You are probably familiar with the Amazon’s personalized recommendations that induce consumers to purchase products. The personalized recommendations are generated by using the detailed records of an individual consumer and corresponding analysis about webpage browsing and consumption. But where is the calculation of the privacy of the consumer?

According to (Mobasher et al., 2000) personalization is any action that tailors experience to a particular individual. The study of Tam and Ho (2006) is more specific by mentioning that personalization is the process of adapting web content to meet the specific needs of users and to maximize business opportunities. Both definitions have in common that personalization involves the tailoring/adapting of certain “subjects” that can be distinguished into products, website features, advertising or communication (Tsekouras, lecture 2).

The benefits of personalization can be analysed from different perspectives. From the consumer perspective, personalized recommendations allows them to see products with more relevance and less effort. From the perspective of the company, personalized recommendations leads to higher conversion and loyalty. However, there are also potential drawbacks: overpersonalization and privacy concerns (Tsekouras, lecture 2). Privacy concerns are increasing because consumers fear that their information will be misused and they don’t like the feeling of being tracked. Due to the drawbacks and benefits, this blog post aims to broaden your knowledge by highlighting both the privacy concerns and the utility of personalization services in e-commerce.

The paper of (Zhu et al., 2017) concentrates on the personalization of services in e-commerce. Although there are benefits associated with personalization, the privacy concerns may drive consumers away. This is referred as the personalization-privacy paradox. Consumers base their decision to disclose information on the trade-off personalization-privacy.

The paper uses the multi-attribute utility theory to analyse the combined outcome of the personalization-privacy paradox (utility). Subsequently, the model is tested by a simulation in Matlab R2012b using random data that depicts the random behavior of consumers. The simulation example is based on the single side of benefit or cost, the benefit-cost analysis and reputation scores. The main findings shows that the value of utility is sensitive to the factors of costs and benefits of information disclosure, the level of privacy concern and company reputation. Company reputation can effectively reduce the perceived risk in the information disclosure for privacy fundamentalists.

As a strength, the paper of (Zhu et al., 2017) is academically relevant: the paper addresses the lack of balanced research that analyses and reconciles the contradiction between privacy and personalization service. The paper therefore uniquely classifies consumers on the basis of their privacy concerns and analyse their behavioural differences. This is depicted in figure 2 of (Zhu et al., 2017) where the following three consumer segments are identified: “privacy unconcerned” (least concerned about privacy and most willing to disclose their personal information), “privacy pragmatist” (they both consider the risk of privacy and benefit of personalization) and “privacy fundamentalist” (mostly considers privacy concerns and have a minimum care about personalization).

Besides the academic relevance, the second strength is managerially relevance. Firstly, the findings of this paper facilitates managers in the creation of a more accurate personalization strategy and privacy management. Secondly, through the use figure 2 managers learn that consumer segments should all be addressed differently to maximize the utility of the product/service: “privacy unconcerned” with the best personalization business model and highest data collection, “privacy pragmatist” with the medium personalization business model and lesser data collection, “privacy fundamentalist” with standard products and utmost least data collection. Thirdly, managers can improve their brand image to address the privacy fundamentalist because the findings show that company reputation can effectively reduce the perceived risk in the information disclosure.

Let’s illustrate the managerial findings with an instance, Groupon.com. Consumers receive location-based promotion but they dislike the feeling that their online footprints are hacked. By weighting the personalization-privacy paradox, they ultimately decide to cancel their account. Through the use of the findings of (Zhu et al., 2017), the cancellation of privacy fundamentalist can be prevented by improving the corporate reputation of Groupon.com.

The weakness of the findings of paper of (Zhu et al., 2017) is the low external validity. The researchers used a simulation to test their framework and it is therefore questionable whether the findings are fully applicable to the real-world. Future research should therefore use real data that quantifies consumer’s preference on the basis of past transaction records, so that the consumer is placed in the proper customer segment. Furthermore, future research could also be done by examining how the website should be designed to address the personalization needs of different customer segments. Thirdly, the relationship between the identified customer segments and profit can be examined.

In what consumer segment do you think that you belong? According to your customer segment, do you agree with your levels of personalization and privacy as suggested by (Zhu et al., 2017)?

References 
Mobasher, B., Cooley, R., & Srivastava, J. (2000). Automatic personalization based on web usage mining. Communications of the ACM, 43(8), 142-151.

Tam, K. Y., & Ho, S. Y. (2006). Understanding the impact of web personalization on user information processing and decision outcomes. MIS quarterly, 865-890.

 Zhu, H., Ou, C. X., van den Heuvel, W. J. A. M., & Liu, H. (2017). Privacy calculus and its utility for personalization services in e-commerce: An analysis of consumer decision-making. Information & Management, 54(4), 427-437.