Better survey design is: (a) easy (b) difficult (c) don't know: stuck for an answer? Use this seven-element model to ensure that your customer questionnaires are complete and achieve intended objective.

[ILLUSTRATION OMITTED]

Nothing is more important to a bank's success than a rigorous understanding of its customers--that is, who they are and what they want. One of the most valuable tools that a financial institution can use to understand its customers is the questionnaire.

Composing a customer survey is no easy task, however. Banks that are unfamiliar with the art of survey design will frequently resort to questionnaire templates originally designed for use in other financial institutions. The problem with the use of "canned" surveys is that they may not address the unique information needs of your particular bank.

Financial institutions that attempt to develop their own questionnaires from scratch quickly discover that the process involves a lot of compromise, squabbling--and frustration. The resulting questionnaire usually ends up providing less than satisfactory information.

In this article, we will offer a model for creating simple, high-quality in-house surveys. We will outline the seven elements that cover all the components, in one form or another, in every effective customer questionnaire. By employing this model, banks will be able to create surveys that cover all the important bases--including questions for special needs.

What questions should a survey ask?

Anyone attempting to develop or compose a questionnaire first must become familiar with the precepts of survey methods, including how to develop the objectives and the questions; how to administer the questionnaire; how to determine sample sizes; and how to design the wording of the selected questions.

One issue that is rarely addressed is what questions to ask in the first place.

Some time ago, we undertook a study on this issue. We started by examining several hundred marketing and consumer surveys. We discovered common elements: All questions could be categorized in seven distinct categories.

Based on this research, we developed a model for the creation of questionnaires that incorporates all seven of these design categories. For several years, we have employed this design model in our own professional research and as consultants with other research endeavors. We have tested the model by requiring others developing a questionnaire to proceed without knowledge of the model and then review their results against the model. And, we have engaged one research group with the model and another group without the model and compared their results and their assessments of the difficulty of the task. We have concluded that the use of the design model greatly facilitates the creation of surveys and also improves the quality of the final product.

The survey model

This model (we call it The Ambrose/Anstey Questionnaire Design Model) consists of seven different elements. The listing of elements is not in any particular order nor is the questionnaire designer instructed to follow the model according to this sequence. Likewise, the categories are not mutually exclusive, as you may find an issue could fit into one category as easily as it fits into another. The intent and value of the model is that it directs and focuses the development of the survey instrument to the various categories of questions that comprise all questionnaires and, having addressed and satisfied each category, the questionnaire is considerably more complete. Concurrently, employing this category model greatly facilitates the design process and grants a greater degree of assurance that the resulting survey instrument is complete.

This model is appropriate for the surveying of different banking populations: depositors, investors, commercial, retail, large versus small accounts, active and prospective customers, and window customers compared to postal customers. We have not encountered a research situation in which the model is not appropriate.

The seven elements:

1. Demographics: the numeric and discrete measures defining the characteristics of a bank customer or prospective customer. The demographic measures are not limited to those collected in the U.S. Census, but should be those dimensions that are most meaningful to the bank and the survey objectives.

2. Behaviors: categorized measurements of the actual activities of the individual or business. These measurements are numeric, categorical and discrete values. The purpose is to determine the past and current banking relationships in a form that clarifies through categorization the various interactions the bank experiences with its banking customers.

3. Attitudes: assessments by the respondents in a scalar form of their impressions and levels of satisfactions. Attitudes are not absolute values but are relative and therefore must be compiled using a scalar form of question where a categorized question would be inappropriate and invalid.

4. Knowledge: measurement of the levels of understanding of factors that are considered important in a banking relationship. These questions can be either scalar or discrete in their design. Likewise, fictitious issues can be included to determine the level of competence of the respondents to complete this or other portions of the questionnaire.

5. Predispositions: evaluation of the tendencies of respondent toward behavior changes. Because of the impossibility of respondents to answer these questions with preciseness or absolute certainty, a scalar design is required. Predisposition questions have many of the ,same characteristics as attitudinal questions.

6. Propositional or conditional: determining the probabilities associated with possible behaviors of events if conditions were to change or changes were introduced. One might consider these questions as testing the hypothetical nature of banking going forward whether these conditions occur naturally or are influenced by the banking institution. These issues are best designed using a scalar form of question.

7. Intentions: declaration of expected behaviors, conditions, decisions or issues of the customers in the future. Respondents provide their best possible estimates of their future and its relationships to the bank and financial industry. As these are estimates, questions that provide a latitude of response or a scalar response are most appropriate.

Some people might object that the model could create confusion as one might not be able to identify a clear delineation between the categories. If I go to the branch bank twice a week, is that a "demographic" or a "behavior" question? Or, if I desire to purchase a new automobile next year, is that an "intention" or an "attitudinal" question?

