Mitigating the Negative Effects of Customer Anxiety by Facilitating Access to Human Contact

Prior research in social psychology has shown that when people feel anxious, they seek advice from others. Yet, companies that operate in high-anxiety settings (like financial services, healthcare, and education) are increasingly deploying self-service technologies (SSTs), through which anxious customers transact without access to human contact. These companies may therefore face a classic efficiency–service trade-off where gains in operational efficiency through automation also hamper service outcomes by neglecting customer anxiety. This paper, set in the high-anxiety domain of financial services, investigates the value of facilitating access to human contact during SST usage. In a field experiment conducted within the context of a credit union’s SST loan-approval process, an invitation to connect with a human loan agent increases the customer uptake rate of approved loans by 24%. Subsequent controlled laboratory experiments, which explicitly investigate the linkages among anxiety, choice satisfaction, and firm trust during SST encounters show that compared with participants without anxiety, those who are anxious consistently report lower satisfaction with their choices and less trust in the firm, regardless of the source of their anxiety. When anxiety is related to the SST encounter, facilitating access to human contact dampens anxiety’s negative effects on choice satisfaction and, by extension, increases firm trust. We further observe that these benefits of facilitating access to human contact arise even though very few customers choose to access a person. Taken together, this research reveals an opportunity for firms to maintain operational efficiency in high-anxiety self-service settings without compromising customer experiences and service outcomes. This paper was accepted by Charles Corbett, operations management. Funding: Funding for this research, apart from internal business costs incurred by our field study partner to support the experiment, was primarily provided by Harvard Business School and partially provided by Dartmouth College. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2022.01029 .