What Are the Principles of Market Research?
Organizations need market information to grow and better serve their consumers. Marketing research is the process by which organizations procure the information that enables better decision-making as they advance toward these goals.
Other People Are Reading
-
Objectives
-
All marketing research begins with an objective. An objective can be as simple as "Which formula for our soft-drink brand tastes best to our consumers?" or as complex as "What should be our five-year strategy for expanding our computer software brand into other categories?"
If a research objective is articulated effectively, other details of the project fall into place, including whom to talk to, what specific technique to use, what questions to ask and how the information will be used.
Sampling Plan
-
Since it is impractical to talk with everyone whose opinion a firm might want, the organization instead speaks to a sample of those individuals, with the expectation that this sample will estimate the opinions of the entire population.
There are many sample-creation methods, and which is used depends on the objectives of a project. In general, larger samples are better: Although all samples entail error, with larger samples, error is reduced. You can then be more confident your sample measurements approximate those of your population.
-
Instrument
-
Research instruments--that is, how the information is to be collected--will vary depending on the objective(s). In some cases, a qualitative approach might be sufficient, so a set of four focus groups might be appropriate. In other cases, a quantitative approach is needed, so you might need a survey to be answered by hundreds of people.
A simple rule of thumb is to determine whether the answer you seek is something that can be quantified. If so, you will likely field a quantitative-based survey with a large sample. You can also examine where you are in an overall project timeline. In early stages, frequently only guidance or preliminary information is needed, and therefore the research can be qualitative or smaller in scale. At later stages, when significant resource allocation decisions need to be made, larger quantitative research would likely be appropriate.
Action Standards
-
Action standards (a.k.a. decision criteria) are the articulation of how the research will be used--specifically, what decision is being made with the research, and what criteria will be used to make that decision.
In many cases, it is helpful to have benchmarks. Sometimes, companies have their own internal benchmarks; at other times, research suppliers have historical benchmarks based on years of testing.
Let's say Company A is interested in replacing its current product with one of two potentially improved products, and plans to execute a test with users of the current product, where the current product is compared to the two potential improvements. Here, the current product is the benchmark, and the action standard could read that if one of the two new potential products significantly outscores the current product on a key measure, like preference, that new product would replace the current product.
Error and Risk
-
No matter how well a marketing research test is designed and executed, there is always some risk the information you receive is faulty or even false, and therefore could lead you to make a wrong decision.
There are two basic types of risk. Looking at the Company A example above, there is a risk that if one of the new products outperforms the current product, the new product is actually inferior to the current product and should not be introduced. This risk is also referred to a "Type I error."
Conversely, if neither of the two new products outscores the current product, there is also a risk that one or both is actually superior to the current one, and should be introduced. This risk is also referred to as "Type II error."
Although it is impossible to eliminate, risk can be reduced, or at least managed. The easiest way to reduce both kinds of risk is to increase the sample size of a test. If one type is error is deemed more grievous than the other, an action standard can be manipulated so that type of error is minimized--but the trade-off will be that the other type of error will increase. In practice, most research tests seek to balance risk between the two types of errors.
Analysis
-
The most basic analysis consists of simple descriptive statistics--that is, what percentage of people answered this way or that way to our questions. This output is frequently accompanied by significance testing. For example, in an advertising test with two new ads and a current ad as a benchmark, both new ads could be compared to the current on a measure such as viewer interest, with an action standard that an alternative ad will only be produced and broadcast if it outperforms the current ad on this measure at a 90 percent confidence level.
Another common analytic technique is cross-tabulating descriptive statistics. With this technique, you can examine subsets of interest-- such as men, households with children, or current brand users--and make comparison between them and to your total sample.
There are many more elaborate analytical techniques, such as regression, factor analysis and discriminant analysis. These tools are used less frequently, but can come into play depending on the research objectives.
Overall Design Considerations
-
In designing marketing research, researchers must balance several factors and recommend the most appropriate research for the question at hand, the risk involved and, of course, the budget.
An important skill for a researcher is scaling a project appropriately. For example, if a brand wants to introduce a new flavor or size of an existing product, that is a far less risky proposition than if it wants to introduce a totally new product in a different category. The latter situation would likely call for a larger, more comprehensive test.
Factors that drive research cost include the sample size, the difficulty of recruiting the right people, the length of the survey, and the level of analysis required after the fact.
-
