1. Define the target population
In defining the target population the important terms should be carefully and unambiguously defined. The target population may be divided into various segments which may have differing characteristics. For example, all television owners may be three or four income brackets.
2. Select the sampling scheme and sample size
There are several sampling schemes: simple random sampling, cluster sampling, sequential sampling, stratified sampling, systematic sampling, and non-probability sampling.
3. Develop the Questionnaire
The questionnaire is the principal instrument for eliciting information from the sample of respondents.
The effectiveness of the questionnaire as a device for eliciting the desired information depends on its length, the types of questions, and the wording of questions. Developing the questionnaire requires understanding of the product/ service and its usage, imagination, insights into human behavior, appreciation of subtle nuances, and familiarity with the tools of descriptive and inferential statistics to be used later for analysis. It also requires knowledge of psychological scaling techniques if the same are employed for obtaining information relating to attitudes, motivations, and psychological traits.
Industry and trade market surveys, in comparison to customer survey, generally involve more technical and specialized questions.
4. Recruit and Train the Field Investigators
Recruiting and training of field investigators must be planned well since it can be time consuming. Great care must be taken in recruiting the right kind of investigators and imparting the proper kind of training to them.
5. Obtain Information as per the Questionnaire from the sample of Respondents
Respondents may be interviewed personally, telephonically, or by mail for obtaining information. Personal interviews ensure a high rate of response. They are, however, expensive and likely to result in biased responses because of the presence of the interviewer.
6. Scrutinize the Information Gathered
Information gathered should be thoroughly scrutinized to eliminate data which is internally inconsistent and which is of dubious validity. For example, a respondent with a high income and large family may say that he lives in a n one room tenement. Such information, probably inaccurate, should be deleted. Sometimes data inconsistencies may be revealed only after some analysis.
7. Analyze and Interpret the Information
Information gathered in the survey needs to be analysed and interpreted with care and imagination. After tabulating it as per a plan of analysis, suitable statistical investigation may be conducted, wherever possible and necessary. For purposes of statistical analysis, a variety of methods are available. These may be divided into two broad categories: parametric method and non-parametric methods.