How to conduct a survey
If you are an entrepreneur and would like to conduct a market analysis, then a survey is one of the most useful tools that can assist you.
So you have decided to start your own business, or bring a new product or service to the market, where do you begin?
Well, for starters, we would always suggest doing a market study. A market study incorporates various research methods and techniques, and one of the most useful research tools is the survey.
Why is the survey useful?
According to the Center for Sociological Research Methods, a survey “is one of the most important areas of measurement in applied social research”. Survey research covers a broad area of measurement procedures but the nature of survey research is to ask questions and receive answers from respondents - which can include anything from a web-based survey to face-to-face interviews.
There are a number of advantages - and also disadvantages - of the survey, and its important that you are aware of both when conducting market research.
Firstly, some survey advantages:
- The survey is easy to apply
- Surveys are a quick way to gather data from a large population
- The data obtained is reliable
- Conducting surveys are usually low cost
- Little or no observer subjectivity (or bias)
- Results are precise, and usually statistical (which makes it easier to identify patterns)
- Interpreting and analysing surveys are simple, and easy to understand
Some disadvantages to surveys:
- Inflexible design (questions are set and pre-described)
- Rigidity (if respondents wish to elaborate or tell you something more about the product or service but are unable to)
- Sample respondents might not be sufficiently representative
- The survey questions might not be correctly understood or fully comprehended by the respondents
- Depending on the type of survey, it can end up being expensive
Just like any other research methods, you will find both good and bad aspects about the tools. Therefore, depending on your intention and purpose of research (for example if you would like to know the behaviour of a large number of consumer preference), then a survey can be the most useful research tool.
Define the objective of your survey
Naturally, if the survey is part of a market study or analysis, then it is likely that the objectives have already been defined. Nevertheless, it is still worth remembering that your goals should comply with the SMART acronym. By following the SMART acronym, you reinforce that your objective is Specific, Measurable, Achievable, Realistic, and defined in Time.
The objective of your research is important when designing the survey questions. It is therefore important that you know what you are looking for, otherwise you could be asking incorrect or irrelevant questions.
Define your survey questionnaire
Once you have defined the objectives of your survey, you need to construct the appropriate questions that will best help you find the answers you are looking for.
For example, each survey question should:
- Correspond to a single answer (the question should not be ambiguous)
- Be relevant and timely - referring only to the objectives of the study
- Formulated with precision - concise and understandable questions that evaluate the aspect of your study
Furthermore, the type of questions can be classified as either open or closed-ended:
Open-ended questions: allow for diversion from the question, invites the respondent to answer in their own words, and usually lead to a discussion. This can be more difficult when coding, since the response cannot be measured in numerical terms. For example - “How would you explain your highschool experience?”.
Closed-ended questions: questions which can usually be answered “yes” or “no”. These questions are more defined, and usually the respondent can choose from a series of answers. These results are easier to code, and tend to be quite precise. For example - “How would you rate our service?” A) Good, B) Ok, or C) Bad.
Define the sample of your survey
Now we are getting to the more complex part of surveys. The sample of your survey is very important for the validity and reliability of your data - but don't worry, we will guide you through the process so it doesn’t seem so daunting!
When we talk about a sample in statistical terms, we are referring to the group of respondents. The sample group is essential to your research, as you can not count on the entire population as respondents.
A well-defined sample is the basis of any survey, if you select a bad criteria as your sample, then no matter how well your questions are designed, the results will be almost useless. Why is this? Well, put simply, your respondents should be relative and representative of the market segment you are targeting.
Of course, the size of your sample matters too - and usually the larger the sample the more accurate the results - but the sample size doesn't even matter if your sample segment is not representative of your target market segment.
As for the sample size, naturally the larger it is, the smaller the margin of error. The margin of error is a statistical magnitude that indicates the inaccuracy generalisations that exists by asking a sample population rather than the entire population. In other words, if the results of a survey show that 100 people rated the service as ‘Bad’, and we have defined the margin of error as 5%, then we are indicating that between 95 and 105 people rated the service as ‘Bad’.
To determine your necessary sample size, you can use this formula:
n = (Z²pq) / d²
Don’t worry - it is not as scary as it looks, and we will go through what each variable stands for:
- Z = the confidence level. This indicates the reliability of your results, and the recommended confidence level is 95% (or 0.95), therefore Z squared = 0.95² = 1.96
- p = the estimated proportion of individuals that have the characteristics of your study in the population. Since this number is usually unknown, we use p = q = 0.5.
- q = the proportion of individuals who do not have the characteristics. Hence, q = 1-p, therefore q = 1-0.5 = 0,5.
- d = the precision of your study (or margin of error in terms of proportion). In general, we use a 5% margin of error, therefore d = 0.05.
As an example:
You would like to know with 95% confidence level and 5% margin of error how many people you would need to survey to test whether a service is rated Good, Ok, or Bad, then you can use this formula:
n = (1.96² x 0.5 x 0.5) / 0.05² = 384,16
Therefore, you would need to survey 385 people.
Now let's assume that you do know the general population, that is, the size of your target audience. Then the formula gets a little bit more complicated, and will look like this:
n = (NZ²pq) / d² (N-1) + Z²pq
Where N is the general population. Using the same example, if you know the amount of people who have tried that service before is 15000 people, then you can use this formula:
n = (15000 x 1.96² x 0.5 x 0.5) / (0.05² x (15000-1) + (1.96² x 0.5 x 0.5))= 374.59
Therefore you would need to survey 375 people.
This formula is enough to provide you with an indicated sample size for your survey in market research. However, if you would like to dig deeper into the sample size you can consider changing either your confidence level (z), or margin of error (d), or both.
Preparing the fieldwork for your study
Fieldwork addresses applying the survey to the selected sample. It is advised that you clearly define the time in which you will carry out your research.
Fieldwork varies depending on the type of survey. For example, an online survey can be effective to reach a larger segment and means you are not physically present with your respondents, a telephone survey takes up more time (especially if the selected sample is large), and a face to face survey is time consuming as it involves physically engaging with each of your respondents. Since all survey types require some level of resources, the fastest and usually cheapest option is the online survey, since it is also easier to analyse the results from your computer data.
After conducting your survey
Once you have done the fieldwork and received your results, now you need to code them. Coding data simply means converting your ordinal and nominal scale into categories with numerical values for you to easily analyse your results and notice any trends or patterns. For example, the variable gender can be coded in the following way: 1 = woman; 2 = man.
For your coding, you can use Excel to analyse your survey questionnaires.
Interpreting your survey data, and preparing conclusions
The final step in your survey market research is to analyse your data and interpret it in the area of the objectives you set.
Depending on the way you have prepared the questionnaire, each question should correspond to a conclusion, and the whole questionnaire should lead you to some general conclusions. For example, if the questionnaire shows that 70% of respondents are not satisfied with the service provided, you can conclude that the general population views the service as ‘Bad’.
Your survey results can help draw your attention to aspects which you had not foreseen in your business idea, and even open the discovery for a potential niche market to be addressed. The advantage of this is that it helps you as an entrepreneur when making business decisions, even if it means reviewing the core of your business.
If at this point you still have doubts about surveys and market analysis, don't worry because it is completely normal. Your first survey can be daunting, but with practice you will soon get the hang of it, and see that surveys are a great way to conduct market research for a specific field of interest.
Debitoor is an online invoicing software perfect for small businesses, entrepreneurs, freelancers, and sole traders. It is simple and efficient to use, and helps you maintain an overview of your business at any point in time. With Debitoor you can manage your billing, expenses, customers and products, and track your business growth.