Over the past 15 years, Opinionmeter International has distributed thousands of satisfaction surveys through its wide-range of interactive survey devices. Here are a few of the key points we've found it important to consider whenever preparing your survey instrument.
Preparation - Take the necessary time to structure the survey correctly. Before beginning to draft your survey questionnaire, you should be able to answer this question: What is the objective of the survey and how will the survey data be used?
Short and Sweet - Often companies fall into the trap of wanting to ask all their burning questions in one long, complex survey which then results in a high non-response rate. Much better to break up the survey questions into smaller more focused satisfaction surveys, each with a clear objective. Lengthy, complex surveys create friction between you the researcher and your respondent. An easy survey is one that doesn't take much time and makes the benefits of the responding obvious. The more steps a survey entails, the harder it is to complete. Don't make your respondents work too hard to give you their feedback! Instead why not undertake a number of focused satisfaction surveys throughout the year with a maximum of 15 questions, with 6-10 questions being the goal.
Question wording - Be careful not to assume the survey participant has a level of knowledge that may not exist. Make sure you are using vocabulary that your customers understand. Don't get caught up using company or industry specific vernacular or acronyms. Be clear so your entire respondent base (the "population" you're surveying) can understand your survey questions. In fact, it often makes sense to "test-drive" a survey with some outside your organization (family or friends) before you launch it to the entire survey participant population.
Whenever possible avoid making the survey respondent uncomfortable by using threatening or overly personal questions. If you ask questions that require too much thought or seem a little too personal, your survey respondents will be less likely to participate.
One-dimensional questions - A common problem with survey questions is cramming multiple dimensions into a single question. For example, if a question asks, "Please rate the wait time and courtesy of our staff today?" The respondent would be unsure how to respond. What if the wait time was far too long, but the staff was extremely courteous? Questions with multiple dimensions prevent the participant from providing a specific response due to lack of focus.
Eliminate Ambiguous Responses - Make sure you provide clear response options for each question in the survey. Check that you have covered every eventuality with your response range. For example: "How did you travel here today? A) Car, B) Public Transport, c) Motorcycle d) Walked…. Someone who has travelled by bicycle would get to this question and not complete the survey. If you're not confident you can provide all possible options, you can always use "Other" as a final response option.
Single Response vs. Multiple Response questions - Why make someone choose only one response if they have multiple reason for liking your products or services? Rather than limiting customers to one response (multiple choice), listen to all that they have to say. In some cases, a ranking system may be more appropriate (more on this later).
Never imply an answer - Be careful to avoid influencing the survey respondent with leading questions or heavy use of adjectives. For example, "which do you prefer most about the new decor of our restaurant?" Who's to say they prefer any of the new attributes of the new décor? The issue with adjectives is they can mean different things to different people and should be used sparingly. If you want honest, meaningful answers, you should be as specific as possible. Branching logic and skip patterns - Questions that are dependent on the previous answer are referred to as "Branching questions" (sometimes also referred to as "skip patterns" within a survey). Branching is a great way to be specific with various sub-sets of your survey population; however it's best to keep branching to a minimum with short, simple surveys.
Ranking questions - When you want to determine the relative importance of various options, for example product features, ranking questions can be very useful. Keep in mind however; ranking questions do take a fair amount of thinking on the part of the respondent. For this reason, we normally recommend limiting your ranking options to 5 or less.
Demographic questions –Use demographic questions to segment your respondents. 90% of your population maybe happy with quality of food, but 100% of males under 25 may not be. If you are using technology for your satisfaction survey you will also be able to segment by date and time and duration of the survey. Demographic questions are best placed at the end of the survey.
Test, test, test – In the end, testing is the only way to know what's going to work with any market research initiative. The best rule of thumb is never to assume anything and test everything. It's always best to test your survey on a handful of people outside of your company to ensure you haven't assumed knowledge or unnecessarily used confusing industry jargon.
