Using Choice Modeling in Service Management

Product-development and marketing managers know that customers make purchasing decisions on the basis of many criteria, including service quality, delivery speed and price. But since no company can excel in all aspects of service delivery simultaneously, companies must make trade-offs on the basis of what they do best, what criteria matter most to their customers, and what their competitors are offering.

An August 2002 white paper, “Understanding Customer Value Drivers: A Key to Successful Service Management,” offers managers a way to assess customers' choices in order to optimize return on their product-development and customer-service investments. Authors Rohit Verma, associate professor of operations management at the University of Utah's Eccles School of Business, and Gerhard Plaschka, associate professor of management at DePaul University's Kellstadt Graduate School of Business, base their framework upon choice modeling (CM) — an approach first introduced in the 1970s by Daniel L. McFadden, Nobel Prize-winning professor of economics at the University of California, Berkeley.

McFadden's research focused on both the economic reasons for individual choices and the ways researchers could measure and predict these choices. He began with the macroeconomic theory that says people will act to maximize their self-interest and applied it to the most complex microeconomic choices, such as why people take certain jobs, whether they marry, and how they travel to work. McFadden's work subsequently led to applications in various fields, including marketing, natural resource economics and transportation, as well as service management.

Verma and Plaschka believe that most companies lack a clear understanding of their relevant customer and market drivers and, as a result, their product or marketing managers tend to take a scattershot approach, hoping that at least one product-service offering will succeed. Using McFadden's theories and subsequent work on experimental choice analysis (introduced by Jordan Louviere of the University of Technology, Sydney, and other researchers), the authors developed a three-step process to assess and challenge management preconceptions of how customer offerings will be received.

CM links theoretical behavior —observed in experiments, surveys and other forms of stated preferences — with behavior observed in real-life situations. Unlike conjoint analysis (in which study respondents rank their preferences using experimental profiles), CM requires that respondents make choices in simulated situations derived from realistic variations of expected market offerings. The process comprises three steps.

First, using qualitative market assessment, customer interviews, case studies, industry data, focus groups and other information sources, managers compile a list of market-value drivers that they believe influence customers' buying decisions. For example, a hotel manager might identify choice drivers and value levels such as “hotel type” (motel, bed & breakfast inn, boutique hotel, convention hotel), “loyalty program” (hotel points, merchant points, airline frequent-flier miles), “amenities” (in-room business center, central business kiosk, anytime check-in/check-out), “eating options” (full-service restaurant, breakfast-only restaurant, in-room kitchenette), and “price” (weekday rate, weekend rate, all-inclusive pricing options). Participants must collectively finalize a list of drivers, limited enough not to overwhelm potential study respondents but realistic enough to reflect the actual market.

Next, managers construct “choice” experiments, asking respondents to select a set of options from a bundle of product and service attributes. For example, a hotel operator might describe two possible hotels, listing a number of market drivers and specific value levels for each. Respondents then answer questions about the choices — for instance, “If these two hotels were your only alternatives, which one would you choose?” “If Hotel #1 were your only option, would you go there?” “What do you consider the most and least attractive features of each hotel?”

Because the number of potential attribute configurations can multiply quickly — 10 drivers with four value levels each can generate more than a million alternatives — experiment designers can employ statistical techniques (such as fractional factorial design and blocking) to estimate the impact of major drivers; then they can divide these choice exercises into statistically equivalent subgroups.

In the final phase, econometric models (based on discrete choice analysis) are used to identify key patterns in the survey responses, providing relative weighting for each market value driver and for interactions among drivers. Managers can then select the optimal combination of operations and market drivers in order to develop a profitable and sustainable value proposition that, under normal competitive constraints, will maximally leverage their available resources.

Like other modeling processes, choice modeling is subject to the “garbage in, garbage out” rule. It generates useful information only if the assumptions behind the selection of value drivers, the experiment design and the data-collection methods are sound. If these conditions are met, Verma and Plaschka suggest, CM can yield valuable insights for market-strategy development by revealing customer clusters, suggesting the potential impact of changing the levels of value drivers, assessing overall brand equity and identifying barriers to brand switching. Moreover, CM can reveal any salient differences between managers' beliefs about customers' needs and their actual needs.

In a number of projects, the authors have applied CM to several industries, including telecommunications, financial services, industrial services, outdoor recreation and banking, among others. They conclude that CM analysis can help managers design attractive offerings and hone their company's operations. For example, if CM reveals that hotel customers value anytime check-in/check-out over a customer-loyalty program, managers could improve the back-end processes that support faster check-in/check-out service. For managers eager for reliable feedback on how consumers view their offerings, the authors say, CM provides a rigorous way to turn that feedback into profitable and sustainable strategies for retaining or capturing market share and profitability.

For more information on the paper or related work, contact the authors at rverma@allgrp.com and gplaschka@ allgrp.com.