MARKETING RESEARCH DESIGN

MARKETING RESEARCH PROCEDURE      

Marketing research is undertaken in order to improve the understanding about a marketing situation or problem and consequently improve the quality of decision-making related to it. The usefulness of the marketing research output will depend upon the way the research has been designed and implemented at each stage of the process. There are five steps in every marketing research process:                                                

A) PROBLEM DEFINITION

B) RESEARCH DESIGN

C) FIELD WORK

D) DATA ANALYSIS

E) REPORT PRESENTATION AND IMPLEMENTATION 

 

B) RESEARCH DESIGN:

If you have stated your problem correctly and precisely, you should be able to spell out the precise objectives for research. Now you are in a position to prepare your research design. The research design spells out how you are going to achieve the stated research objectives. The data collection methods, the specific research, instrument and the sampling plan that you will use for collecting data and the corresponding cost are the elements that constitute the research design.

Data Collection Methods: A great deal of data is regularly collected and disseminated by international bodies, International Labour Organisation, World Bank, International Monetary Fund, Government and its many agencies including Planning Commission, Central Statistical Organisation, Reserve bank of India, Census Commission, private research organisations, and trade associations. This kind of data which has already been collected by another organisation and not by you is known as secondary data. This secondary data already exists in an accessible form; it only has to be located. You must first check whether any secondary data is available on the subject matter into which you are researching and make use of it, since it will save considerable time and money. But the data must be scrutinised properly since it was originally collected perhaps for another purpose. The data must also be checked for reliability, relevance and accuracy.

When secondary data is not available or it is not reliable, you would need to collect original data to suit your objectives. Original data collected specifically for a current research are known as primary data. Primary data can be collected from customers, retailers, distributors, manufacturers or other information sources. Primary data may be collected through any of the three methods: observation, survey and experimentation.

In the observation method, the researcher gathers information by observing. This method is generally used to observe buyer behaviour in a shop or to assess the impact of shelf placement and point of purchase promotional material. For instance you may like to observe the movement of shopping traffic through a department store, the number of shoppers who stopped before a particular display etc.

The obvious limitation of the observation method is that it allows observation of only overt behaviour. It provides no clues why a customer behaved in a particular manner, what product attributes appealed most to him, whether he would like to buy the product again etc. Such data can be generated by using the survey method. The survey method can also yield information about the socio-economic profile of your customers. The survey may either be conducted in a small group of customers through the focus group interview or may cover a large number of customers with the help of a questionnaire. In the focus group interview five to fifteen customers are invited for a discussion on a specific product or _.a specific aspect of the product. The customers' comments provide valuable insight into their thinking which can help the manager to fine tune his marketing strategy to suit different customer segments. Surveys conducted with the help of questionnaire often take off from the focus group interview which yields excellent clues for designing the questionnaire. The questionnaire-based surveys yield not only qualitative but also quantitative data which have statistical validity.

The third method of collecting data is through experimentation. This is basically a simulation of the real-life situation, but in a controlled environment in which you systematically introduce certain elements to study their impact. This method is used for finding the best sales-training technique, the best price level, the most effective advertisement campaign. However, its use requires an extremely skilled researcher to ensure useful results. Also, this method is expensive.

Research Instrument: In the observation method, the researcher may use a camera, tape recorder or tally sheet (a sheet in which the number of times an event occurs is recorded). Whatever the instrument used, the researcher must ensure that the instrument is appropriate to the occasion and is reliable.

In the survey method the most commonly used instrument is the questionnaire. This is a written and organised format containing all the questions relevant to soliciting the required information. The construction of a questionnaire requires great skill. To check that the questionnaire serves the necessary purpose, it should be tested on a limited scale and this is technically known as a pilot survey. The objective of a pilot survey is to weed out unnecessary questions, questions which are difficult to answer, and improve the phrasing of certain questions which are difficult to comprehend.

