Editor’s note: Michael Latta is executive director of YTMBA, a Wilmington, Del., research and consulting firm, and is the William J. Baxley, Jr. applied business professor and associate professor of marketing in the E. Craig Wall Sr. College of Business Administration at Coastal Carolina University, Conway, S.C.

Marketing research has been described by many practitioners as a business where “The urgent drives out the important.” Many brand managers feel extreme time and budget pressures when making decisions about promotion campaigns for target markets and providing segment profiles to both ad agencies and upper management. This situation leads to a need for rapid-response data collection methods.

Increasingly, marketing research professionals have turned to multiple sample frames to obtain consumers in short-timeline projects. Surveys that used to take a month to deliver results now are expected in half the time or less and always at lower costs. Multiple-sample-frame surveys provide broader access to highly specialized or target consumers in simultaneous field periods meeting rapid response requirements.

Recently, panels and proprietary databases have been used to complete rapid-response surveys. Consumer panels have become available for both Web and phone surveys provided by companies such as National Family Opinion (NFO) and Greenfield Online. These panels are composed of specific consumers who have agreed to participate in surveys at certain times over an extended time period. In addition to panels, product category-specific databases have been built by companies for their own use through toll-free numbers attached to consumer relationship management (CRM) programs; business reply cards from magazines; journal advertising; product Web site registration information; surveys; and a variety of other sources. This trend is reflected in reports of increased use of client-supplied samples and panel samples supplied by marketing research companies doing field work for consumer surveys.

In tracking the use of multiple sample-frame-surveys from 2004 to 2007 (Macer and Wilson, 2008), research has shown a high use of client-supplied (77 percent), and third-party or access panels (74 percent) as sources of samples for multiple-sample-frame surveys. These high utilization rates are taken to indicate consumer research is an important use of panels and client databases for marketing research surveys, with access panels showing the greatest increase in utilization (+18 percent) over the four-year period. The present survey involved these two types of sample frames (client and access panels) in a rapid-response consumer segmentation study.

Better address the issues

The company commissioning this study has been serving the consumer market with a “gold standard” product since 1978. During that time many marketing research studies had been undertaken involving consumers in this market, but no research study had focused on segmenting the market. The company now wanted to better address social and psychological issues unique to the target market, and needed to do so quickly to inform the strategic marketing plan for the next year.

The overall client goals of this study were to understand consumer needs, information requirements, concerns and problems and to provide the company with publishable results. More specifically, the objectives of this research were to determine consumers’ current knowledge of and attitudes toward their problems; to gain greater insight into common concerns or problems of the target market; to identify concerns and misunderstandings concerning the company’s product; and to identify methods and resources favored for communication by the target market.

Method

In order to meet the objectives of this research, a division of National Family Opinion was commissioned to conduct a two-phased study. The qualitative stage consisted of five focus groups to define the survey instrument. For the quantitative portion, a survey was conducted among a national sample of 504 consumers. Results of Phase I (qualitative focus groups) research were utilized to structure content areas for quantification.

The quantitative consumer sample was drawn from two frames: a proprietary toll-free call-in service database of 1,913 consumers maintained by the company and a 2,208-member NFO consumer access panel.

Consumers represent a random mix from the above databases. Computer-assisted telephone interviewing was used to collect data from consumers in this survey. Each interview lasted approximately 25-30 minutes and each consumer received an honorarium for cooperation.

Results

Sample demographics

Because of the nature of the client’s marketing interest, all consumers were women and approximately 60 percent of the sample is between 60 and 75 years of age. Half of the consumers have a high-school education or less. Six in 10 are married; about one in four are widowed. The bulk of these consumers have incomes below $40,000. Four in 10 have incomes less than $20,000 while a similar number have incomes of $20,000 to $40,000. Ethnically, the consumers in this sample are primarily Caucasian (91 percent). The minority population is somewhat underrepresented by African-Americans (4 percent) and Native Americans (4 percent). Consumers cut across various residential settings. Slightly over four in 10 live in the suburbs, one-third live in urban areas while about one-fourth classifies their area as rural.

Survey performance measures

The four standard performance measures of survey sampling appear in Table 1. As can be seen from the data in the table, the NFO Consumer Access Panel (NFO-CAP) was more efficient in converting dials into contacts (44 percent vs. 39 percent) and converting contacts into completions (64 percent vs. 35 percent) compared to the company toll-free number.

Figure 1 presents the cumulative completes for the individual sample frames and the total sample. As can be seen in the rapid growth of completes, the survey was finished in four days, and could have been cut off at three days with no loss of precision.

Discussion

Academics and practitioners alike are concerned about consumer origin and its relationship to results from a survey. For example, it has been argued (Krosnick, 1999), that survey research methods need a change in standard operating practices such as systematic, representative sampling and high response rates with post-survey weighting used to maximize representativeness. It would be nice if clients would pay for such insurances of data quality. But that kind expenditure only comes with government contracts.

On the practitioner side, as noted by Nelems (2007), the data could be weighted to reflect any differences in the two sample frame profiles, but typically in samples of this size it does not make a meaningful difference in conclusions one might draw from the results of the main measures. In addition, the chief statistician at Knowledge Networks argues that survey researchers need to be clear about the nature of the access panel used for sampling (DiSogra, 2008).

The differences in sample performance between volunteer access panels (VAPs) and non-volunteer access panels (NVAPs) are real and important according to this argument. Clearly, the sample performance reported here indicates the NFO-CAP (NVAP) sample frame was somewhat better than the company toll-free number (VAP) sample frame in efficiency of contacts and conversion to completions. But, survey methodologists, clients are only interested in the final combined results.

The director of respondent cooperation at the Council for Marketing and Opinion Research (CMOR) has argued that sample quality issues related to consumer cooperation, consumer engagement, consumer coverage and presence of professional and fraudulent consumer survey participants are all connected to the origins of the consumers in a sample frame (Glazer, 2008). Consistent with CMOR, practitioners of marketing research strive to meet the new standards of operation adopted by the research industry including:

  • improving the representativeness of access panel samples;
  • decreasing the field time for a survey;
  • lowering the costs of research to deliver higher value;
  • being concerned with and protective of privacy of access panel members; and
  • complying with any privacy legislation.

In the context of the sample performance reported here, these concerns seem to be more intense for the company toll-free number sample frame than the NFO-CAP sample frame. However, use of both sample frames yielded a field period of only four days, leveraged the company sample frame in providing lower costs and added value in meeting rapid response requirements by getting the right people for segmentation as the client requested.


References

DiSogra, C. (2008), “Trains, Planes and Automobiles.” Alert! Magazine, August, 36-37.

Glazer, P. (2008), “Data Quality Seen Through Respondents.” Alert! Magazine, August, 35.

Krosnick, J.A. (1999), “Survey Research”, Annual Review of Psychology, 50, 537-567.

Macer, T. and Wilson, S. (2008), “Do Something About PowerPoint!” Quirk’s Marketing Research Review, March 2008, 60-65.

Nelems, J.H. (2007), “Telephone vs. Internet Data Collection: A Case Study.” Quirk’s Marketing Research Review, December 2007, 22-24.