Process and learnings
User research should be part of everything you create. If you are not able to do the research you won't understand what the problems are the users have and you will never be able to solve them. There are many ways to conduct user research. These are some of the types of research I have experience in:
- Desk research - check what the competition is doing, what are best practices and what can be quick wins.
- Data research - look into the data and try to find the bottlenecks in your flow. Where do users drop off?
- User testing - let the user go through your application/website and see where they run into problems.
- User interview - with all other research you can get the answer to what the problem is, but talking to your users will answer the question 'why?'
I will discuss the last two in more detail.
- Preparing the user tests & interviews
- Moderator or observer
- Creating evaluation reports
- Creating actionpoints for teams to make sure the learnings were used in day to day work
At Zylom, a couple of years ago, user testing was something completely new and together with 2 colleagues we set up a user testing process where users come in the office every two weeks to test our products. I led the team in how to conduct the tests and made sure we had something we wanted to test.
What I've learned in user testing
- Make sure the user is relaxed. They shouldn't feel it's a test, they can't do anything wrong.
- Have a couple of assignments. Don't make them too long.
- Use different kind of tasks. Impression, exploratory or directed tasks.
I've conducted user interviews where multiple users were in the room or where it was a 1 on 1 talk. There was always an observer in the room who made notes. At the end of the interview the observer had the option to ask more questions. It's very important how you formulate a question, you can easily lead someone to an answer and that is not what you want. Try to always ask open questions to get the most honest response.
What I've learned in user interviews
- It's ok if it's quiet.
- Listen, probe, and validate.
- Don't explain everything. You don't have to defend your designs.