in the time it took for members to actually transition, after controlling for non-network features.

  3. Discretize the network features selected in Step 2 into binary features and quantify their relationship with members’ career mobility.

In the following sections, we will describe the crux of each step. For more technical details, please see our companion paper (link).

Defining career mobility
For the purposes of this study, we do not focus on the “engagement” or “viral” aspects of a network. Rather, we focus squarely on its professional implications. To that end, we measure career mobility as the time it takes for members to move onto their next jobs. This can be a move within the member’s current employer or a career transition to another employer. Specifically, career mobility is measured by the time elapsed between the start of an existing job (j) and the start of the subsequent job (j+1)

t-sub-i-equals-start-date-sub-j-plus-one-minus-start-date-sub-j

Our analysis is based on the career progressions of a random sample of 3 million members from across the world who held at least one full-time job between July 1, 2017, and December 31, 2019. The dates were chosen to avoid confounding data from changes in job seeker behavior during the global pandemic.

Step 1: Accounting for the effect of demographics and experience
Although professional networks are important, they are by no means the only factor that affects career mobility. For example, demographics such as age, gender, education, and country of residence, as well as professional features such as industry, job function, seniority, and level of experience, all factor into your career mobility. In order to make more of an apples-to-apples comparison later on, we control for a total of 18 features that are not directly related to members’ professional networks but may affect their career mobility in the first step of our analysis. In particular, we use a survival model to estimate the expected amount of time it takes for members to transition to their next job based on their demographics, professional experience, and other non-network related features.

Step 2: Identifying the most important network features for career mobility
In the second step of our analysis, we sought to identify a shortlist of the most important network features for career mobility. To this end, we constructed a list of 24 features that capture different facets of one’s professional network. For example, in addition to the raw count of members’ LinkedIn connections, we constructed features that represent their connections’ absolute/relative seniority, employment status, company, industry, and job function; we also considered looser definitions of “professional network,” such as whether or not the member is in a group, how many companies the member is following, and how many professional endorsements the member has received from their network.

To understand the relationship between members’ networks and their career mobility, we look at the residual variation in career mobility that can’t be explained by non-network features in Step 1 and try to identify the network features that are the most explanatory of this residual difference. The following are the top three network features that were selected by LASSO as the most predictive of the residual difference in career mobility after controlling for a myriad of non-network demographic, education, and career-related features.

  1. Number of connections from companies other than the member’s current employer

  2. Whether the member is in at least one group on LinkedIn

  3. Number of companies that the member follows on LinkedIn

Step 3: Quantifying the relationship between top network features and career mobility
In the final step of our analysis, we quantify the relationship between the residual variation in members’ career mobility (not explained by their demographics, education, and professional experiences in Step 1) and the top three network features identified in Step 2 using an OLS regression. After controlling for members’ non-network features, we observe a statistically significant relationship between all three of the network features from Step 2 and members’ career mobility. We further discretized these network features in order to identify the threshold to cross that corresponds with the greatest gain in career mobility. From this, we arrive at a checklist of items that we believe can give members some guidelines as to how to build a professional network on LinkedIn that can accelerate their career progression.

The Professional Network Checklist

By comparing members with similar demographics but varying professional network structures, we identified three network features that bear the strongest relationships with career mobility. Relative to comparable members who do not satisfy these network criteria, those who do, on average, realize between 7.1% and 22.9% higher career mobility speeds (defined as the inverse of the time it takes members to transition to their next job from the start date of the current job).

table-view-of-results-previously-described-in-intro

These findings underscore the importance of building a “diverse” network, and they illustrate that your professional “network” is not restricted to direct connections with individuals. LinkedIn offers our members a number of tools (connections, groups, follows) through which they can access information that can potentially accelerate their careers. Through this blog, we offer evidence of how effective usage of these tools can help with members’ career mobility, as well as provide guidance on how professionals can best utilize our tools to accelerate their own careers.

Acknowledgements

We would like to thank Parvez Ahammad, Sofus Macskassy, Ya Xu, Meg Garlinghouse, Bari Lemberger, Paul Ko, Karin Kimbrough, Stephen Lynch, and Meg Hoppe for insightful comments.