Before recruiting respondents, we must define a target group. What people do we need to meet, or who needs to fill in a questionnaire? What are their characteristics, education, and experience with digital products? On one hand, the target group should be sufficiently homogeneous (e. g. all respondents do shopping at e‑shops at least once a month), but on the other hand sufficiently different to obtain information from different types of respondents (e. g. include men and women from different income groups, with a different level of education, etc.). Thanks to this, most potential and current clients of our product will be sufficiently represented.
We can get respondents in several ways
- The easiest and cheapest one is to ask our friends and family members. We must make sure that they are not our colleagues from the IT environment if this is not our goal, but ordinary users similar to potential customers.
- Another option is to reach out to our current customers if our relationship with them allows it. Clients know our product and can bring valuable ideas and knowledge that will help with our research. We can also address customers who still haven't decided but have us on their mind – just upload research information to the website or social networks.
- The third option is to get help from an agency that deals with this professionally. We can leave recruitment to them and avoid the need to get respondents. This method is the most expensive – the recruitment of one respondent can range from one to tens of thousands of crowns. Recruitment through an external agency is helpful if you need to find a specific target group that would be difficult to find yourself.
How many respondents do you need
The number of respondents we need depends on the type of research. For example, in a questionnaire survey, we will work with tens to thousands of answers. We obtain a specific number by considering the size of a group we are working with, the level of accuracy, and the range of errors we can accept. Calculators that are used for such a calculation will help, e. g. Survey Monkey.
For qualitative methods, we will need a smaller group. For example, during in-depth interviews or user tests, we work with 5 to 8 respondents. It is similar for focus group purposes. All this information is related to only one target group. If we conduct research with several research groups, the number of respondents increases proportionally.
How respondents behave
After successful recruitment, the research itself awaits us. It is good to avoid a few of the most common mistakes. Always keep in mind that:
- Our respondents do not think the way we do
Users don't usually appreciate things that are important to us (such as a banner ad on the main screen) and can use different words or product paths than we designed. They might also think about the research problem differently than we would expect. This may be due to their technical knowledge, personal experience, and other circumstances. Therefore, it is necessary to confront the product with users several times in the design process.
- Our respondents often do not know why they behave the way they do
Users sometimes do not know or do not want to say the reasons for their behavior. Sometimes they even come up with the reasons. Therefore, it is important to observe them and find out their behavior when confronted with the product. Likewise, they are often afraid to tell us their opinion or do not remember the essential details that are important to our research. In this case, it is important to assure the respondents that we do not judge them in any case and that we are interested in their opinion.
- The best prediction for future behavior is past behavior
Opinion polls are often inaccurate and differ from the actual behavior of respondents. Therefore, it is not appropriate to ask respondents whether they would use our product or appreciate the feature we implement in it. Instead, our role is to identify respondents' past behavior and predict their future behavior.
- The behavior of our respondents depends on the context
Different respondents will behave differently in the same situation. Therefore, it is ideal to conduct research in their natural environment, in situations that interest us. User testing in laboratories is often sufficient, but by moving our research into the natural environment of the respondents, we can also capture the influences that we would miss in the laboratory (e. g. stimuli that interfere with the use of the product or transition to service). If we can't conduct research in the natural environment, we should try to make it as real as possible (e. g. tell the respondents to imagine buying insurance, give them real money on the card, and let them arrange it). In in-depth interviews, it is useful to ask questions in such a way as to identify contextual influences. E. g. If we ask about ordering food online, we will try to find out all the circumstances that led to it, the elements that facilitate this process, or obstacles that disturb the respondents in this activity and prevent them from satisfying their needs.
Don't conduct the research yourself
If possible, people outside the UX team should be involved in the research. Developers, product owners, or managers perceive investigations completely differently when they not only examine our findings but are directly present at the testing. They see how real users interact with the product or how they perceive the problem you are solving.
In the last article in this series, we will focus on evaluating research and presenting our findings to benefit everyone involved and motivate them to take action.