Elsevier

Journal of Business Research

Volume 124, January 2021, Pages 667-675
Journal of Business Research

Does self-disclosure matter? A dynamic two-stage perspective for the personalization-privacy paradox

https://doi.org/10.1016/j.jbusres.2020.02.006Get rights and content

Abstract

Although marketing managers are relying increasingly on customer data, insight into the best approaches for resolving the personalization–privacy paradox remains limited. Specifically, we argue for the success of a personalization involving the integration of two stages: the self-disclosure stage and the personalization stage. Using a conceptual framework grounded in the foot-in-the-door effect, we argue that compliance with commitment to self-disclosure as the initial small request induces greater compliance with the later target request. The results of a large-scale two-stage field experiment based on a combined propensity score matching and difference-in-difference model show positive causal effects of the act of self-disclosure and the positive effect of the intensity of self-disclosure on purchase responses to personalized promotions. The results also indicate that a combination of privacy assurance and personalization declaration drives customers’ act of self-disclosure and increases the intensity of self-disclosure. Findings empower managers to capitalize on new opportunities in personalization.

Introduction

While there exist tremendous benefits of personalization forbusinesses, such as fostering consumer loyalty (Martin, Borah, & Palmatier, 2017) and increasing retail sales (Luo, Lu, & Li, 2019), the advancement of tracking and profiling technologies in driving personalization has also raised a dilemma for customers—that is, the personalization–privacy paradox1 (Sutanto, Palme, Tan, & Phang, 2013). On the one hand, customers may find online personalization useful because personalized promotions cater to their preference—that is, the value of personalization. On the other hand, customers may feel uncomfortable and anxious when they learn of the inappropriate application of their unauthorized personal information (Martin et al., 2017)—that is, the intrusion of personalization.

Indeed, personalization promotions enforced without a customer’s authorization also exist. This often happens when firms do not deploy privacy policies (e.g., a privacy assurance or personalization declaration) to assuage customers’ privacy concerns; furthermore, they fail to ask for permission to collect customers’ data for personalization. Obviously, much of this is owed to the fact that the rules and regulations for protecting customer data have not yet been standardized or widely used. In this case, lacking a self-disclosure stage may increase customers’ privacy concerns with personalized services and, as a result, negate the appeal of personalization promotions.

The first step of any effective personalization strategy entails the collection of customer data, as these are the driving force behind personalization (Tucker, 2014). In practice, firms use various methods to collect data from their customers, such as orders, surveys, cookies, and data brokers. Among these data collection methods, surveys (e.g., membership registration) are the most effective way of obtaining a fair amount of customers’ personal information because they seek details about individuals’ profile and even some highly sensitive information (e.g., credit card number, residential address). Surveys can also satisfy customers’ request for control and transparency regarding data collection and usage, which have become critical factors in influencing their decisions on whether to share their personal data (Martin et al., 2017). Thus, firms can exploit surveys to obtain a clear picture of their target customers and then use these customers’ disclosed information as insights for a successful personalization campaign.

Drawing on the personalization–privacy paradox, we investigated whether and how the self-disclosure of consumers’ personal information via survey participation would influence their subsequent purchase responses to personalized promotions (Mehta, Ni, Srinivasan, & Sun, 2017). Indeed, many studies have offered solutions to the personalization–privacy paradox, such as privacy assurance and personalization with consent. For example, Özpolat and Jank (2015) highlighted the importance of privacy assurance in influencing purchase behavior, while Goldfarb and Tucker (2011) found that the personalized advertisement increased purchases. Unlike previous studies that devised solutions from the content of privacy policy or personalization strategy (i.e., the external side), the present study tackled the paradox from the perspective of customers’ commitment during the self-disclosure stage (i.e., the internal side). In particular, we argue that the customers’ commitment in disclosing their personal information can alleviate their privacy concerns when confronting with subsequent personalization promotions.

To this end, we constructed a conceptual framework that integrates two separate stages—the self-disclosure stage, at which customers disclose their information to firms, and the personalization stage, at which firms utilize customer data to provide personalized promotions. We adopted the dynamic two-stage framework2 to argue that the consistency between the two stages offers a more precise scope to understand the personalization–privacy paradox than a general positive trade-off between personalization benefits and privacy concerns. Based on the foot-in-the-door (FITD) effect (Freedman & Fraser, 1966), we conceptualize the self-disclosure stage as the initial request and the personalization stage as the later target request by firms. Using a commitment–consistency lens (Freedman & Fraser, 1966), we argue that surveys are effective and useful methods for diminishing the experience of privacy invasion and inducing a positive response to the subsequent personalization request. However, the FITD effect also identifies the critical factor that might mitigate the detrimental effects of privacy concerns involved at the initial stage—a compliance-promoting heuristic (Fennis, Janssen, & Vohs, 2009). A combination of privacy assurance and personalization declaration can, therefore, encourage customers to consent to participating in a survey (i.e., act of self-disclosure) and even drive them to answer more questions in the survey (i.e., intensity of self-disclosure).

