Just as senior managers need not be accountants to understand the principles and lessons of the profit-and-loss statement, balance sheet, and cash-flow budget, neither do they need to be statisticians to understand the principles and lessons of fact-based market research.
The principles for any valid, replicable, projectable market research study are the same.
1. Survey a population that represents the total universe as defined by the company. Whether your market is women who might be interested in a new health club, food-store decision makers responsible for shopping carts, or homemakers who buy ground coffee, ensure that your sample represents the population. For example, an internet survey that is not weighted to control for bias is not representative of any population other than people willing to participate in internet surveys. It understates the proportion of the market who are not connected (i.e., minorities) and people who have neither the time nor the interest in cooperating with survey researchers (i.e., busy executives, hassled moms, professionals in general). In such cases, you must take care to recruit these under represented groups into the study and to weight different types of buyers into their correct proportions in the population.
2. Sample enough people within this universe so that their responses are reasonably stable. Interview 500, for example, and any percentage that you get is stable plus or minus about 5 percent. Interview only 50 people, and the wobble in your data increases to a point at which it is relatively unusable. This is particularly important when you are looking to break down your total sample into subgroups such as users versus nonusers, men versus women, high income versus low income, and the like.
3. Test the questionnaire to ensure that respondents understand what is being asked and are not being prompted to answer one way or another. This is called pretesting and is something that is becoming more and more rare in marketing research studies.
4. Conduct experiments within your survey wherever it makes sense. Expose half the sample to one new product concept, the other half to another. Show respondents three different price levels. Expose people to four different positioning.
5. Analyze the responses in a way that makes logical sense. Don't confuse correlation with causation; just because one thing is related to another does not mean that the first caused the second. |