Evaluating channel performance in multi-channel environments

https://doi.org/10.1016/j.jretconser.2006.02.001Get rights and content

Abstract

Evaluating channel performance is crucial for actively managing multiple sales channels, and requires understanding the customers’ channel preferences. Two key components of channel performance are (i) the existing customers’ intrinsic loyalty to a particular channel and (ii) the channel's ability to attract switching customers. We apply the Colombo and Morrison (Colombo, R., Morrison, D., 1989. A brand switching model with implications for marketing strategies. Marketing Science 8, 89–99) model to assess channel performance along these dimensions. Using data from a large home-shopping company, we analyze the evolution in the performance of its main channels over time, and test for differences in channel performance among different product categories offered by the company, as well as between different customer segments. Based on the results, we derive implications for managers to operate a company's multiple sales channels more effectively.

Introduction

More and more companies become multi-channel operators (Ganesh, 2004; Coelho et al., 2003). Therefore, managers need metrics that help them assess the performance of each individual sales channel, as well as the inter-relationships among the different sales channels in their portfolio. Preferably, these metrics should be grounded in marketing theory and should be objective, based on readily available data, easy to quantify, intuitively appealing, and should have diagnostic value (Ailawadi et al., 2003). Furthermore, these metrics should be based on customers’ channel preferences (Reardon and McCorkle, 2002). A key component of customers’ channel preference is their behavioral loyalty to each sales channel. Even though it was initially argued that customer loyalty would disappear in near-perfect Internet markets (Kuttner, 1998), recent research has convincingly argued that customer loyalty remains central to long-run profitability.1 However, channel loyalty is only one component of sales channel performance, because some customers may not have a strong intrinsic preference for any specific channel, and occasionally (or even regularly) switch among different channels. Therefore, a second key characteristic for sales channel performance is the ability to attract these switching customers (Reichheld and Schefter, 2000).

Insights into these issues can be obtained through the analysis of aggregate switching matrices (see, e.g. Lilien et al. (1992, pp. 41–42) for an in-depth discussion on the construction and interpretation of such matrices). Repeat purchases with the same sales channel then appear in the diagonal elements of the matrix, while switching between the various channels, or defection to competing companies, is reflected in its off-diagonal elements. However, not all observed consecutive purchases through the same sales channel reflect the same level of channel loyalty. Indeed, some repeat purchases could come from customers who buy repeatedly with probability one, and hence do not consider the use of other sales channels or companies. But a repeat purchase could also come from customers who have considered other sales channels or companies, but have not yet switched. Terech et al. (2003) refer to the former as hard-core loyals, and to the latter as soft-loyal customers. Obviously, the former are the more attractive customers for a given sales channel.

Various stochastic models allow for a further decomposition of the aggregate switching matrix beyond the simple diagonal/off-diagonal dichotomy (e.g. Bayus, 1992; McCarthy et al., 1992; Colombo and Morrison, 1989; Grover and Srinivasan, 1987). A well-established model is the one developed by Colombo and Morrison (1989). Its empirically determined parameter estimates have clear managerial interpretation and diagnostic value. Furthermore, its data requirements are low and its implementation and interpretation are straightforward, thereby satisfying the various criteria advocated in Ailawadi et al. (2003).

In this paper, we apply the Colombo and Morrison model to assess multi-channel performance. We illustrate that it allows for more detailed insights into customers’ channel preferences by decomposing the observed switching behavior among a company's sales channels into (i) customers’ intrinsic loyalty to a particular sales channel and (ii) each sales channel's ability to attract potential switchers, referred to as its conquesting power. In doing so, we also distinguish between hard-core loyal and soft-loyal customers to account for different levels of loyalty (Terech et al., 2003). In addition, we not only investigate the overall channel performance, but also evaluate whether there are (i) changes in a channel's loyalty and conquesting power over time, (ii) specific product categories that are better suited to be sold over a particular channel (Inman et al., 2002; Morrison and Roberts, 1998), and (iii) differences between specific customer segments. To that extent, we distinguish between heavy and light users. Lim et al. (2005) argue that light users are more likely to attribute their channel choice to external causes than to an internal cause—such as their intrinsic preference for that sales channel.

