خطرات رقابت برای بلیط قطار - تحقیقات تجربی از رفتار مشتری و عملکرد در صنعت راه آهن

Based on a comprehensive data set of German railway customers we analyze consumers’ choices and particularly subsequent changes of two-part pricing contracts (loyalty cards). In a competing risks framework, we simultaneously estimate effects on three types of contractual events: cancellations, upgrades, and downgrades. Focusing on customer relationship management (CRM) practices, we find several factors affecting these events, some of which railway companies can influence to their advantage. Intuitively, installing auto-renewal procedures for loyalty cards decreases cancellation hazards. However, automated electronic mailings (e.g., reminders and account statements) and advertising (e.g., ticket offers) can be counterproductive and increase the risk of cancellation.

Rail transportation has been the subject of increasing research in recent years because of its far-reaching economic, social, and environmental impacts at multiple societal levels. Numerous methodological and empirical studies have profoundly elaborated ways of handling railway timetabling, modeling and optimization of operations, as well as capacity and pricing issues (e.g., Abril et al., 2008, Batley et al., 2011, Corman et al., 2010, Corman et al., 2012, Hansen, 2007 and Lin et al., 2012). Another stream of research has focused on consumers’ selection criteria and usage preferences for railway travel or other means of transport (e.g., Bhat and Sardesai, 2006, Hess et al., 2007 and Keumi and Murakami, 2012). However, in environments like the transportation and logistics sectors marked by increasing competition, researchers and practitioners alike have emphasized a growing need for customer orientation and relationship management (Ganesan et al., 2009). Reinartz et al. (2004) define CRM at the customer-facing level as a systematic process to manage customer relationship initiation, maintenance, and termination across all customer contact points to maximize the value of the relationship portfolio. Boulding et al. (2005) explain that CRM relates to firm strategy and centers on the development of appropriate (long-term) relationships with specific customers or customer groups, the acquisition of customer knowledge, and the intelligent use of data and technology, to enhance customer loyalty and organizational performance. Accordingly, interest in CRM practices in the transportation sector has grown considerably in recent years (see Daugherty et al., 2009, Ellinger et al., 1999, Grawe et al., 2012, Ramanathan, 2010 and Steven et al., 2012). However, research in the context of two-part pricing schemes is scarce. In addition, previous studies often neglect settings where travel decisions are not only based on consumers’ current preferences, but are also influenced by previous contractual choices. As a consequence, a comprehensive approach towards understanding customers’ travel behavior, and particularly, the determinants of their contractual choices and changes therein and how such decisions can be influenced effectively, is lacking in the literature on (rail) transportation settings. In various international markets for rail transport, and particularly in the German market, severe oil price increases and growing environmental consciousness have helped public transit and rail transport to achieve substantial and profitable growth over the last decade. In 2011, the light rail traffic and the heavy rail sector served more than 2.5 billion railway travelers, most of them traveling by the major German railway company “Deutsche Bahn AG” (DB) (Statista, 2011, p. 16). DB offers passengers the option of purchasing in advance various loyalty cards that act to discount ticket prices for 12 months from the date of issue. These contractual devices are commonly-known as BahnCards and are widely-used in German railway transportation. Schmale et al. (2013) have studied the travel behavior of German BahnCard customers and found that a disproportionately high proportion of customers fall victim to the “flat rate bias” and underuse their BahnCards. Accordingly, based on the premise that customers strive to choose more suitable BahnCard contracts as time passes, in this study, we focus on analyzing the determinants (e.g., customer demographics, usage behavior, pricing, loyalty programs) of consumers’ choices of BahnCard contracts as well as changes to these contractual choices over time. We collated comprehensive travel history data spanning a timeframe of almost 6 years and applied a non-generic competing risks framework. Using a semi-parametric proportional hazards model stratified by failure type, we simultaneously estimated effects on three types of contractual events: cancellation, upgrade, and downgrade of a BahnCard. In addition to identifying reasons for terminating a BahnCard or substituting a different one (which translates into an upgrade or a downgrade scenario), we were also able to quantify the magnitude of effects. We studied these issues based on a large-scale, longitudinal data set comprised of more than four million individual transactions. Accordingly, we find several factors affecting contract choices, some of which railway companies can influence to their advantage. Our study contributes to the literature by offering both theoretical and practical implications. This paper proceeds as follows. In the next section, we describe the characteristics of the loyalty cards studied (here: BahnCards), and describe our data. The third section explains the methodology. We briefly introduce the concept of survival analysis, describe approaches towards competing risks and apply a specific method developed for the context at hand. We then report our main results; the last section offers a discussion and conclusions.

