Marketing Research Process, Part 2

As we started to discuss last month, commonalities among marketing research projects enable us to organize the process into six major steps. The value in characterizing research projects in terms of steps is that, first, it gives non-researchers and researchers an overview of the entire process, and second, it provides the researcher tasks to consider and in what order. Here are the final three steps in the marketing research process.

Step 4: Collecting the Data

Once the questionnaire is finalized, the methodology decided in step 2 is used to launch the survey into the field. Although this step is fairly automated throughout (phone interviews, online surveys) the data collection process, it is important to keep a close eye on the number of people qualified to take the survey, also known as the incidence rate. Because the accuracy of the data depends on meeting the quota (the number of people needed to take the survey), low incidence rate can cause the cost of your study to spiral out of control, since you must do more interviews per survey complete.

Depending on the type of methodology used, collecting the data can take over half of the project�s timeline. Although rising social networks and online panels have led to faster and easier data collection through web, phone interviews continue to generate higher response rates.

     Here are the pros and cons to consider when choosing your survey methodology:




Because they use paper and pencil instruments, mail surveys are a less expensive way to collect data from large number of people. Also, because of the convenience of completing the survey at their own pace, a longer questionnaire design is plausible. Since mail surveys are self-selected and self-administered by the respondent, there’s no follow-up from the researcher. They also have the least amount of control over all other survey methodologies in terms of questionnaire design and clarification of questions. The time getting the survey back by mail, the time entering the data, and time analyzing the data are extensive and the completion rate can be very low.




Computer-Assisted Telephone Interviewing

High quality control with trained interviewers and CATI technology to assist in complex questionnaires. Clarification of unfamiliar words and questions are carried out by interviewers. Data collection is quick and less prone to errors. It’s easier to reach “lower incidence” respondents. It is also less prone to “non-response” bias and more likely to represent the true opinions of the population than any other methodology. In general, the telephone survey needs to be short, usually less than 15 minutes, for respondents to be engaged and to complete it without breaking off. It’s difficult to ask sensitive and personal questions in a telephone interview. Hurdles such as the FTC’s Do Not Call List for telemarketers and inexpensive gadgets such as TeleZappers make reaching respondents more difficult. Cost per telephone interview can be expensive if the completion rate and response rate is low. There’s also the possibility of interviewer bias when probing for answers.



Interactive Voice Response

No interviewer results in lower cost and no interviewer bias. Convenient for respondent to dial in anytime at his convenience. Used in transactional surveys, IVR provides the fastest feedback after interaction with an agent/company. Survey needs to be very short, 5 minutes or less. It tends to have low initial cooperation rates and high break-off rates. No clarification of questions and probing for reasons are possible.




Although there can be high software and hosting start-up costs involved with launching an online survey (unless a professional marketing research firm like Polaris is used), the cost of online is less expensive than a comparable CATI survey. Time spent collecting data is greatly reduced since data is automatically entered in a database and can be exported to other programs. Online surveys are as powerful as CATI, cover international populations, and can display many different media types. Coverage that is representative of the population is still assumed to be inferior to telephone sampling, but this is changing as broadband access increases and landline telephone penetration decreases. Email invites can end-up in junk mail boxes and people are paying less attention to emails as they used to. Self-selection bias exists and privacy concerns may prevent respondents from responding.



The completion rate is high. Respondent is unlikely to “drop off” before completion. The interviewer can acquire more qualitative data, and explore answers with respondent. It’s optimal for difficult to reach populations. Cost is very high due to interviewer training, traveling, and offering respondent incentives. Face-to-face surveys have almost become extinct due to cost compared to telephone or online methods. It has the highest degree of interviewer bias, in verbal and also in this case, non-verbal communications. Also, respondents may feel reluctant to share truthful answers to sensitive topics and give a more socially acceptable response to the interviewer.

Step 5: Performing Data Analysis

Data cleaning and coding for analysis requires careful handling of the data set. Setting-up a quality check process will prevent misrepresentation of the data. Data analysis is where you will get the benefit of having spent a little more time finalizing the questionnaire, because analyzing the data depends heavily on how the questionnaire was constructed. For example, if the data shows 20 percent of women are more dissatisfied than men, what are the reasons for this dissatisfaction? If the questionnaire had a follow-up open-ended question for those who gave a low satisfaction rating, the verbatim can be coded and dissatisfaction by women versus men can be quantified.

Toplines or basic frequency counts of the data can be obtained by standard office suite tools. For more advanced analysis, such as multiple regression or factor analysis, statistical software and the skills to translate the data into actionable items are required. There are infinite ways to slice and dice the data by cross tabulation analysis that will either give you too much or no information at all. In this step, reviewing your objectives again can help you strategize your approach to piecing out the most actionable information.

Creating sound research for information and decision-making is more involved than many managers understand and it is the role of the researcher to impose discipline and structure on the process. By following the six steps of the marketing research process, the research can help managers design the optimal research project to meet business objectives.

Step 6: Reporting and Presentation

The last step of the marketing research process, to prepare and present the final research report, is also one of the most important. This phase cannot be understated because it communicates the study results to the client. If the objectives stated in the early stage of the marketing research process are communicated clearly and the information gathered satisfies those objectives, the client can launch a major and highly successful marketing initiative or product.


All Content � Copyright 2004-2011. Polaris Marketing Research, Inc. All Rights Reserved.

Send inquiries to or call 1-888-816- 8700.