In constructing a questionnaire the important points to be considered are the type of questions to be asked, wording of questions and sequencing of questions. Each question should be checked to evaluate its necessity in terms of fulfilling the research objectives. Furthermore, the questions should be such that the respondent can answer them easily. Questions which require the respondent to answer questions about events which occurred a long time ago or about which he does not have direct knowledge should be avoided since you are not likely to get very accurate response. The questions should have direct relevance to the problem being researched. Too many irrelevant questions will only increase the length of the questionnaire (which would only put off the respondent) and also add to the burden of analysis without yielding any useful result.

The wording of the questions is a very important input in ensuring the correct response. Clearly worded, precise questions are not only easy to understand but they also facilitate the proper response. The wording of the question should be neutral and not attempt to influence or bias the response. This is especially relevant when information is being sought on non physical issues such as motivation, attitudes, and personal values of the respondent. If you want to know the name of the shop from where the respondent bought his last tube of toothpaste, any way that you phrase the question will elicit the same response. Consider the following three alternatives in this context:

a)     Where did you buy this toothpaste?

b)     Can you please tell me the name of the shop from where you bought this toothpaste?

c)     From which shop did you buy this toothpaste?  

On the other hand, suppose you are trying to find out the customer perceptions about the performance of foreign brands of televisions versus Indian brands. The manner in which you phrase the questions is extremely critical as it can influence the response. Consider the following three alternatives:

a)     Do you think there is any difference in the performance of Indian TV sets as compared to foreign sets? (neutral wording)

b)     Don't you think foreign TVs perform better than Indian ones? (interviewer bias)

c)     Most people feel that foreign TVs perform better than Indian ones. Would you agree wit: this statement? (introducing respondent bias)

 

When including questions about qualitative aspects it is better to ask open ended questions rather than close ended questions, and unstructured rather than structured questions.

Open-ended question

"How would you describe the taste of this toothpaste?"

Close-ended question

"Would you describe the taste of this toothpaste as tingling?" Yes/No

Unstructured

Word association: For assessing toothpaste taste you may ask the respondent to give his immediate reaction to the following phrases in context of your specific brand:

Fresh          Tingling       Foamy

Mild            Pleasant      Sharp

In the structured questions you may like to give the respondents a number of answer choices to choose from. This is known as multiple-choice questions.

"Which one of the following words or phrases, in your opinion, best describes the taste of this toothpaste?"

-Fresh

-Tingling

-Spicy

-Minty

-Cool

-Antiseptic

-Medicinal  

A technique which combines both the structured and unstructured type of questions is the question scaling. The respondent is asked to rank his perception of a particular brand, product attribute, company image or any such factor on a scale ranging from extremely good to extremely poor. A typical scale may look as depicted in Figure I  

The advantage with unstructured and open ended questions is that they give the respondent freedom to answer in his own words. And this often provides information and insight about the product which the researcher had not even thought about. The only problem with unstructured questions is that of interpreting the results. The same results may lead to different analysis by different researchers. Unstructured questions also make statistical summaries difficult.

Close-ended and structured questions are easy to summarise and there is no scope for misinterpretation. But the scope of the research gets limited. The respondents have to choose from already given alternative answers, even though none may exactly match the respondent's perception.

The sequencing of the questions in the questionnaire should be such that the opening questions create interest in the respondent and are easy to answer. You would not like your respondent to be put off by posing difficult questions right in the beginning. The questionnaire should gradually move from relatively simple to difficult questions. The questions should be arranged in a logical manner to facilitate the respondent's answers and not confuse him. Personal questions about income, education, profession should be asked in the end since many people may view them as a violation of their privacy.

Sampling Plan: After preparing your questionnaire or your equipment for observation, you have to identify the source of your information, the source is also called the `population' or `universe'. For conducting marketing research you would rarely gather information from the entire population, rather you would select a small group known as sample which has all the characteristics of the population, and conduct research among the sample group. The reasons for not using the population for research are:

a)     the number of units in the population may not be known,

b)     the population units may be too many in number and/or widely dispersed thus making research an extremely time consuming process,

c)     it may be too expensive to include each population item.  

When the number of population items is small and known, (say, the number of cinema halls, colleges, government hospitals in a city) you may use the population as your source of information. But in most cases, a representative group which has all the characteristics of the population and is known as sample is drawn from the population and this is used for conducting research.