This study demonstrates how hazard models combined with data multiplication methods can be used in a context of competing risks to model railway travelers’ choices of two-part pricing contracts in terms of contract cancellation, upgrade and downgrade behavior. Specifically, we estimate an “initial-change” model for annual train ticket contracts (BahnCards) based on the occurrence of cancellation events or loyalty card changes in a large-scale sample from Deutsche Bahn AG (DB). In comparison to other models of this type described in the literature or used in practice, the proposed model takes a customer-centric approach, as it captures the underlying behavior of train passengers. We extend the conventional cancellation risks framework by applying competing risks models to the specific context of customer loyalty cards, which provides a broader perspective on customer retention in the railway sector. Our method originates from survival analysis and has not been transferred to transportation studies on consumer travel behavior before. The model is appropriate to examine consumers’ contractual choices across different segments of the population (male vs. female, young vs. old, business customers vs. consumers, etc.) and across varying usage intensities (low, medium, high). Therefore, it could help transportation firms assess the potential effects of developing and optimizing two-part pricing schemes and of adopting and redesigning various relationship management practices on consumers’ subsequent (inter-temporal) contractual choices and usage decisions. To our knowledge, this study is the first to be based on an extensive and unique longitudinal dataset (more than four million individual customers’ transactions) in the railway market that explicitly considers contractual decisions within two-part travel pricing schemes. Based on this study approach, we can develop some interesting reflections for the field of customer relationship management. The value and the efficiency of diverse forms of customer programs have been the subject of controversial discussions among researchers. For example, Leenheer et al., 2007, Meyer-Warden, 2008 and Noordhoff et al., 2004 and Vesel and Zabkar (2009) study the characteristics of such programs and contribute to our understanding of the effectiveness and economic outcomes of loyalty programs. Despite the considerable proliferation of loyalty programs, Berman (2006) observes that many have not produced the desired results. According to Allaway et al. (2006), little is known about the responsiveness of different customer segments and about differences in the behavior patterns of customers included in such programs. In addition, in recent years the awareness of the need to improve our understanding of the outcomes of CRM practices in the transportation sector and also of the resulting opportunities to influence consumer choices more effectively has grown considerably (see Ellinger et al., 1997, Ellinger et al., 1999, Ramanathan, 2010 and Steven et al., 2012 on various aspects of customer services and loyalty in logistics and transportation settings). As managers and academics become increasingly interested in the “true” value of CRM practices, based on the evidence provided here we point out some systemic flaws in one of the biggest loyalty programs in Germany, and we can highlight some critical leverage points relevant to improving the effectiveness of loyalty programs in rail travel. Our results may suggest a direction for relationship management to take in similar transportation contexts as well. We have highlighted the customer demographics, CRM practices, pricing strategies and BC usage factors that are most influential on consumers’ upgrading, downgrading and cancellation events, and quantify their relative importance, to assist transportation firms to select the factors to focus on when trying to affect either kind of behavior among customers. Specifically, we show that on the one hand, firms might take a stronger segment-specific approach to customers with low usage (BC25) and medium usage intensity (BC50). On the other hand, results also suggest that it could pay to differentiate between customers on a demographic basis when it comes to activities targeted at reducing downgrading events. We point out opportunities to rethink CRM practices, particularly in terms of electronic mailing policies (e.g., automated mailings, and special offers), as in some circumstances, they have counterproductive effects and inhibit customer relationships. We also highlight opportunities to redesign customer loyalty programs to increase retention and upgrading behavior. We have also revealed the effects of pricing and consumers’ previous contract decisions that led to optimal or suboptimal usage of their loyalty cards. Further, we observe that neither cancellation nor upgrade and downgrade hazards are “memoryless”, since change proportions of these events are affected by the track record a customer has with the firm, that is, by the time passed since the (initial) BahnCard purchase. The contributions of the paper to the literature are: First, it adds to the recent literature on transportation and logistics management that emphasizes the growing need for customer orientation and relationship management (e.g., Ganesan et al., 2009, Grawe et al., 2012, Ramanathan, 2010 and Steven et al., 2012) by systematically exploring how such approaches can be managed in practice, as well as highlighting their specific effects on (un)desired business outcomes in the context of rail travel. Second, consumers’ contractual choices and changes therein have not as yet been studied in the railway literature, although two-part pricing arrangements have become widespread in recent years in transportation and other sectors. The literature on (rail) transportation does not yet provide a comprehensive approach to understanding consumers’ travel behavior, and particularly, the determinants of their contractual choices and usage decisions over time. Accordingly, little is known of how such decisions may effectively be influenced. A better understanding of such linkages is essential when discussing the effectiveness of CRM practices, especially in the context of customer loyalty programs, and if proposing business strategies matched to future market development in the transportation sector. We shed light on these issues and provide new theoretical insights by integrating previous research on CRM and two-part pricing schemes for the transportation sector. Third, semi-parametric proportional hazards models stratified by failure type have not previously been developed to study consumers’ contractual choices in a travel behavior (competing risks) context. Based on this unconventional methodological approach (see also, Li et al., 2012, Smith, 2012 and Wen et al., 2012 for recent methodological contributions in the rail sector), our results would help transportation firms assess the potential effects of promoting, re-developing, and fine-tuning their two-part pricing schemes on consumers’ subsequent contractual choices and usage decisions, both initially and over an extended period. Therefore the current research offers some practical implications for railway companies, particularly, concerning the creation and management of customer loyalty programs and customer relationship management practices when using two-part pricing schemes. Fourth, since we examined an entire population of railway passengers in Western Europe based on extensive longitudinal data, we suggest that findings should be sufficiently robust to be transferable to similar cultural and economic settings. Then, extending the focus beyond the current study context, results may also help forecast performance outcomes and have implications for policy (e.g., on CRM practices and pricing strategies) for suppliers of comparable transportation services, as well as strategic guidance for firms applying two-part pricing schemes in general (e.g., concerning customer segmentation, and effects of demographics and varying usage intensities). However, the study is not without limitations. Future studies could potentially enhance the model framework applied here by adding insights into the consumer perspective by modeling preferences towards alternative means of travel as they relate to contractual choices. A worthwhile focus may be on underlying consumer characteristics and motivations, on cultural aspects, habits, or lifestyle choices (including for example, consumers’ cultural heritage, environmental consciousness, susceptibility to social influence, household income, etc.). Of course, broader changes of lifestyles and attitudes within the population would change our results as well. Nevertheless, environmental consciousness and sensitivity towards oil prices have been steadily increasing throughout the last decade, so that the factors driving consumers’ travel behavior are not expected to be reversed anytime soon (Statista, 2011). Still, our model could also be extended to include aspects related to the market environment and competition among infrastructure providers, thereby assessing contractual choices against the background of economic and regional equity issues in a wider context. From a methodological viewpoint, the adaption of our model towards a multi-stage approach might also prove an interesting variation in future research.