Those more detailed analyses provide multi-channel managers with additional insights into the performance of a company's sales channels. Thus, analyzing changes over time enables managers to evaluate whether customers start to migrate from one sales channel to another. Moreover, knowing the product-channel association helps to better tailor the assortment to the different sales channels, and additional insights into different customer segments allow managers to better target their multi-channel marketing activities.

The remainder of the paper is organized as follows: Section 2 outlines the research methodology. Section 3 describes the data set. Empirical results are given in Section 4, and Section 5 provides conclusions and implications of our results.

Section snippets

Method

We use the model developed by Colombo and Morrison (1989) to examine customers’ behavioral channel loyalty and inter-channel switching behavior. The model is based on the assumption that there are two groups of customers:

  • customers who are intrinsically loyal and stay with the same sales channel, called hard-core loyals, and

  • customers who potentially switch from one sales channel to another on every purchase occasion. They choose between the available sales channels according to a zero-order

Data description

We observe data from a large European home-shopping company that owns two direct sales channels: a call center and an Internet channel. Customers have two possibilities to purchase a product. First, they can watch TV shows featuring a subset of the company's products on a privately owned broadcasting channel. The company alternates between a number of shows, which each focuses on specific product categories such as cosmetics or jewelry. During the shows, the call center's phone number is

Empirical findings

As indicated in Table 1, the overall number of transactions is decreasing (−2.9%). However, the number of transactions conducted on the Internet channel is gradually increasing, in both absolute and relative terms.

An intuitive metric to evaluate channel loyalty would be to consider the diagonal elements of the aggregate switching matrix, which represent the hard-core loyal as well as the soft-loyal customers. Using this metric, 94.7% of the current users of the call center and 67.9% of the

Conclusions

In this paper, we proposed the Colombo and Morrison model to assess the performance of the various channels in a company's portfolio. Its implementation and interpretation are relatively simple, and the data requirements are low. As such, the method is applicable even in companies that only have a limited amount of data available. Moreover, we have shown how additional insights can be gained by splitting the data set in several sub-samples, e.g., according to different time periods, segments or

References (35)

  • J. Chiang

    A simultaneous approach to the whether, what and how much to buy questions

    Marketing Science

    (1991)
  • F. Coelho et al.

    Exploratory evidence of channel performance in single vs multiple channel strategies

    International Journal of Retail and Distribution Management

    (2003)
  • R. Colombo et al.

    A brand switching model with implications for marketing strategies

    Marketing Science

    (1989)
  • P. Danaher et al.

    A comparison of online and offline consumer brand loyalty

    Marketing Science

    (2003)
  • R. Dholakia et al.

    Switching to electronic stores: consumer characteristics and the perception of shopping benefits

    International Journal of Retail and Distribution Management

    (2002)
  • Flavián, C., Guinalíu, M., Gurrea, R., 2006. The influence of familiarity and usability on loyalty to online...
  • J. Ganesh

    Managing customer preferences in a multi-channel environment using web services

    International Journal of Retail and Distribution Management

    (2004)
  • Cited by (64)

    • Financial consequences of adding bricks to clicks

      2023, International Journal of Research in Marketing
    • Online channel adoption in supermarket retailing

      2021, Journal of Retailing and Consumer Services
    • Multichannel integration quality: A systematic review and agenda for future research

      2019, Journal of Retailing and Consumer Services
      Citation Excerpt :

      Managers need these metrics to assist them in analyzing the performance of each marketing channel and also the interrelationship of different channels (Gensler et al., 2007). Gensler, Dekimpe and Skiera (2007) used brand switching model (Colombo and Morrison, 1989) in regards to channel loyalty to evaluate channel performance. Neslin and Shankar (2009) propose to include a cross-elasticity matrix to assess the impact of one channel on sales activity of another channel.

    View all citing articles on Scopus
    View full text