[ILLUSTRATION OMITTED]

The assignment of these questions to a specific category is not important. You could make an argument that either answer is correct. The importance of the categories of the model is not the need to assign questions to a category but to prompt the designer of the survey instrument to include the question. The prompt should be: "Is there a demographic factor that can be relevant in fulfilling the objectives and answering the issues of the study?" That is the exact purpose and limit of the categories and of the model.

The simplicity of this model can be deceiving. One's first impression is "So what?" "Isn't this just a simple itemization and does not help in resolving the complex problems in questionnaire design?" We agree with the simplicity. It is not a repeatable model, nor is it a model premised upon parameters, formulas or abstract conditions. It is easily understood and easily followed. Regardless, part of the value is its simplicity. In addition, value is derived by the completeness of its categorizations and by its contribution in avoiding and precluding many of the complex struggles that surface in questionnaire construction.

Applied to the banking industry

Using the listed categories of the design model, the two tables shown on the previous pages are a demonstration of the application to the design of banking related questions. Because of the clear distinction between retail customers and commercial customers, we have developed representative questions for both segments. However, before reviewing the questions, we must remember the questions are illustrative only and may not be universally applicable, nor are these questions exhaustive of the examples that could be constructed for each of the categories of the model.

All, some or none of those questions or categories may be relevant to a particular bank. The value in following the model is "the question of the category is asked and answered"--not that a questionnaire must rigidly include all categories in its design.

Obviously, just as there may be no relevant questions in a given category, there may also be multiple questions in a category. If we were to examine motivations extensively, we would probably have many questions that are attitudinal in nature. Or, if we were attempting to develop a profile of our customer base to match it to a large data file of prospective customers, we would ask a variety of demographic and behavioral questions that match the data file.

In reviewing this model against existing questionnaires, we have not found any question that is employed in surveying customers that is not addressed by the model. That gives us the comfort in knowing the model is exhaustive and fully inclusive of all relevant types of questions.

The model does not specify the wording of questions, design response scales or provide a method of sequencing the questions in a study. Those issues and concerns are well covered by the existing documentation of questionnaire designs and marketing research. Nor does the model instruct on the proper method of managing open-ended questions, selecting a representative sample from the study population or assessing responses statistically. Those dimensions of research are also well documented. The model does facilitate the design of the study instrument and assures the designer the questionnaire is more complete than would have been possible without use of the model.

A principled process

Questionnaire design is a challenging and complex task. It has long been recognized that developing a questionnaire is not a science but largely an art. The rules and guidance that have been developed over the years have been intended to move the process to be less reliant upon artful creativity and more oriented to a principled process.

We believe our model is a contribution toward moving questionnaire design from the artistic side of the scale to the rule-based (scientific) side. This model is not sufficient to accomplish that totally but does assure the designs of questionnaire are more complete and reflective of objectives than those developed without the model. And, in moving from the artistic orientation, the model greatly simplifies the process without compromising the designers' desires for independent or customized inclusions.

David M. Ambrose and John R. Anstey are faculty members of the Department of Marketing and Management, College of Business, University of Nebraska at Omaha, Neb. E-mails: ambrose1@cox. net and [e-mail]

Chart #1

Consumer/Retail Customer Survey

Examples of retail customer survey questions that
employ the seven categories of questions.

Categories

Demographics        In what year were you born?

Behavioral          Approximately how many visits
                    each year do you make to you
                    main bank?

Attitudinal         On a scale of 1 to 6 (1 is low,
                    6 is high), how important is a
                    drive-up window at the bank?

Knowledge           What does National Bank charge
                    you for each check you write?

Predispositions     What are the most important
                    reasons your neighbor selected
                    their checking account bank?

Conditional/        If you were to consider another
Prepositional       bank, what factors would be
                    most important to you?

Future Intentions   How likely are you to open a
                    savings account for your child
                    (or children) within the next six
                    months? (On a scale of 1 to 6)

Chart #2

Commercial Customer Survey

Examples of commercial customer survey questions that
employ the seven categories of questions.

Categories

Demographics        What was the corporation's total
                    gross revenue in dollars for last year?

Behavioral          Approximately how many
                    international currency transactions
                    do you have each year?

Atitudinal          What is your level of satisfaction
                    not having the actual checks
                    returned to you each month?
                    (1 is low, 6 is high).

Knowledge           Which of the following services
                    are provided by National Bank?
                    Safe Deposit           Y   N   D K
                    24-hour drop box       Y   N   D K
                    AR collection boxes    Y   N   D K

Predispositions     How important is the relocation of
                    National Bank to your continued
                    relationship with the bank (on a
                    scale of 1 to 6).

Conditional/        If you were to require the services
Prepositional       of a corporate trust officer, what
                    different financial institutions
                    would you consider?

Future Intentions   Do you expect in the next year
                    to open additional retail locations
                    that would require that you consider
                    a different bank than National Bank?
                    (on a scale of 1 to 6).