- •How to conduct user research
- •Joel Marsh Chapter 1. What Is User Research?
- •Different People Will Say That User Research Happens at Different Stages in the Process
- •Subjective Research
- •Objective Research
- •The Less Obvious the Problem, the More People You Need to Find It
- •Get Users Who Fit the Profile and a Couple Who Don’t
- •Users Are Not a Crystal Ball
- •Don’t ask users how they want their problems solved
- •Consensus Is Not a ux Strategy
- •Chapter 3. How Many Users Do You Need?
- •Chapter 4. How to Ask Questions
- •Three Basic Types of Questions
- •Do Not Help
- •True Story: People Will Lie to Get Your Approval
- •Chapter 6. Interviews
- •If you want to ask some questions or see people try something you have designed, you need to meet them face-to-face. What Is an Interview?
- •Interviews are good because...
- •Interviews are bad because...
- •Surveys are bad because...
- •You should use surveys when...
- •Chapter 8. Card Sorting
- •What Is Card Sorting?
- •Card sorting is good because...
- •Card sorting is bad because...
- •Useful profile
- •Why it’s useful
- •Think of “Ideal” Users; Several of Them!
Do Not Help
It will be very tempting to help a user who is confused or uncomfortable because they don’t understand. Resist that temptation! The moment you help a user, or give them a hint, or point to something useful, the test is ruined because you are now testing the design instead of them. Letting users fail is sometimes the most useful result you can get.
True Story: People Will Lie to Get Your Approval
Don’t forget that you are in the room. People will lie to cover embarrassment, or act helpless to get your help, or tell you the design is great when it isn’t—just because you’re sitting there.
Always approach user testing as if you expect users to lie about something, even if they don’t realize it, because you can’t trust a user any farther than you can throw them. There we go: best metaphor ever. Thanks, Mom!
Chapter 6. Interviews
If you want to ask some questions or see people try something you have designed, you need to meet them face-to-face. What Is an Interview?
An interview is a set of questions, created by you before the interview, asked to a user, in person.
Interviews are good because...
You can ask follow-up questions; find out if your questions are confusing; give people tasks to complete; and get long, open answers to questions that might be harder to answer in writing.
You can also watch users and get nonverbal clues, and you can learn about things where time limits are built into the experience, like games, quizzes, or real-time messaging.
You can hand-pick the testers.
Interviews are bad because...
You are there, so testers might adjust their behavior and opinions to get approval from you.
It is harder to get real people to come to a place that is good for you, so you will usually end up testing with fewer users.
The social nature of a face-to-face interview is not good for embarrassing or private products and services, like that site you buy all your latex body suits from.
Introverted users probably can’t imagine anything worse than a face-to-face interview.
You should do interviews when...
You need to test subjective stuff with many steps or decisions involved, like navigating around a site to select the perfect latex body suit. Or if you need to ask follow up questions depending on the user’s behavior, like “which ball-gag goes with that body suit, in your opinion?”
Chapter 7. Surveys
Face-to-face interviews aren’t always an option, so sometimes you need to send questions out to people. But be careful, there are things a survey can do, and things it can’t.
What Is a Survey?
A survey is a set of questions that a user answers by filling out a form on paper or online, privately, and sometimes anonymously. It’s like a Buzzfeed quiz, but instead of finding out which Harry Potter character you are, you just give feedback.
Surveys are good because...
They allow users to participate privately and therefore be more honest.
Every user gets precisely the same questions, and you (the designer) can’t screw it up by asking the questions wrong.
It is easy and cheap to get thousands of people to answer a survey. If you do that in person then you also need portable toilets, food trucks, bands to perform... it’s a hassle.
Nobody has to feel disappointed, because a UX survey never says you are most like Malfoy.
