Is conducting a survey as easy as just asking questions? Well, yes and no. A survey is a series of questions you ask a particular group of people. Common marketing research goals are to somehow better the business or product. There are, however some pitfalls that stop research from being productive. Below are some general tips on how to create a survey.
1. What are the goals of the project? Just asking questions for the sake of ‘doing research’ is expensive and wastes time for all involved. Any project needs to have a specific question it is trying to answer through the research. Examples: Customer opinions on a current (or new) product. Employees’ job satisfaction levels. Testing new advertising promotions before launching the campaign
2. Who is the target sample? Who you ask the questions to is integral to research if you want reliable results. You may want to get answers from current customers, non-customers, your own employees, or a segment of your geographical area. It is important to get a representative sample so that results do not show a bias and can be generalized to the greater population.
- Personal Interviews offer great qualitative incites, however they are expensive, require well-trained interviewers and are time consuming to conduct and then to code later.
- Telephone Surveys are usually reliable ways to contact respondents, however current trends show that landline phones are on the way out and mobile phone numbers are not readily available; they are also time consuming like interviews.
- Mail Surveys are less expensive and offer more visual elements to the respondent, however they can take longer to receive results and multiple requests to any respondent increases costs, especially if addresses are not current.
- Computer Direct Interviews & Web Page Surveys are surveys conducted on a particular website. They offer anonymity to respondents and the computer can cater the survey based on previous responses, however you will only get respondents with access to computers and the Internet who are already on the site.
- Email Surveys are FAST & GREEN, however limited to those with Internet access and email addresses (this is becoming less of a factor, but should be considered for certain sample groups). Being categorized as SPAM can also be a problem.
- Introductions should appear in a cover letter or front page of the survey and include a survey title, identify who you are (or disguised as), the intentions of the survey, and how respondents are to complete it (such as offering a link, address, or phone number). Appeal to a respondent’s altruistic senses: Your answers will help us make a better product, company, buying experience, etc. Their answers are important to you and here is where to tell them.
- Question Order needs to keep your respondent engaged. Open with questions that are interesting and focused on your topic to create a rapport. Group questions by topic; if there are several in one group split them up with a smaller set of questions so the respondent does not get frustrated and quit. If you have recall questions make sure the information is not presented in an earlier question. Demographic and sensitive questions should be put towards the back of the survey.
- Keep it short; overwhelming respondents with too many question or complicated instructions will decrease response rates. If a question does not measure something pertinent to the project drop it! Be consistent in how you set up questions; common types are multiple choice, Lickert (rating) scales, and open-ended questions.
- Simple Symbolism can help respondents maneuver through the survey. Be sure to use common symbols anyone in the sample would understand. Arrows and stop signs are universally understood.
- Pretest, PREtest, PRETEST your survey with a small group to ensure that the instructions are clear, your questions make sense, and answers are exhaustive. Test links to the survey, format and ease of use for the medium.
