New: card sorting is now available! Click to learn more.

User Sense

7 Expert tips for moderating user tests and interviews

Whether you already have some experience running moderated user tests or want to get started with them for the first time, in this article we'll give you seven concrete tips to get you started.  

1 - Put the respondent at ease 

Many people find it scary to do things they haven't done before and this is no different for participants in usability research. As a result, respondents are often nervous at the start of the test, especially those who are less digitally savvy.  

Clear communication prior to the test helps to reassure respondents. Think for example of sending the instructions by e-mail and follow up by telephone the day before the test.  

Use small talk to reassure the respondent. For example, talk briefly about the weather or how their day has been so far and let them know that they cannot make mistakes during the survey. This will help you break the ice, making the rest of the testing session easier.  

2 - Avoid steering questions  

The biggest and most common pitfall of novice moderators is asking leading questions. This is because the conversations you have during a moderated user test are different from the conversations you have in your daily life and because you already know how the website works. Whereas in daily life you often want to hint at your opinion or are looking for common ground with your conversation partner, as a moderator you want to remain neutral.  

In other words, asking leading questions is something that comes naturally. Yet, as a moderator, you want to avoid them because it can affect the results of the study.  

Therefore, it is good to be aware of what constitutes a guiding question. Example of guiding questions: 

  1. Did you find that easy to do?  
  2. We see that many visitors drop out on this page. What do you think of this page?  
  3. Research shows that many marketers work overtime. Is that true for you too?  
  4. So this functionality fits well with what you're looking for?  

Better is: 

  1. How did you find it going?  
  2. Take your time going through this page. What do you think of this page?  
  3. Can you tell me more about what your work week looks like?  
  4. What do you think of this functionality?  

3 - Do a test run 

Test script created and aligned with all stakeholders? Then make sure you ask some colleagues to do a test run with you. They can then assume the role of tester. There are quite a few advantages to this: 

  • You get an indication of how long the test session will take, so you can delete or add questions. 
  • You become familiar with the test script and discover which parts require revision so that the test sessions will run smoothly.  
  • Helps you build confidence or get back into it if you haven't been a moderator for a long time.  
  • Gives you a chance to get feedback on how the tester (in this case, a colleague) perceived you as a moderator. 

4 - Introduce 'awkward' silences

Whereas in real life we try to avoid uncomfortable silences as much as possible (or just enjoy them when watching First Dates), dropping silences is one of the success factors during a moderated user test.  

By dropping silences, you give the respondent time to think and share additional information with you. Crazy tip: make sure you have a cup of tea, coffee or glass of water and take a sip each time during a silence. This makes the silences feel less awkward and gives the respondent time to share his or her thoughts.  

5 - Let the respondent wander off 

Do you notice during a session that a respondent is going in completely the wrong direction? Then make sure that you do not intervene too early. It is often interesting to see which pages the respondent visits and how he/she searches for the information. Is the respondent really lost? Then of course you can intervene. But in general, the more freedom you give the respondent, the more you will get out of the survey.  

6 - Try to answer questions with questions 

Nobody likes to make mistakes. And even though it is often not the user's fault, users often blame themselves. It often happens that respondents ask "Am I on the right page now?" or "I don't think I'm doing something quite right, do I?". 

To really understand user behavior, it is best to answer such questions with a different question. For example, "What makes you wonder if you're on the right page?" or "What's the reason you think you're not doing it right?".  

7 - Recruit a reserve respondent 

In practice it is almost inevitable that a respondent will be unable to attend. To prevent the numbers from being too small to draw conclusions from the survey, it is often a good idea to schedule a reserve respondent. With five test sessions, often 6 respondents are recruited and with ten test sessions 12 respondents.