The objective of this thesis was to propose solutions for mitigating potential conflicts of interests regarding online privacy and data use between companies and customers. A particular emphasis was placed on the business model of multi-channel retailing that dominates e-commerce.

In Chapter 2 it was shown that cross-channel effects exist between a company’s physical store network and its e-shop. The perceived size and reputation of physical stores had a significant influence on consumers’ perceived trust in the e-shop. Perceived privacy had the most important influence on the development of consumer trust in our model. The results motivated our further research on privacy and multi-channel retailing.

Chapter 3 developed a Web analysis framework with 82 analyses for Web sites. New conversion success metrics and customer segmentation approaches have been proposed. A particular emphasis has been placed on the development of metrics and analytics for multi-channel retailers. The metrics have been calculated for a data sample of Web user and usage data from a European multi-channel retailer and an information Web site. Implications of the results have been discussed and recommendations for improving business success online have been derived.

Chapter 4 integrated privacy requirements into the analysis process. A privacy-preserving Web analysis tool has been developed for the analyses defined in Chapter 3. The tool indicates when business analyses are not compliant with legal privacy regulations or P3P specifications and thus supports the privacy management within a company. A syntactical extension of P3P for known inference problems and legal regulations has been proposed.

Chapter 5 provided an overview of consumer privacy concerns. A meta-study of 30 privacy surveys emphasized the importance of consumer privacy regarding Web user and usage data. Moreover, the impact of privacy concerns on user-adaptive systems has been discussed. Possible solutions to privacy-preserving personalization have been suggested.

Chapter 6 proposed a new user interface design approach, in which the privacy practices of a Web site were explicated in a contextualized manner and users’ benefits in providing personal data clearly explained. A user experiment has been conducted that compared two versions of a personalized Web store: one with a traditional global disclosure and one that additionally provides contextualized explanations of privacy practices and personalization benefits. Subjects in the second condition were significantly more willing to share personal data with the Web site, rated its privacy practices and the perceived benefit resulting from data disclosure significantly higher, and also made considerably more purchases.

Regarding the structural equation model in Chapter 2 future work should further focus on the interactive influence between e-shop and physical stores. It could be that a reciprocal effect from the Internet on physical stores exists. A comparison of the mutual effects should further explain why multi-channel retailing is such a successful business strategy. Moreover, the integration of further “media channels” (mail, television) and “institutional channels” (call center, sales force) would be an interesting research aspect. Further work should also analyze the impact of cultural differences on privacy perceptions [cf. Jarvenpaa, 1999].

The Web analysis framework in Chapter 3 proposed five categories of Web analyses with 82 metrics and analytics. The framework could be enhanced by integrating further metrics and analytics. In particular, cost-related and detailed product-related analyses would be a useful extension.

The analysis framework was tested on data from a retailer who sells consumer electronics, which belong to a product category that is successfully sold on the Internet [Omwando, 2002]. A discussion of the impact of product characteristics such as search and experience attributes on Internet suitability has been discussed in related work [cf. Nelson, 1974 Phau and Poon, 2000 Subramaniam, et al., 2000 Wright and Lynch, 1995]. Further work should discuss the impact of product characteristics on the analysis results in more detail. For example, in Teltzrow et al. [2003] we have shown that consumers tend to increasingly pick up products with increasing product weight and price.

In Chapter 4, a further development of the privacy-preserving analysis prototype is envisioned. The Web service can be improved by codifying different privacy policies that can be extended to customers from which they can choose one according to their desired privacy level. From a legal viewpoint, it would be interesting to develop a set of Web analyses that meet the requirements of privacy regulations in different countries. A matching and combination of service providers’ policies, user-defined P3P preferences and legal restrictions would help to better protect the customer’s informational self-determination.

An extension of the business model could be the exchange of analysis results between several companies using the framework. Further privacy questions would arise that need to be solved in the context of a three-party business case [Boyens, 2004]. Developing encryption techniques for an analysis Web service that guarantees anonymized transfer of data back and forth from a company to a service provider is also a research question that should be addressed in further work.

In Chapter 5 we discussed the impact of privacy on personalization. Further research should focus on solutions to privacy-preserving personalization [Kobsa, 2003]. Experimental studies should further analyze what company commitments and/or technical solutions increase consumer trust in online personalization.

Regarding our privacy communication design in Chapter 6, additional factors should be tested that may also play a role in users’ data disclosure behavior. In particular, further experiments should explore whether the reputation of a Web site, the stringency of a Web site’s data handling practices, the visibility of contextual explanations or the placing of references to the full privacy policy have an impact on data disclosure and willingness to buy.