Taking orders over the phone

Editor's note: Cheryl Ross is market research manager with Spiegel, Inc., Oak Brook, Illinois.

Success as a retailer, especially an apparel retailer, has always had a lot to do with good instincts: an eye for the trends and a knack for picking the ones that would dominate a season.

In 1986, Spiegel, Inc. management realized that the game was in the process of becoming much more complex. The market was fragmenting, the competition intensifying and trends proliferating. Instincts would remain crucial but instincts supported by knowledge, specifically market knowledge would be the key to success. That same year Spiegel created its consumer research department to forge a direct link to consumers and gain a continuous, rapid source of up-to-the-minute market knowledge.

As a direct marketer, Spiegel had made excellent use of market testing and its vast database. Using the wealth of database information to develop programs and refining programs before they were rolled out on a major scale was key to much of the company's increase in efficiency. But the company had plans to go beyond cataloging into multi-channel specialty retailing with stores from coast to coast. The company also had plans to go beyond its core market (upscale, educated, working women) and to penetrate new market segments (men, younger consumers, ethnic markets). To achieve these goals, Spiegel needed to merge database information--how consumers behave -with solid consumer research information - why they behave that way -to gain a clear understanding of consumers.

As the learning from initial studies was applied, and as it became clear how significantly profits improved by using the action plans produced from the research, the number of requests from within the company for research studies exploded.

Remain within budget

We wanted to be able to meet this stepped-up demand, especially now that the "key players" at Spiegel had begun asking the kinds of questions we were confident would result in highly actionable plans. Yet we wanted to remain within the budget we had established before demand had risen so greatly. Obviously, the solution was to cut costs dramatically but without affecting the integrity of the data.

My staff and I began closely investigating the research firms whose costs were lower than their competition, as reflected in the more attractive bids they submitted. This exploration revealed that the most cost-efficient firms were those that had leveraged a newly emerging technology - using computers to help in the research process through a computer-assisted telephone interviewing (CATI) system.

We investigated a number of competing firms offering CATI software and spoke with firms using each of the strongest of these software packages. We selected ACS-QUERY based on two key criteria: user-friendliness and the responsiveness of the software producer, Analytical Computer Service.

The primary reason we sought a user-friendly system was to be able to program the questionnaires quickly and to easily pull off responses for analysis as soon as they were available. We felt it was more important to staff our research department with people highly skilled in analysis, management, and marketing strategy than with technical experts devoted to programming and running the computer system.

Specifically, rather than devoting our strongest human resources to producing data, we wanted to focus them on determining what consumer information made the most sense to obtain. The emphasis was to be on analyzing and interpreting the data, distilling the most meaningful findings, and developing solid action plans. This would permit us to translate findings into results that have a direct, measurable impact on profits.

Second, we knew that we were going to test the system by "borrowing" personnel from our customer service division. People would be interviewing only part-time. Thus, we wanted them to be able to use the system almost immediately.

We also realized that because the technology was new -both to us and to the industry -we would have to find a software vendor that was sensitive and willing to react to both new technologies and to innovative uses of the software. We already had given some thought to future needs such as automatic telephone dialers, connecting multiple facilities through a computer network, in-store interviewing, respondent administered surveys, and graphic interfaces. In short, we wanted a software supplier that would respond and grow with us as we developed.

Test phase

Initially, we wanted to test the system's capability to cut costs and produce high quality work with minimal investment. Therefore, we tested ACS-QUERY SOLO, which runs on standalone PCs. While it lacks the sampling capabilities and quota control features that are available on the full system, it gave us the opportunity to try out the system at a relatively low cost.

We tested the system for 18 months by borrowing staff from our customer service center, training 33 individuals to operate the system, and measuring cost savings. Direct, head-to-head comparisons were made by performing parallel studies. That is, we would split a study, with half conducted internally and the other half conducted by an outside vendor. We sought to measure both the cost savings and the quality of the data.

These direct comparisons revealed that research costs dropped to 55% of their previous levels. We also found, for reasons described later in this article, that the quality of the data actually improved.

Full roll-out

Based on the overwhelming success of the tests we recommended (and management agreed) to quickly move from "test" to "full rollout" phase to capitalize on the cost savings and quality enhancements.

We purchased 7 PCs, upgraded to ACS-QUERY (which includes sampling capabilities, quota control and call management features), installed telephone lines, and hired a staff of interviewers to conduct Spiegel studies on an on-going basis. This was in December 1990. Within two months the system was installed, the staff was trained and our CATI system was up and running.

Immensely pleased

Since then, we have been immensely pleased with the results. In fact, while the benefit originally sought -that of cutting costs -has been achieved, three additional benefits have also accrued which are proving equally or more valuable: one, greater depth of understanding; two, accelerated pace of operation, producing research studies much faster; and three, integration of database information reflecting customer behavior with consumer research learning revealing the motivations underlying that behavior. A description of the reasons for these changes and the benefits we derive from them follows.

The first advantage we realized from our CATI system was that the information gathered and provided to management is much more illuminating than data collected using paper and pencil surveys. That is, we get deeper insights from both the open-end and closed-end questions. This is because of three features that CATI offers.

First, the CATI system enables us to develop more complex question patterns with a tighter focus. For example, if we learn that respondent A is a catalog shopper who works full time and has children, we may want to ask her a different set of questions than those we would ask respondent B, whom we discover never shops from catalogs.

