10 Tips for writing survey questions for market research

Be Careful What You Ask For

I’ve heard it too many times. “We want to do a survey, and here are the survey questions we want to ask.”

My response is always the same: “Time out. Tell me what you want to learn from the answers. Then, we’ll figure out the questions.”

The process of designing an effective survey is a combination of math, linguistics, and psychology. Ignoring any one of those things can lead to missed opportunities and poor decisions. Some could be minor annoyances, but some may lead to missteps in strategy that will be costly if not realized until it’s too late.

Here are 10 of the most common pitfalls when it comes to writing market research survey questions and tips for avoiding them.

Common mistakes to avoid when writing market research survey questions

1. Failure to think like your audience: Make sure your wording is clear and understandable. Industry jargon may be permissible in a B2B survey, but could be problematic in a consumer survey. Also, remember the importance of the order of the questions. Surveys that don’t flow logically can confuse and frustrate respondents. If they feel like there was a lack of effort in designing it, they may put less effort into taking it.
2. Too few answer choices: Sometimes we think we know what the answers will be, but isn’t the point of research to learn and confirm? To avoid leaving out important response options, it helps to do some qualitative research (1:1 interviews or focus groups) before drafting the survey. It doesn’t need to be elaborate or expensive, but it can uncover issues you may not have considered.
3. Too many answer choices: On the flip side, excessively long lists of answer choices won’t get thoughtful consideration. Strike a reasonable balance and, where appropriate, include an “Other – please specify” option for respondents to write-in something that’s not covered.
4. Too many open-ended questions: Asking respondents to elaborate in their own words can yield valuable insights, but asking too many causes respondent fatigue, superficial answers, and drop-out.
5. Insufficient incentive: It is important tochoose the type and level of incentive that will secure the sought-after number of responses. There are several factors that should impact that decision. Who is the audience? How hard are they to reach? How valuable is their time? How long realistically will it take them to complete the survey? How interested and engaged will they be with the content?
6. Order bias on multiple choice questions: Even with a well-designed survey, some respondents will try to speed things up by picking an answer near the top of the list. Randomly rotating the order of the answer options reduces the negative impact of this on your data.
7. Using the wrong question type: Choosing the right format for each question results in better data and better insights. For example, if you ask them to use a 10-point scale to rate the importance of price, quality, and service, you may fall victim to “everything is important” syndrome—a bunch of high scores with little differentiation and no real insights. In some cases, other approaches like ranking, weighting, or MaxDiff (best/worst scaling) will yield more meaningful results.
8. Single- vs. multiple-response format: In some cases, asking respondents to choose one answer from a list of options is best. But sometimes there may be multiple considerations that can be masked by forcing a single response. One way to combat this is to have them choose their top choices, and in some cases, rank them.
9. A lack of attention to detail: Small mistakes in survey design can lead to frustration for respondents and potentially incorrect data for you. For example, watch out for numeric ranges that have gaps or overlaps, or are so broad that respondents can’t be broken out into meaningful segments.
10. Forcing a response on potentially sensitive questions: Requiring a response to a question helps avoid spotty, inconsistent data. However, in cases where the information might be considered sensitive or proprietary, requiring a response can lead respondents to falsify their answer or drop out of the survey.

So, how do you avoid these pitfalls?

Take your time, think through each question, don’t be afraid to do multiple rounds of edits, and pretest the survey (internally and externally) with people who aren’t as close to it as you are. When it comes to research, no one likes to hear, “That’s really not what we were trying to find out.” Wouldn’t you rather know that before the launch?

Getting some help from a market research professional is always a good idea. Drop us a line to get in touch with the market research team at FARM.