The Marketing Research Process
1. The key to good research is good thinking about the decisions facing the consumer marketer and following this through into the research process ).
Analyse the market and the competition
1. Examine the tools the marketer has available (e.g. new product, pricing, method of distribution);
2. Work out what is needed to achieve to be financially successful
3. Consider what the competition might do;
4. Examine the social and economic environment (e.g. how much time and money will households have for the new service).
Identify strategic marketing decisions.
1. Usually there is a hierarchy of decisions to make
2. These decisions have to be realistic in the context of the company (e.g. budgets).
Identify the information needed to make decisions.
1. The higher the risk, the more likely that consumer research is required.
2. Some of this information will come from intra-company sources (e.g. the costs of starting a new service, reports from sales staff), but much will need to be collected externally.
Analyse available secondary data.
1. Often some or all of the information required is already available)
2. Consumer research is expensive; there is no point duplicating what is already known.
Design a primary research study.
1. Assuming key information is not available from secondary sources, consumer research may be required.
2. Qualitative techniques such as depth interviews, focus group discussions (and more rarely) projective techniques are often used, either on their own or in combination with quantitative methods.
3. Depth interviews consist of one on one interviewing where a topic is explored in detail, so the researcher can understand the feelings, thinking and understanding of the consumer. The researcher has a topic list, rather than a structured questionnaire, to guide the interview.
4. Projective techniques (e.g. the Thematic Apperception Test) assume that consumers will display their true (even unconscious) feelings when reacting to an ambiguous stimulus presented by the researcher.
5. In a focus group, the researcher is hoping that the interaction between people in the group will uncover many of the basic ideas, feelings and reactions to a topic of interest. The researcher acts as a group moderator, guiding discussion into areas of interest, while ensuring that the researcher’s views do not influence the group.
6. All three qualitative methods call for great interpretive skills on the part of the researcher.
7. Quantitative methods are required where the marketer requires a good representation of the market (e.g. how many people will buy a new service), or where sophisticated data models need to be constructed.
8. Major techniques include observation, experimental research and surveys.
9. In observational research, data is collected by observing what customers do rather than by directly collecting data from them. Some observational methods can be qualitative (e.g. judging the consumer’s look of delight when receiving a gift), but may also be quantitative (e.g. recording the length of time a buyer/consumer in the supermarket glances at a new product display).
10. Shadow shopping is an observational technique where the researcher pretends to be a customer in order to observe the service that is actually offered
11. Experimental designs are used where a direct comparison is required, e.g. in taste tests of rival soft drink ranges. Typically the researcher systematically varies the independent variable (e.g. type of product, order of presentation of alternatives) to gauge differences in the dependent variable.
12. While experimental studies allow the effects of independent variables to be measured accurately, they are sometimes conducted in laboratory settings and may be unrepresentative of the actual marketing environment.
13. Direct marketers can experiment with systematically varied offerings (e.g. price premium) to matched lists of consumers, to gauge their effectiveness.
14. Survey methods are one of the most commonly used research techniques. Surveys ask questions – its what most people think marketing research is
15. Contrary to what some people believe, a survey is not a necessary part of market research! It is a tool like any other, to be used if it provides the right information.
16. Many surveys use personal interviews either in the home or in public places, such as shopping centers (often called intercept interviews).
17. Mail surveys are conducted by sending self-completion questionnaires to a sample of respondents. Group interviews also can use self-completion questionnaires, as can employee surveys. Diary panels collect data over a long period. The Internet is now being used to conduct self-completion surveys among specific target groups.
18. Telephone surveys have become the most common survey method, largely because of the high penetration of telephones into consumer homes, the speed of fieldwork and the ability to obtain accurate samples.
19. CATI (Computer Assisted Telephone Interviewing) allows the research process to be automated.
20. The main survey methods vary in terms of cost and timeliness and to some extent, the type of questions which can be asked (e.g. telephone interviews do not allow the interviewer to show the consumer a copy of a magazine advertisement before asking questions about it.) All three techniques can produce comparable accuracy if properly conducted.
21. Surveys are designed to represent a population of consumers, by taking a sample. Basically samples are of two types, probability and non-probability.
