Introduction

In today's competitive business world, understanding and leveraging the power of the four Ps of marketing, namely product, price, place and promotion has become one of the major factors that determines a company's marketing future survival or success. However, these four Ps have all been modified by technological evolutions and especially the Web and they have been converted into the four Ps of e-marketing, namely permission, privacy, profiling and personalization (Goldsmith, 1999). According to (Mobasher & Dai, 2003) 'personalization is the provision to each individual of tailored information, products, or services' and the subject of this chapter is to study the significant role of personalization in intelligent e-marketing applications.

The roots of personalization are traced back to the introduction of adaptive hypermedia applications in Brusilovsky (1996, 2001). Adaptive hypermedia were introduced as an alternative to the traditional 'one- size -fits-all' approach, building a model of the goals, preferences and knowledge of each individual user, and using this model throughout the interaction, in order to adapt to the user 's specific needs (Perkowitz & Etzioni, 1997, 2000a).

Personalization is a broad scientific and technological area, also covering recommender systems, customization, one-to-one marketing, and adaptive Web sites (Blom, 2000; Mulvenna et al., 2000; Schafer et al., 2001). It is the process of gathering and storing information about Web site visitors , analyzing the information in order to learn users' patterns, habits and preferences, and based on this analysis, delivering the right information to each visitor at the right time (Eirinaki & Vazirgiannis, 2003).

Personalization can be deployed by e-businesses (any Web site whose operation is related to commercial purposes, e.g., e-shop, e-procurement, e-auction, e-mall, portal, etc.) and be used as the key technology for managing customer relationships, running targeted advertisement campaigns , promoting products or services and customizing Web site content (Perner & Fiss, 2002). Besides, according to the forth of the five mutable laws for Web marketing (Wilson, 1999) the objective of e-commerce sites should be to 'pull people to the site by its attractive content, then push quality information to them regularly via e-mail.' However, personalized content apart from e-mail can be advertising, recommended items, screen layout, news, or anything else accessed via commercial Web site.

The digital channels that can be deployed by an e-business to reach its customers are numerous : Web, e-mail, ftp, chat, search, voice over IP, interactive multimedia, etc. The power of these digital channels is significant, since they can dramatically change the way the e-business listens, understands and responds to its customers and suppliers. Moreover, these channels facilitate instantaneous responses to the information gathered by various touch-points ( buttons ' selection, hyperlinks , transactions, log files, etc.) without the need for human intervention (Yao et al., 2001).

Web personalization may be implemented in the form of:

• personalized content presentation and delivery (inserting or removing thematic units/sections/paragraphs, optional explanations or detailed information, personalized recommendations/offers/prices/products/services).

• personalized structure (sorting, hiding, unhiding, adding, removing or highlighting links), e.g., yahoo.com (Manber et al., 2000).

• personalized Web site layout presentation and media format (from images to text, from text to audio).

The above can be based on the analysis of data from online and off-line sources using a variety of techniques: association rules mining, clustering algorithms, classification techniques, collaborative filtering, patterns discovery, prediction, log files analysis, etc. (Vassiliou et al., 2002).

The benefits for the businesses can be significant and include among others increasing site usability, converting users to buyers , retaining current customers, re-engaging customers, penetrating new markets, etc. (Nielsen, 1994). On the other hand, personalization applied to the e-commerce domain is an effective way to provide added value to customers by knowing and serving them as individuals. Customers need to feel that they have a unique personal relationship with the businesses they come in contact with. Users have now the ability to visit Web sites that allow them to find information or choose products and services to buy fast and easily. Moreover, they can receive e-mail, newsletters or other information that they attach importance to. And even though a significant percentage of online users give their permission to receive these kinds of information from marketers, they tend to unsubscribe and take back their permission when their preferences and needs are not met and they end up receiving useless or irritating information. This observation brings up the importance of quality and precision in producing and delivering personalization, as a means for assuring customer faith and loyalty.

There are two general approaches to personalization in e-commerce applications:

• Buyer driven: the customer subscribes to different services, fills in forms or questionnaires, rates products, participates in surveys, etc.

• Seller driven: the adaptation is controlled by the owner of the e-shop. Content is served up using pre-defined business rules, targeted online advertising, as well as product cross-selling and up-selling. It typically uses a rules engine which captures and encodes business rules, describes workflows and automates manual tasks .

The basic steps comprising the personalization process (depicted in Figure 3-1) are the following (Cooley et al., 1999a; Srivastava et al., 2000):

• Data collection.

• Data processing.

• Personalization output.

  
  Figure 3-1: Personalization process

This chapter aims to define Web personalization for intelligent e-marketing applications, describe the tasks that typically comprise the personalization process and demonstrate the close relation between personalization and Web mining. Moreover, our goal is to illustrate the future trends in the field and, in this way, suggest directions that may trigger new scientific research.