To the best of our knowledge, ours is the first study to take an initial step toward conceptualizing a framework that dissolves the personalization–privacy paradox via the customer survey channel and to examine the effects of self-disclosure—that is, the act of self-disclosure and the intensity of self-disclosure—on the purchase conversions of personalized promotions as well as the antecedents of self-disclosure—that is, privacy assurance and personalization declaration. Drawing on the conceptual framework, we formulated three hypotheses. In particular, we first examine the effect of the act of self-disclosure on the effectiveness of personalized promotions at the customer level. Then we investigate the relationship between the intensity of self-disclosure and the purchase amount. Finally, we examine how both privacy assurance and personalization declaration affect survey participation at the self-disclosure stage.

To test the hypotheses, we developed a dynamic two-stage procedure to design our large-scale field experiment. The experiment design enabled an identification strategy that could directly account for the reasons of self-disclosure. The focus of the study was on customers’ self-disclosure solicited by privacy policies (i.e., privacy assurance and personalization declaration). Actual transaction data at the customer level and a large-scale two-stage field experiment in which customers experienced an exogenous event (i.e., self-disclosure request) allowed us to compare the purchase responses of a treatment group versus those of a control group before and after the event. Our findings suggest that customers who disclosed their personal information spent significantly more than those who refused to self-disclose. Moreover, our findings indicate that setting a combination of privacy assurance and personalization declaration in advance can effectively increase self-disclosure during the first stage. These findings can be accounted for by the FITD effect and commitment–consistency principle (Freedman & Fraser, 1966). As the initial request, customers participated in the survey by answering personal questions during the first self-disclosure stage. Then they entered the second personalization stage and received a personalized promotion. Following the commitment–consistency principle, self-disclosure customers would be more likely to grant the firm’s target request by making a purchase of the personalized product or service. In addition, we supposed that the more questions customers answered in the survey, the more information they would disclose for the subsequent personalization, thus achieving a larger FITD effect with more positive purchase responses to the personalized promotions. Moreover, the combination of privacy assurance and personalization declaration as the intervention strategy of the FITD effect could encourage customers’ self-disclosure, because the combination can enhance their control and the transparency of the data collection and usage.

Overall, our contribution lies in our focus on investigating the personalization–privacy paradox through a self-disclosure lens. First, beyond the static cross-sectional perspective, we complement the existing research on the personalization–privacy paradox by taking a dynamic two-stage perspective that integrates both the first self-disclosure stage and the second personalization stage. Second, the previous “trade-off” perspective is limited in explaining the contradiction of simultaneously increasing engagement in personalization and privacy concerns (Tucker, 2014). Our study provides a new perspective to split personalization into two separate stages including self-disclosure stage and personalization promotion stage. Moreover, different from the existing trade-off perspective, we find that self-disclosure does not negatively influence consumers’ purchase decisions but, surprisingly, triggers more subsequent purchases of the personalized product or service. This finding challenges the previous trade-off perspective, which highlights the sacrifice of data collection in place of the balance between utility pursuit and privacy concerns (Sutanto et al., 2013). Besides, rather than examine the effect of data collection at the firm level, we focus on the individual customer level. Our findings provide a deep understanding of customer analytics from the perspective of customers’ behavior patterns.

The remainder of this paper is organized as follows. We first review previous literature on the personalization–privacy paradox and self-disclosure. Then we describe our theoretical foundation and conceptual framework from the dynamic two-stage perspective by integrating the first self-disclosure stage and the second personalization stage. And we present our research model with three hypotheses, methodology, and results. The paper concludes with a summary of our findings, implications, and directions for future research.