The branching process can become complex, as you learn more about the respondent and want to direct certain questions to ever-shrinking subsets of people. For example, if we discover in question 3 that the respondent is a trier-rejector, and in question 5 that the respondent purchased apparel rather than home furnishings from Spiegel, and in question 11 that the respondent currently frequents Spiegel's Eddie Bauer and Honeybee stores, we may want to ask questions 12-15. Otherwise, we may wish to ask only question 14. Obviously, this kind of logic is unwieldy in a paper and pencil survey and results in lengthy surveys and lost data because incorrect surveys have to be discarded. CATI eliminates this problem.

Second, since the pace of the interview is faster when the interviewer uses CATI, more time is available for probing open-end questions. This is critical to Spiegel because we often find that greater insights are gleaned from answers to open end questions than from closed end questions.

The essence of the CATI system is to assign to computers the functions they are designed for and best able to perform, and to allow the interviewers to focus full attention on what humans can do, such as achieving rapport with consumers, gaining an increased level of participation and enthusiasm for taking part in a research study, and probing for more information.

Third, we also can gain deeper insights into respondents' answers because we can integrate previous answers into subsequent questions. This approach has been used for unaided/ aided awareness questions; however, we have been able to go further by asking such questions as: "You said that you consider the three most important factors in deciding which store or catalog to buy from to be style, price, and ease of return. Can you please evaluate the following three stores on each of these features?" Focusing respondent attention on specifics rather than on generalities produces far more incisive answers.

Speed

The second major benefit was that the pace of completing research studies increased dramatically. CATI permits the interview itself to be performed much faster than paper and pencil surveys. In 15 minutes the interviews are producing as much information as the previous 35 minute surveys.

Performing a portion of our research internally accelerates the process significantly by removing the need to deal with an outside firm that has other "urgent" projects, or whose other clients may represent a greater portion of the firm's revenue.

Using CATI also reduces the tedium of the process, and therefore the fatigue of the interviewers, which translates into greater interviewer enthusiasm and output. Individual interviews are performed faster, which means response rates are higher, and each study is completed more quickly.

CATI also eliminates the time-consuming keypunching step of the research process. In addition, CATI makes it possible to get a "quick read" on answers. Because the computer can collapse and summarize the answers nightly, it is possible to gain insights during the investigation, rather than awaiting its final completion. This affords the added benefit of evaluating the content of the survey mid-stream.

Sometimes an assessment of the first 300 interviews reveals that the answers are clear enough to permit eliminating several questions and adding new ones to further deepen the learning.

The net effect of this speed is that management now views research as a real decision-making tool. Rather than being forced to use research only for questions for which we can wait several months for answers, CATI is our tool to make consumer input a "real time" possibility.

Database integration

The third major benefit of CATI was the ability to integrate our database information into the survey we are conducting.

Our database contains behavioral information on our customers that we compile and analyze regularly. For example, we track which ad, promotion, product, or catalog brought the most new customers to Spiegel, to identify what the characteristics of those customers are. We determine which vehicle for attracting new customers produces heavy buyers, light buyers, or trier-rejectors. We also determine what type of products, catalogs and promotions each of these types of customers finds most compelling, whether they obtain and use Spiegel credit cards, whether they begin buying from Spiegel subsidiaries such as Eddie Bauer, Honeybee, Crayola Kids, whether they buy early or late, at full price or on sale, etc.

This behavioral record is an excellent source of research because it provides an accurate record of what people actually do, rather than relying on what they say or recall having done. While this behavioral record is useful, it has a crucial limitation: analyzing past actions alone does not provide an understanding of how they behaved in the total competitive arena - where else they shopped, how much, and why. Nor does it lend insight into why they behaved as they did at Spiegel, or how they would respond to future promotions or new products or catalogs. As Marshall McLuhan would say, it is kind of like driving a car by looking through the rear view mirror - it only reflects what happened in the past, rather than what will happen next.

However, we have found that looking at past behavior combined with current attitudes together produces phenomenal synergy and a clear understanding of what motivates our consumers. This helps us discover not only what they bought, but why. In so doing, we can determine not only which catalog, promotion, ad or product works best, but also how we can improve it. Similarly, we have traveled up the learning curve from promotions or catalogs that were not successful, by discovering what made them weak, and whether that weakness is fixable or something we can learn from and thereby render future efforts more successful.

This is achieved by integrating database information right into the survey. For example: "Mr. Jones, we noticed that you have a Spiegel credit card, which you use when buying for yourself, but when you purchase gifts from Spiegel you use your American Express card. Why is that?"

Summary

Our goal prior to purchasing a CATI system was to complete the conversion of Spiegel into a consumer-driven company in a cost-effective manner. We looked for user-friendly software and a vendor that would support both our current and future needs.

Our mandate was - and is - to strive not only to understand how people behave, but also why they behave that way. When something fails, our charter is to find out why and whether we can fix it. Even more important, when a new promotion, catalog, or advertisement works, we use research to develop an action plan to tell us how we can improve it, thereby boosting the success of the products, catalogs, promotions and ads on an ever increasing cycle.

The CATI system has accelerated our efforts to gain this understanding. The rapid-fire turnaround has made research more valuable to management by providing a direct line to consumers. The cost saving has enabled Spiegel to conduct more surveys while increasing quality and decreasing costs. The flexibility of the system has made it possible to get more in-depth answers. The result is that Spiegel's products, promotions, services, and catalogs have become more consumer driven. We ask consumers what they want and then we deliver it.