When cross-selling backfires: modeling customer reactions to sales attempts Zeynep Akşin, Evrim Güneş, Lerzan Örmeci Koç University Hazal Özden, Koç Sistem July 2007 Outline (cid:131) Motivation and research question (cid:131) Brief look at the literature (cid:131) Modeling framework (cid:131) Brief survey results (cid:131) Models with different forms of negative reaction (cid:131) Conclusion July 2007 Cross-selling (cid:131) Important revenue generation tool in customer relationship management (cid:131) The practice of selling additional products and services to existing customers (cid:131) According to a McKinsey report estimate (2006) cross-selling can generate as much as 10% of the revenues through a bank branch network (cid:131) Call centers can attempt to cross-sell to 60% of its callers (Anton, 2005) July 2007 Customer reactions to cross-selling Customer reaction “This new loan option Try to Cross-sell? is exactly what I need!” + $$€€ “I don’t want another sales pitch, just transfer the money! ” Lost time, annoyance •(+) Retention: Marple and Zimmerman, 1999 •(+) Reduce churn: Kamakura et al. 2003 •(-) Switch: Kamakura et al. 2003 July 2007 Our research questions (cid:131) How can we model negative reactions to cross- selling attempts? (cid:131) Would such reactions influence optimal cross- selling policies? (cid:131) If yes, how? July 2007 Brief look at related literature (cid:131) Descriptive models of customer relationship – Schmittlein, Morrison and Colombo, 1987 – Schmittlein and Peterson, 1994 – Netzer, 2004 (cid:131) Optimizing customer equity – Ho, Park and Zhou, 2005 – Rust, Lemon and Zeithaml, 2004 – Venkatesan and Kumar, 2004 – Ching, Wong, Altman, 2004; Sun and Li, 2005 – Sun, Li and Zhou, 2006 July 2007 Relation between a customer and the firm P - f 1 = ) t p e c c a ( P Exponential lifetime with rate µ P ( d e cli n e ) = P Poisson f with rate λ Customer reacts: • lowers utilization of service? • quits relationship earlier? • less inclined to accept Like in HPZ 05 future attempts? Assumptions July 2007 of SMC 87 Which type of reaction do we focus on? (cid:131) Web based survey – 104 respondents with prior exposure to call center cross-selling – Convenience sample: average age 32.8; 43% female (cid:131) Is the probability of failure affected by the number of previous failures (i), and the number of previous contacts (j)? (cid:131) Does a failed cross-sell offer affect the rate of contacting the call center, which is a measure of λ, and the probability of leaving the bank (as a measure of µ)? July 2007 Survey results (cid:131) Increasing i (number of failures) increases the probability of failure (cid:131) The majority of the subjects stated that they would not change their contact rate and quit rate as a function of failed attempts (cid:190) We focus on modeling negative customer reactions as a lower inclination to accept future attempts (cid:190) We consider different effects the number of contacts may have on customer behavior July 2007 A model of the relation between a customer and the firm (cid:131) Our aim: To maximize the total discounted value generated during the lifetime of a customer relationship with the firm. (cid:131) Assumption: Failure has a cumulative effect while success allows starting over “fresh”. (cid:131) State: – Number of customer contacts (j) – Number of attempts (i) (cid:131) State space: S = { (i,j) : 0 ≤ i ≤ j } July 2007