The Perils of Over Connection
Employee turnover is costly, disruptive, and—despite all the data we collect—very hard to predict. Most companies point to external reasons when someone leaves: a better offer, a higher salary, a relocation. But my research suggests that the real drivers of turnover often lie within the organization itself, hidden in the structure of workplace relationships.
Over the past few years, I have studied how employees’ positions in the internal social network of a firm—the informal web of advice, collaboration, and support—can shape their likelihood of staying or leaving. Results have been quite surprising. We tend to assume that being central in a workplace network, that is, being the go-to person in a company, is always a good thing. It signals trust, competence, and influence. Yet, my research uncovers a more complex picture: employees who are highly central are also the ones most exposed to interruptions, overload, and eventual burnout. Specifically, combining field, archival, and experimental data, I consistently show that the association between social networks and turnover is non-linear. While employees at the very periphery of the network are more likely to leave because they feel excluded or underutilized, those at the very center are also at higher risk of quitting. Their constant involvement, nonstop flow of requests, and pressure to be always available can lead to burnout. The safest position? Somewhere in the middle. Those who are well-connected but not overwhelmed are the most likely to stay.
That centrality can backfire is not always visible on the surface. These employees rarely complain. They are seen as high performers and are often rewarded for it. But over time, the very thing that makes them indispensable, such as constant requests and continuous connectivity, can become a source of fatigue. With little time for reflection or recovery, they face what the literature refers to as information overload. When these individuals leave, it is not just a loss of talent; it creates a ripple effect across the whole organization.
This is a blind spot in many retention strategies. We invest heavily in wellness programs, engagement surveys, and tailored benefits, which are all important efforts. But we often miss the relational structure that shapes how work actually gets done. Traditional people-analytics models focus on individual attributes. But a social network approach may reveal that risk is relational. Some of the most at-risk employees are not isolated or disengaged: they are overconnected.
The solution is not to discourage collaboration or responsiveness, but to be more intentional about how we design work. Mapping the social networks within an organization can help us identify who is carrying an unsustainable share of the relational load. These employees may need more support, protected time, or redistributed responsibilities. To be clear, this does not mean turning every human resource department into a team of network analysts. But it does mean acknowledging that social connections and turnover are entangled in ways we have only just begun to understand.
As firms deal with post-pandemic work arrangements, hybrid models, and evolving expectations, this is a chance to rethink not just where people work, but how they are connected with each other. Sometimes, keeping your best people does not require more perks, but rather a better design of informal relationships.
