Primary sources of data for market research


Primary data is of two types

1. Census
2. Sample

Census refers to the collecting of data from the entire population. The most common form of census is the Indian population census or compilation of voters’ list in an area. Census takes a long time and hence is not suitable for most marketing researchers. However, one may observe its usage only where customers are limited and have diverse needs. An aircraft manufacturer like Boeing or Airbus Industrie can use the data from census surveys as the number of airlines in the world is limited. Census data may not be useful for a tire manufacturer or a soft-drink or a textile firm who cater to large markets. To overcome this problem, most marketing research and consequent decisions are based on sample responses.

Sample refers to a pie taken from a population. If this pie is drawn carefully, the conclusion can be extrapolated to the entire population and generalization can be made. The key then is a representative sample. Some of the decisions to be taken are the size of the sample (number of people to be contacted), how their responses will be tabulated, analyzed or interpreted (sample stratification), and how the sample will be drawn (sampling procedure)

The decision on sample size is based on the time available to the researcher and decision-maker. It is also dependant on the nature of the problem or the research design. An exploratory research involving focus groups could be done on a sample of 100-150 customers. So long as these customers match the demographics and psychographics of the target market, conclusions from this sample can be generalized but a gallop poll forecasting the success of different parties in next general elections is done on a much larger sample size. Location of the sample also needs to be considered. This is often based again on the nature of the problem, product positioning and firm location. For example, to study the problems of suburban train commuters in Mumbai a sample drawn from the Western and Central suburbs can help. But to study the problems and expectations of train commuters in India, the sample will have to be drawn nationally with representatives from rural and urban commuters, male and female commuters, and travelers in different income age, education and occupation groups, etc. This is because problems and expectations are likely to differ across the length and breadth of India.

Sample stratification refers to the basis used for segregating responses. Often demographic, geographical location and psychographic variables are used to stratify the sample responses. Each cluster of sample respondents is called strata. This helps the researcher test differences in behavioral patterns among different strata.

The key to a reliable conclusion is the sampling procedure or the manner in which a sample is drawn.

There are different ways as described below:


It sampling refers to collecting data from a number of respondents drawn randomly. For example, the researcher may decide to collect information from every third person he meets at Mumbai CST Station. This method helps in eliminating researcher bias and any other respondent bias which may lead to an error in conclusions. The problem with this is that it is not necessary that every third person is the target customer for the firm’s product and hence the responses may be incoherent and not useful. Hence this is not a very commonly used procedure.
To overcome this problem, the researcher may decide to use the purposive random sampling technique.

Purposive and Judgmental

As opposed to random sample, purposive and judgmental sampling procedure involves the researcher assessing a respondent and the value of information from such a respondent. Based on this assessment the researcher decides to include a respondent in the sample. This method suffers from researcher bias and hence conclusions can never be generalized.

Here, the researcher decides on who meets the target market’s definition and then collect data. For example, the researcher may decide to collect information from every third women aged between 21 years to 35 years if he or she (i.e. the researcher) is doing research on cosmetics. At other times the researchers may ask the question to every third person on whether he/she uses the product and based on the reply may either continue investigation or close it.

Many a time, the researchers may have to use their own judgment in selecting a respondent. Even though he/she might have been asked to knock at the second flat of the second floor of every alternate building on the left of the road, he may find that there are not enough buildings on the left but there are plenty on the right of the road. In such a case he uses his judgment and considers the purpose of the research and target market definition. Such sample then is called the purposive, judgmental randomly drawn sample.

There can also be other methods designed to suit a firm’s product exclusively with certain guidelines given to the surveyor or collector of data. After collection the information is compiled statistically and conclusions drawn after making suitable calculations.

Whatever method is adopted at the end of the day it is the product quality, pricing, utility and after sales service that matters to increase the market share particularly if it is a new entrant to the market. However optimistic the research may be the marketer must ensure the aforesaid aspects to make it consumer oriented and for customer patronage.