Section snippets

Personalization–Privacy paradox

The literature on personalization has identified the critical privacy issues induced by the trade-off between personalization benefits and privacy risks, which is called the personalization–privacy paradox. Personalization strategies can influence consumers’ behavior positively and increase their loyalty. However, privacy concerns arise when firms inappropriately use customers’ personal information and share their sensitive data with third parties without the customers’ consent (Martin et al.,

Foot-in-the-door effect and conceptual framework

The FITD effect posits that customers are more likely to satisfy a larger target request if they have complied with a small initial request than those with refusal of the initial request (Freedman & Fraser, 1966). Prior studies proposed the commitment–consistency principle to explain the FITD effect (Cialdini, Wosinska, Barrett, Butner, & Gornik-Durose, 1999). The commitment to a request enables the accomplishment of later acts related to the initial commitment to move in a consistent direction

Research setting

We conducted a large-scale two-stage field experiment with a leading retailer (which wishes to remain anonymous) in southern China. This retailer carries various product categories, including men’s and women’s apparel, accessories, general merchandise, cosmetics, foods, and electrical appliances. A total of 120,000 customers were randomly selected to participate in this study from the registered members’ database. During the first stage, the retailer sent a short message service (SMS) to the

Descriptive statistics and the causal effect of the act of self-disclosure

PSM uses a set of customer-specific characteristics to calculate the propensity score, which is the probability of the customer purchasing during the post-self-disclosure period (Janakiraman et al., 2018). Those customer-specific characteristics are used to create well-matched treatment group and control group with similar covariate distributions, thereby taking care of bias induced by covariates. Though the selection bias due to unobserved factors cannot be completely eliminated, we chose both

Discussion and implications

In this study, we developed a dynamic two-stage research model of self-disclosure to address the personalization–privacy paradox with respect to whether and to what extent personal information disclosure can result in greater consumer purchase conversion on subsequent personalization, as well as what intervention strategies can encourage more participation in self-disclosure. Based on a dynamic two-stage model, data from a large-scale field experiment yielded the following notable findings.

Acknowledgement

The authors would like to thank the guest editors and the reviewers for their valuable suggestions. The research is supported by the Key Program of National Natural Science of China (Grant No. 71832010, 71832004), the Research Grant Council of Hong Kong SAR (No. CityU 11502218), and the National Natural Science Foundation of China (Grant No. 91646121, 71672164).

Fue Zeng is a professor of Marketing in Economics and Management School, Wuhan University. Her current research interests include Marketing Ethics, Industrial Marketing, Network Marketing, and related marketing issues. She has published over 20 journal articles, including research papers in Journal of the Academy of Marketing Science, Journal of Business Research, Journal of Business Ethics, Harvard Business Review, Industrial Marketing Management, Management International Review, Journal of

References (26)

  • M.C. Dong et al.

    Network embeddedness as a dependence-balancing mechanism in developing markets: Differential effects for channel partners with asymmetric dependencies

    Journal of the Academy of Marketing Science

    (2019)
  • B.M. Fennis et al.

    Acts of benevolence: A limited-resource account of compliance with charitable requests

    Journal of Consumer Research

    (2009)
  • J.L. Freedman et al.

    Compliance without pressure: The foot-in-the-door technique

    Journal of Personality and Social Psychology

    (1966)
  • Cited by (0)

    Fue Zeng is a professor of Marketing in Economics and Management School, Wuhan University. Her current research interests include Marketing Ethics, Industrial Marketing, Network Marketing, and related marketing issues. She has published over 20 journal articles, including research papers in Journal of the Academy of Marketing Science, Journal of Business Research, Journal of Business Ethics, Harvard Business Review, Industrial Marketing Management, Management International Review, Journal of Business-to-Business Marketing, among others.

    Qing Ye is a current PhD student of Marketing in Economics and Management School, Wuhan University. Her research interests include Big Data Privacy, Big Data Analytics, and Marketing Ethics. Her research work has been presented in Proceedings of 2019 INFORMS Marketing Science Conference.

    Jing Li is an assistant professor at the School of Business in Nanjing University. She obtained her PhD degree from the Department of Management and Marketing at The Hong Kong Polytechnic University. Her research interests include Social Media, Digital Marketing and E-commerce. Her research work has been published in Information Systems Research.

    Zhilin Yang is a professor of Marketing in College of Business, City University of Hong Kong. His current research interests include Network in Marketing, Governance Strategies in Marketing Channels, Customer Value and Customer Loyalty, Service Quality and Customer Satisfaction in Electronic Commerce. He has published over 75 articles in various decent academic journals, including Journal of Marketing, Journal of Marketing Research, Journal of Management, Journal of International Business Studies, Journal of Service Management, Journal of Informetrics, Journal of Business Research, Journal of Business Ethics, Service Business, Journal of International Marketing, International Marketing Review, Journal of Advertising Research, Marketing Research, Psychology & Marketing, among others.

    View full text