It can be hard to create solid user research questions that are easy to measure. Learn how to use desired outcome statements to create research questions with a strong experimental design.
- When you do user research, you need to focus so that you know what you're looking for. Unless you're performing very early-stage, formative, exploratory work, you and your product team should work out what you want to be able to say about users before you even think about the methods you'll use to find the answers. The best way to ask summative research questions is to start with what you'd like the answer to be, what I call desired outcome statements. Start by writing down a statement that you'd like to be able to make. This helps you work out how you're going to get the answer. For instance, users preferred option A over option B. Either an A/B test or a survey could give you this type of comparison. Because we're asking for preference, we'll use a survey. Next, you need to make sure that your statement is measurable. What does it mean to prefer something? We could show both options and ask, but then we don't know how much one option was preferred. Instead, let's ask for a rating for option A and a rating for option B because then we can compare those ratings. A good measure might be satisfaction with the outcome or confidence that the task was successful. Now, we can refine our desired outcome statement to be 80% of users were more satisfied with option A than option B. But we're still not quite there. How much of a difference is it important to have? For instance, users might show a strong preference or might really not care. Let's say you decide to refine the desired outcome statement with this difference. Now it might be 80% of users were more satisfied with option A than option B by at least two points on a five-point scale. Now, ask yourself what could introduce bias. For instance, if everyone saw option A first, it might prejudice them. Once you've taken different kinds of bias into account, you add this to your desired outcome statement. 80% of users were more satisfied with option A than option B in a randomized presentation by at least two points on a five-point scale. And now you have your desired outcome. You can turn it into a research question by writing it in a way that you think it would be measured. Users will be asked to perform a task with both option A and option B. Presentation order will be randomized. After performing the task, users will mark their satisfaction with each option on a five-point scale ranging from very dissatisfied to very satisfied. Now, that looks like a research question with a strong experimental design behind it. Answering this question will give the team good insight into their product development decisions. And, if they don't get the answer they were looking for, they can't just say that the research was wrong. This little example shows you how, by working backwards from the desired answer to the question, you can be sure that you choose the right research methods and that you get valuable results. You can find out more about creating good user testing task questions in my course, UX Foundations: Usability Testing.