Article Preview
TopIntroduction
Many organizations are increasingly using BI tools such as data mining that work on top of data warehouse to build analytics especially domains like customer relationship management (CRM) to identify, attract, understand, serve and retain the customers (Ngai, Xiu, & Chau, 2009). The customers of hyper competitive retail businesses such as telecom, e-tailing, insurance, media and entertainment have been pampered with choices. New products and services are introduced on ongoing basis and time to market has been substantially reduced. Changes in job profiles, income levels, locations, family statuses, life styles along with availability of various new options and choices, their tastes and requirements keep changing. Analysis that is done based on what they did many years back may not be relevant in the current context. Analysis should be done on ongoing basis tracking every activity the customer does, looking for opportunity and taking immediate action or within reasonable time. In competitive environments, the business firms require continuous analysis of data generated through various activities for insights, only episodic and ad-hoc or craft-based will not suffice (Davenport, Harris, & Morison, 2010). Delays in recognizing, interpreting and acting on insights and changes are critical emerging impediments to competitiveness (Prahalad & Krishnan, 2008).
BI many a times includes tasks such as classifying millions of customers into loyal and non-loyal, clustering or segmenting them into few groups like platinum, gold, iron, lead etc., however, in the worst-case scenario each one can have an unique behavioural patterns based on his or her demographic profile, interactions or transactions that he or she does with the firm. It may not be that effective to apply the broader group rule(s) or behaviour(s) to the customer or try to fit the customer into a group to understand and serve him or her. The bottom line is: even a firm is dealing with millions of customers the firm must analyse, understand and serve one customer at a time (Prahalad & Krishnan, 2008) to get better customer life time value.
Firms like Amazon, Netflix offer unique personalised experiences to their users based on their past transactions, likes and dislikes. Large number of products and contents (catalogues) are made available online and they are personalised and recommended. So that each user gets what he likes or tastes are. It helps to form groups of niches: people who like particular kind of product or product (Anderson, 2006). Various filtering techniques (Felfernig, Friedrich, & Schmidt-Thieme, 2007) have been used to implement personalization to offer unique experiences to individuals. While personalisation has been successfully applied in domains like e-Commerce (Adomavicius & Tuzhilin 2005; Balabanovic & Shoham, 1997; Burke, 2007; Felfernig et al., 2007; Mandl, Felfernig, Teppan, & Schubert, 2011), one of the challenges is, offering it on various mobile interfaces like SMS, voice and WAP (wireless application protocol). The interfaces like SMS and voice are less interactive not much of information is available compared to WAP kind of interfaces where click-streams can be tracked like on the Web. Another challenge is the size of mobile screen. The personalised items (we used the common term item to refer to product, service, content, campaign, offer etc.) to be targeted and pushed to the individual mobile subscriber must be very relevant based on his or her changing likes and dislikes, preferences and needs, objectives and constraints.