22. Probability samples are the better method to get an accurate representation of the population. The relationship between the sample and the population is known, there is no bias in selection and sampling error can be controlled.
23. Probability samples are of four main types:
(a) Simple random sample – each consumer has an equal chance of being selected (like a Lotto ball in a prize draw);
(b) Systematic random sample – from a list of consumers, every ninth person is selected (e.g. a sample from a mailing list) starting from a random start point;
(c) Stratified random sample – the population is segmented into mutually exclusive groups (strata) of importance to the marketer and a systematic or random sample drawn from each. Weighting may be used to represent the overall population;
(d) Cluster sample – the population is divided into a large number of groups (such as Census collection Districts). A sample of these groups is selected and multiple (clusters of) interviews are conducted.
24. Sampling error is calculable for probability samples. For example, a sample of 400 consumers will have a sampling error of no more than ±7% (with a 95% confidence interval).
25. Non-probability samples are also of four main types:
(a) Convenience sample – easily available consumers are selected (such as students in consumer behaviour classes);
(b) Judgment sample – the researcher attempts to construct a broadly representative sample from among those consumers available (e.g. experts in the field);
(c) Quota sample – the researcher arranges for a defined number of interviews to be conducted in each of several categories of respondents
(d) Snowball sample – in rare or hard to find populations, the researcher will ask each respondent to refer the researcher to other members of the population.
26. The key to effective surveys, are the questions (or questionnaire) which the researcher develops.
27. Questionnaires need to be brief, clear and interesting if respondents are to provide reliable information.
28. The wording and order of questions can severely affect the accuracy of an answer. Typically questionnaire design is a bigger source of error than sampling. For this reason, questionnaires are often pre-tested prior to use.
29. Interviewer bias is another problem that can affect accuracy.
30. Questions can be either disguised or non-disguised, depending on whether the researcher wishes to disclose the purpose of the research. For example, if consumers knew which political party was conducting an opinion poll, they may alter their responses.
31. Questions can be either closed or open-ended. In a closed question, a pre-defined set of responses is defined (such as an attitude scale). In an open-ended question consumers reply in their own words and may be encouraged to elaborate on their answers.
32. Scales are standardised sets of questions, typically used to measure attitudes. Three common types are Likert, semantic differential and rank-order scales.
33. Likert scales typically ask consumers to specify the extent to which they agree or disagree with a set of statements provided by the researcher. A key point is a balance of negative and positive categories (such as level of satisfaction or dissatisfaction).
34. Semantic differential scales ask consumers to rate a stimulus item (e.g. a petrol company) on a set of bipolar scales (e.g. clean vs. dirty).
35. Rank-order scales ask consumers to list stimulus items (e.g. television programs) in order of preference.
36. Consumer market research is normally conducted by professional companies.
Train fieldworkers – collectors
1. Research must be carried out exactly as designed by everyone
Collect Primary Data
1. Qualitative research requires skilled researchers to collect and analyse the data.
2. Quantitative data collection requires considerable attention to verify consumer participation and to check the accuracy of responses collected.
1. Almost all quantitative data is analysed by computer, using a mixture of simple (e.g. cross-tabulations) and complex data analysis methods.
2. Computers are sometimes used to analyse qualitative data
3. The analysis used reflects the nature of the data collected and the purpose for its collection.
4. Give interpretations – if/as required
1. The researcher aims to communicate the key findings to marketing decision makers in a succinct way that goes to the heart of their decision making. This after all was the purpose of the research.
2. Typically a report will include:
(a) an executive summary, highlighting the reasons the research was conducted, the main method used and the key outcomes. This summary is designed for business people who have neither the time nor the willingness to read a lengthy report;
(b) a summary report with the main findings and conclusions and recommendations;
(c) details of the research outcomes and the method used to collect information;
(d) attachments showing the research instruments used (such as discussion guides and questionnaires);
(e) computer tables showing the main cross-tabulations.
An outline of many of the topics and ideas to be found in the MAANZ Marketing Research program. Visit http://www.marketing.org.au – Full Diploma (PG and UG) courses and/or computer based short courses – MXPress.
Contact us on firstname.lastname@example.org