You don’t have to be a usability expert to improve your product through user testing. How many usability problems you’ll find in your tests depends primarily on your test scenario. In this blog post, you’ll find helpful tips and tricks for useful test scenarios that we’ve collected over the last decade by creating and assessing thousands of user tests. Additionally, you’ll find pre-formulated test scenarios for different types of webpages that you can easily copy and use as a basis for your own tests.
Always Start with the Context
This is one of the most important tips for creating your test scenarios: Your test participants should be able to empathize with the situation as well as possible and be influenced as little as possible. They should behave as they would do outside of your test situation. The best way to do this is to describe the most realistic situation (the context) in which your product is to be used.
Thus, before starting with your first, specific test tasks, you should share with the testers the situation they are in when dealing with your product. The following structure is well suited for this:
- Introduce yourself…
- You would like to…
- That’s why you have to…
- This leads to…
Imagine that you’re planning your summer vacation for your wife and six-year-old son. You’d like to spend your 14 vacation days in Greece this year because your neighbor told you about the beautiful sandy beaches there. That’s why you’re looking for a suitable hotel for you and your family. You’re considering the last two weeks of August for your vacation because your wife has also got leave during this time.
After the context has been explained, we can finally proceed with the first specific task, such as booking a hotel with a good price-performance ratio in the specified period. The big advantage, if we explain the context well, is that these three sentences are often sufficient for our test participants to derive details for our test tasks on their own, and that we can formulate them in a very open manner, which in turn leads to our not influencing our participants too much with our test tasks. Thus, for example, in our task we no longer need to include finding a “family-friendly hotel”. We can leave our testers to get to this conclusion in their own words in view of the context.
Don’t Ask About Usability
It may sound contradictory, but the worst way to test the usability of your product is to ask your user. Maybe that’s why it’s one of the most common mistakes beginners make with their first tests.
Wrong: Please rate the shop for usability: Is the shop user-friendly?
Wrong: How easy to use is our website?
Wrong: Is the search form intuitive for you?
Having users rate the usability of a website doesn’t work. There are three reasons for this:
- People would rather tell you what you want to hear or what they consider you expect them to tell you in this situation.
- Human memory is faulty and highly selective; for example, we like to remember positive rather than negative experiences.
- People rationalize their behavior. “I was distracted at the beginning, but now I fully understand it.”
To learn something about the usability of your product, you have to give people a representative task and watch them while using the product. Thus, in the end, you’re the one who assesses the usability of your product based on data.
Right: Imagine your girlfriend’s birthday is coming up next week and you’re looking for a gift. You plan to spend a maximum of 50 €.
Don’t Ask About the Future
What’s more obvious in a user test than to ask if people would use your offer also outside of the test situation, i.e. “in the real world”? In the last months, we’ve seen a lot of tasks that were very similar to this one:
Wrong: Would you buy something on this page in the future?
Don’t ask your subjects about what they will do in the future, because they simply don’t know. On the one hand, people overestimate how much time they actually have in the future, and they also underestimate the effort actually required to use your product. Thus, the question about a hypothetical future won’t give you any usable results. Instead of asking about the future, ask about the past. This way you’ll learn about actual behavior of your test participants in the past. And past behavior is the best indicator of a similar behavior in the future.
Right: Have you purchased a similar product in the last 3 months? If yes, please briefly describe your experience. If not, please explain briefly why not.
You’ll learn a lot more if you understand what has influenced the behavior of your test participants in the past than if you make them explore the crystal globe for a nebulous future.
Don’t Ask for Opinions
User testing is not about asking people’s opinions, but about discovering usability problems through their behavior. Thus, you should absolutely avoid questions like the one below:
Wrong: How do you like this page?
Even if five test participants answer this question, we won’t get any usable results. The fact that all five of them find our site visually appealing isn’t representative with such a small number of individual opinions. If you really want to know how good your page looks, send a survey (a quantitative method) to a statistically representative number (preferably hundreds) of people. User tests are not the right method to conduct market research – even if many people believe it is, they’re wrong.
Avoid Leading Questions
Leading questions influence test participants by providing the solution to the question. Since we want to influence our participants in our tests as little as possible so as to observe the most realistic possible behavior, we should take care to avoid such leading questions:
Wrong: Did you understand that you can filter for specific products on our website?
This task is leading in two ways. On the one hand, it will be difficult for your testers to tell you that they “didn’t understand” something – even if they actually didn’t – because nobody likes to admit one’s own incomprehension. On the other hand, by expressing this question, you are implicitly saying that it’s possible to filter for products on your site. Thus, with your question, you’re anticipating the answer. A better way would be like this:
Right: Try to get displayed only those products that you’re actually interested in. Please don’t use the search feature for this purpose.
So, you see, we don’t need to talk about “filtering” at all. This way, we don’t give our test participants any indication about the solution to the problem, but simply observe whether they use the filter function on our site (if they have the task of displaying only those products that “actually interest them”).
Avoiding leading questions is often not so easy and it takes a little finesse to set tasks so that they anticipate as little as possible.