Having decided to use a sample, your next step is to draw up the sampling plan. There are four aspects to the sampling plan:

-who is to be surveyed (sampling unit)

-how many are to be surveyed (sample size)

-how are they to be selected (sampling procedure)

-how are they to be reached (sampling media).  

The choice of sampling unit will depend on the product with which you are dealing and the kind of information you need. In case of a product such as lipstick if you need information on the reasons which motivate a customer to buy your brand, your sampling unit would obviously be the actual user, i.e., a woman. But would the population comprise all the women?

Obviously not, because all women do not use lipsticks. You then need to collect information about women who use lipsticks in terms of their socio-economic background, education, occupational profile (student, housewife, professional), age and marital status. The sample which you choose must be representative of the universe in terms of all these characteristics. If you want to find out the monthly sale of all brands of lipsticks in a particular market, your ^'sampling unit would be the distributors or retail outlets which deal in cosmetics. Suppose the product being researched into is toys for the under 7-years age category. Who would constitute your sampling unit: the child who actually plays with the toys or the parents who exert a strong influence in the final decision to purchase a particular toy? Here you would have to consider not only the kind of information that you need, but also who is most likely to have it and his ability to communicate, and choose your unit accordingly.

In deciding on the sampling size, you have to make a trade-off between the desired accuracy of the results and your budget. The larger the sample, the more accurate are the results likely to be, but the cost would also be correspondingly high. Another factor affecting the sample size is the kind of research which is being conducted. In exploratory research even a small sample may be sufficient. In focus-group interviews and motivation-research studies, very small sample sizes are sufficient because here the emphasis is on qualitative aspects rather than accuracy of numbers.

The choice of sampling procedure is between two kinds: probability sampling and non-probability sampling. In the former, each item of the universe has an equal chance of being selected as a sample unit. In non-probability sampling, the researcher selects the units to be included in the sample. Non-probability sampling is mostly used in exploratory research where a true representation of the universe is not important, But where true representation is important, probability or random sampling is used. Random sampling enables the researcher to make an accurate estimate of the population characteristic but it is more expensive than non-random sampling. The cost that you can bear and the degree of accuracy which you require have to he weighed to arrive at a decision.

The fourth element in the sampling plan is the sampling procedure. How should you reach your sample units: personally, by mail or by telephone. Personal interviewing is most suited when there are many questions to be asked and it is important to ensure that the questions are understood properly. Thus, wherever the questions are little complex, personal interviewing should be used. This is also the best method to ensure that correct answers are given which can be corroborated by the interviewer through observation. But this technique requires a skilled interviewer and a great deal of administration and supervision. Also, it is the most expensive of the three methods.

The mail questionnaire is extremely appropriate when your sampling units are distributed over a wide geographical area and the cost of reaching them personally is very high. However, the return rate of mail questionnaires is usually very low, ranging between three to seven per cent. On an average, you would have to mail 1000 questionnaires to get back thirty filled up questionnaires. Another drawback is that you have no way of checking the authenticity and accuracy of the response. The respondent may fill totally wrong information and you may never be able to detect it.

The telephone interviews combine advantages of both personal and mail interviews. It allows you to clarify questions which may not be clearly understood by the respondent and to reach a widely scattered sample at a relatively low cost. But the obvious disadvantage is that your sample is restricted to the people who have telephones. Also, you, cannot conduct very long interviews over the telephone.

Cost: No information can be collected without incurring cost. Before undertaking a research project its cost should be calculated and assessed against the benefits it would yield in improving the quality of decision-making. If the benefits outweigh the cost, it is certainly worthwhile initiating the research. There are four kinds of costs involved in marketing research.

Cost of data collection: The actual cost incurred for collecting the data, which may comprise the research organisation's fee, staff time, printing and postage of questionnaire, computer time, etc.

Cost of time delays: The more time it takes to provide the research results, the longer the dependent decision (s) is delayed. In the meanwhile, the opportunity may be lost or it may become less attractive.

Risk of adverse environment change: While the decision is pending unfavourable conditions may set in (entry of competition) and consequently the returns may be lower.

Cost of error: Sometimes, by chance or because of some bias or wrong choice of sampling units, there could be an error in the results leading to expensive consequences for the company concerned.