What is market research?

In simple terms, marketing research is the systematic process of collecting, analyzing, and presenting objective, useful marketing information to assist management in problem solving and decision making. We believe it can be defined as the process of understanding: it helps you understand what important groups of stakeholders think of you, what they feel about you, and what you mean to them - and it also highlights what you can do to modify their perceptions in your favour. Market research should help management to identify, service, and satisfy the needs and desires of its customers.

What are the benefits of market research?

Research helps you uncover and understand factors that drive product or service choice, and result in customer or stakeholder satisfaction. Even trivial attributes can have a marked impact on the way people view a particular offering. Once you know how your stakeholders judge what you provide, you can modify the offering or your presentation of the offering so that the users find greater value in it - adding up to improved customer satisfaction. Research has helped countless organizations fine tune their approach to their markets to provide dramatic acceleration in profitability.

What if I don't do any research?

People tend to be either their own harshest critic or their own greatest admirer - either approach lacks the objectivity necessary to really understand how the market views your offering. Opinions are all well and good, but there is no substitute for knowing what is really going on in making effective, confident decisions about your organization.

When could it help my company?

The main reason many companies first turn to market research is when they strike a real problem. Why has that star product stopped growing? Why have product returns increased? The distribution base shrunk? Profits fallen? And so on. But there are many more scenarios in which insightful and dedicated research can offer dramatic return. Here are a few examples:

• When you lack all the information you need to make a decision.
• When you are weighing alternatives.
• When there is conflict in the organization about which direction to take.
• When you detect symptoms of problems from your customer base.
• When you embark on something different, such as new products, new packaging or new advertising campaigns.
• When you wish to raise financing
• When you are contemplating a merger or acquisition

Glossary of research terms
You'll find this a useful resource for definitive research terminology explained Quirks.com: Research Resources [Glossary]

An overview of sample sizes

We often hear questions relating to the statistical validity of sample sizes, and in itself this can be a complex subject. Put simply, this means that the statistical validity of a sample is a straightforward function of sample size: in effect, the larger the sample, the more accurate the results will be.

"Accuracy" refers to the "range of error" for the sample at a given confidence level. The most common confidence level used is the 95% level which means that if sampling were conducted the same way 100 times, results would fall within defined limits on 95 occasions.

Every sample size has a specific range of accuracy at the 95% confidence level. For example, a sample of 100 corresponds to plus or minus 9.8 percentage points, 200 to ±6.9 points, and so on. Interestingly, doubling the sample size does not double the accuracy. Also samples smaller than 30 are not statistically meaningful - in other words, the range of error is sufficiently large that you cannot interpret the results with confidence that they are reasonably representative of the population being studied .

Which sample size is right for your project? Many factors come into play in this decision, including the scope of the project, the requirement to look at sub-samples, and the significance of the design-making that will result from the research.

An overview of margin of error

This table shows the maximum range of error for frequently used samples sizes at the 95% confidence level: