How to Use Surveys as a User Experience (UX) Research Method
As a user experience (UX) research tool, surveys can be an effective starting point for UX research—especially if your resources are limited and your stakeholders are skeptical. But, because surveys are cheap and easy to administer to get lots of results, product teams sometimes use them in place of conducting more extensive UX research. This is problematic because surveys have two major limitations:
It’s difficult to put together an effective survey. You have to carefully word each question and, because people won’t spend time on a long survey, you should limit your questions to 10 at the very most.
You can’t ask follow-up questions. So a respondent might say something really interesting but you can’t draw out deeper insights like you would be able to do in a one-on-one interview.
Because of these two limitations, surveys should be used with caution. That said, from a user experience research perspective, I find surveys to be helpful in these two scenarios:
To gather preliminary research from a large group of participants with the intention of doing more focused UX research based on the results. For example, when we have exploratory research questions and a large customer base dispersed geographically. The answers from the surveys will inform how we approach UX research that requires more resources, such as interviews or usability studies.
In tandem with one-on-one interviews, usability studies, or diary studies. For example, a diary study often includes survey-type questions (such as multiple choice questions) but also asks users to supplement that with photos and videos, often during or after using a product and/or conducting a relevant task.
Assuming you’re going to do a survey, here are some guidelines for creating an effective one:
Decide what you want to get out of the survey. What are your research questions? Keep in mind that surveys are not the best UX research method for every research question. If our team wanted to know what roadblocks were preventing users from completing a task using our platform, we would rely on a usability study—not a survey.
Find participants who are representative or actual users. There is no point in sending out a survey to someone who is not likely to use your product. Quantitative studies demonstrating statistical significance require more survey respondents than qualitative surveys. We recommend reading SurveyMonkey’s article on calculating a sample size.
Limit the survey to 10 questions or less. You can’t ask everything. Quality is better than quantity. Participants are more likely to give thoughtful answers when there are less questions. If you aren’t going to take action based on the answer to a survey question, don’t ask it.
Mix in a variety of questions (for example, multiple choice, open-ended, and scale questions) to keep survey respondents on their toes. You’ve been guilty of lazily filling in answers to surveys and your participants will be, too. Don’t bore them by repeating the same types of questions.
Ask questions about specific past behaviors as opposed to relying on participants to predict their future behavior. For example, “In a few sentences, describe the last time you…” or “How many times did you order takeout in the past seven days?”
Test your survey before you send it out. The question may make sense to you but it could confuse other people. Test your survey with a few people—ideally people who are similar to survey participants—to ensure you don’t end up sending out a survey where you’ll have to scrap the responses.
The goal of the survey is to answer your research questions and provide actionable insights for moving forward—even if that insight is that you need more information. Remember, surveys can be used to provide data that helps you decide where you should focus your efforts for more in-depth research, such as interviews or usability studies. Online survey tools allow you to administer surveys and filter and compare results. Nearly all survey tools show charts and visuals within their platform for close-ended questions. Some tools will even tag the text from open-ended questions for you. Regardless of your results, remember that surveys should not be the only tool in your user experience research toolbox.