T.L. HEADLEY, MBA, MA, BA, AAMS
Principal and President
In today’s business world it is imperative that a company know its customer base. Who is your customer? What are their likes and dislikes, wants and needs? What are you doing right? What are you doing wrong?
There are two ways to answer these questions. A business owner can gaze into crystal ball and rely on psychic abilities to find the answers, or they can do the hard work of research. If they choose research they are still faced with the question of whether or not the researcher is competent and whether the research is good or bad.
A company would probably do better relying on the crystal ball or tea leaves rather than bad research. Only through good, quality research can a company lessen the risks of doing business.
For many people, polling is like alchemy. It is something they don’t quite understand and are not quite sure they trust. If they do know about it, they assume it is very expensive and their business simply can’t afford it. The end result is more gazing into the crystal ball and failing businesses.
The Genesis Group can take the mystery out of public opinion polling. We can provide the service as affordable prices and help take some of the guesswork out of doing business.
CHARLESTON – For a business owner interested in finding out what is customer base is thinking, the magic number is 384.
384?
That is the number of responses needed to accurately predict the behavior of populations of 100,000 or more people. It is the cornerstone of survey analysis – a key component of good market research.
Many business owners mistakenly think survey analysis is either too hard to conduct properly or simply too expensive.
The truth is, neither view is correct. However it isn’t as simple as picking up the phone and dialing 400 people.
The first question for a business owner is whether they really need a scientific survey. There are other ways to gather market research, such as a focus group or content analysis, but these methods do not necessarily provide an accurate view of the general population.
If you determine that survey analysis is needed, there are a few other decisions to make. Each decision affects the cost and accuracy of the survey.
The first decision is how accurate the survey results must be. This is known as the “margin of error.”
For example, if you want to determine the potential number of people who will buy a new newspaper, you would probably want to get as close an estimate as possible. Is plus or minus 10 percent good enough? Or do you need a 5 percent margin of error?
Another calculation is what is known as the “confidence interval.” This is essentially the likelihood that the given percentage is correct. Most surveys aim for a confidence interval of 95 percent. For the purpose of this article, we will assume a business owner wants a 95 percent confidence interval. The numbers we will cite use this as a baseline.
Once you determine the confidence interval and the needed margin of error, you can use this to determine the size of the sample needed to meet these needs.
For populations of 100,000 or more people, the minimum number of respondents needed is 384 people. It is important to note that 384 is the number needed of people who accurately respond to the survey. This does not include those who fail to respond or who only partially respond.
A rule of thumb is to double the number of respondents needed to estimate the number of people who will need to be contacted to achieve the number of responses. In other words, it may be necessary to contact 768 people to get 384 responses.
But how do you know who to call or contact? Do you just dial the first 768 people in the phone book? Or do you just send out 768 letters to people along a certain street?
This is where many people give up.
Once you determine how many people are needed, you must determine which method of survey you will use – mail, phone, direct intercept or contact? The most cost effective method is mail – however it is not the most dependable.
A phone survey provides a good view of most populations – however there are more and more people who don’t have “land-line” phones, choosing instead to use cell phones. These numbers are not readily available to those performing surveys and a large portion of the population (particularly young and mobile people) could be missed.
Direct intercept or contact provides perhaps the best, most accurate sampling, but is the most expensive method.
That said, the differences of reliability are relatively small and most surveys rely on either mail or phone as the method.
If you choose the phone survey method, you will need to determine whether you have the capability of conducting the survey in-house or will need to hire a research firm. If you choose to conduct the survey in-house, whether phone or mail, you will need to know how to generate random sampling.
What is random sampling?
Basically this means that you determine your contacts (phone or mail) in such a way that each individual in the entire population has an equal chance to be chosen. It sounds simple, and it can be, but many people make it harder than it has to be or else use “opportunity samples” – temporary groups of people, such as those you might find at a movie theatre on Friday night.
A true random sample can be achieved by using a random number generator. These are readily available on line and in numerous statistical text books. It is essentially a table of numbers, generated at random.
If you are using a random number table (often 10 by 10 square for 100 individual numbers), you determine a pattern, such as two over and one down. Then you close you eyes and choose a first number at random. The next number is chosen by moving according to the pattern. This is a simplification, but is essentially correct. These numbers are then used in groups of four to generate the four ending numbers of a phone number.
Keep in mind that this is a rough explanation. Those interested in getting a good sample should use a random number generator.
From this point, the business owner can easily make the appropriate number of calls needed to generate a valid response.
Next: Creating a Survey: Asking the right questions is vital.
The Mind's Eye 