As mentioned in a previous post, putting users at the center of your design process can help you ensure an optimal user experience. There is growing interest and adoption of a user-centered design approach opposed to a systems-based approach. In this post, I will explore how segmenting your audiences and interviewing them, will help you better define the problem you are trying to solve.
What is User-Centered Design?
User-centered design (UCD), as the name suggests, is a design and development methodology that focuses on building a web site or product for a group of target audiences. UCD incorporates research activities to answer questions about users and their tasks and goals. The findings from this research informs design decisions about the final web site design.
UCD activities seek to answer the following questions:
- Who are the target audiences / users of the web site?
- What are the users’ tasks and goals?
- What is the users’ experience with the web site / application, and others like it?
- What information might the users need, and in what form do they need it?
- How do users think the web site / application should work?
The first step to understanding your audiences is identifying them. Stop and ask yourself who you are trying to reach and how you will attain your organizational goals once you do reach them. Write down the full list of audiences and take a few minutes to describe each type:
- How often are they online?
- Basic demographics: age, occupation, income level, technological experience, career-level, gender, etc.
- Technology: connection speeds, browser & platform specifications
- What are the audiences’ goals / needs?
For each audience type, write down the goals that they may want to accomplish on your web site. What tasks will the users need to complete on your site to complete that goal? List them next to each goal (Search for information, become a member, download a report, etc). Are the tasks unique to each role, or is there any overlap? Tasked-based segmentation allows you to group audiences by the goals that they are attempting to accomplish rather than role, thus potentially saving you from segmenting your audiences into many sub-roles with similar goals.
Audience Research: Getting Started with Interviews
Not sure where to start? Stop by your company’s marketing department (or equivalent) – it is a great place to help you determine your site’s target audiences. The marketing department has probably already conducted a fair amount of research on demographics and technological requirements of your target audiences. The team sponsoring the design will also have a sense of who the users are and why they want to reach those users. Schedule a few one-on-one interviews with members of this team to gather as much information about the target audiences as possible. Listen carefully to what they have to say about the audiences and use that information to help you decide whom to talk to later.
Once you’ve determined who your potential web site visitors are, the next step is to go talk to them. I recommend preparing a script for your interview. Consider asking the interviewee the following questions:
- What other web sites do you visit?
- How do you currently use this site?
- Do you use any competitor’s products?
- Are you interested in potential features?
- How do you know when the site has new content?
- Do you register for newsletters?
- Do you use a non-web version of the product (mobile, print, etc.)?
Remember that you are not looking for statistical evidence, you are trying to figure out the mental model people may have for the site you are about to design. When possible, it is best to use actual users of your site; however, you can use guerrilla techniques, such as asking friends and family questions about the subject matter.
A Good interview is neutral and probing. The following are some tips to help you be a strong interviewer:
- Consider how you word questions. Don’t ask questions that can be answered with a “Yes” or “No” answer. You want the interviewee to provide you with as much information as possible while answering a question.
- Keep your opinions to yourself. You do not want to steer the conversation in any way. You are interested in the interviewee’s thoughts and do not want to influence their responses.
- Encourage responses. The key to a good interview is keeping the person talking on a subject. You want to find out as much as you can, so let the interviewee know you are listening with encouraging, yet neutral responses, such as “I see,” “hmmm,” or “oh?”
- Ask follow-up questions. Similar to above, keep the conversation going. For example, ask the interviewee to “Tell me more.” Incomplete sentences, such as “and your goal was…” are great follow up questions.
- Ask interviewees to show, rather than tell, when possible. Actions speak louder than words, so encourage your interviewee to show you how they use the current web site to answer a question. Another effective technique is to ask users to sketch out the functionality that they are proposing.
Document all your findings in detailed notes so that you can reference them later.
Analyzing Your Results:
After the interview, read over your notes and look for recurring themes. Make a special note of problems or issues that had repeatedly come up. You are also interested in clues about how people think about the subject matter. The language that people use to describe the subject matter, especially if repeated across interviewees, could become labels for your web site’s content categorization.
While analyzing your results, it is important to keep in mind that when an interviewee is requesting a new feature, or changes to the current design, you will need to evaluate if the request is unique or representative of a wider set of users. If you are unsure, talk to more users until you are certain.
The next step is to take your findings and design a new prototype of the web sites. Stay tuned for a future post, that will discuss how to evaluate that prototype with user feedback.
For more information on user-centered design, audience segmentation, or interviewing consider this resources: