Monday, 16 March 2009 at 5:00PM
Panel:
- Mark Trammell - Digg
- Juliette Melton - Lumos Labs
- Nate Bolt - Bolt|Peters
- Carla Borsoi - Ask.com
- Andy Budd - Clearleft Ltd
Who has people doing user research? Most of the audience does.
What tools do you use?
Budd: Talking directly to users to see how they go about doing what they do. Borsol: 1-on-1 interviews, heavy use of analytics and then surveys. Melton:
Budd: Eye-tracking is not need most of the time to figure out what is getting used and not used. It is usually quite obvious to see in the design.
There is an anti-focus group feeling permeating the panel. Budd notes that determining attitudes is a benefit of focus groups.
Those that don’t do user research and just design for themselves, but what they have going for them is that they are designing for a very specific group or person. In the case of Apple - Jobs or 37Signals - themselves. If you think you understand your users without research you run the risk of over or underestimation of your users. You may never try anything new because you get stuck in the rut of what you have always done. Apple’s research is actually that they build lots of prototypes. You cannot look at Apple and say that research is not necessary. Genius design is rare - most people need user research feedback.
Design, good designs do not come via epiphany most often, iteration and user feedback creates good design.
Remote testing: viewing via screen sharing or watching the user remotely. Bolt is writing a book about this [link]. It allows users to be in their natural habitat while they are shopping/perform whatever task being tested. Remote testing can be handy if you are developing for a remote audience in a remote market. They may have context pressures that would not indicate using a different demographic for testing (i.e., local to the designer).
Getting the team involved: Show the team what is not working. Silverback is a great tool for watching a user struggle with an interface, for example. It motivates the team to help real people. Then show the team a success from the decisions that were made based upon testing. It makes it real for the next iteration. People like numbers. Your research is giving you that so give them soundbite. Budd is a bit annoyed at having to bribe your design team to care. Hire different people, he says. He is a bit Bluesky, methinks. Me: There are many times that designers are not the only team members. Support people/staff that are part of the larger organization, for example.
Don’t ask permission to do your job. A good rule of thumb.
Quantitative vs. qualitative
How do we pay for it all?
There are a plethora of cheap and dirty tools out there: card sorts, Silverback, SurveyMonkey, etc.
Microsoft’s method when writing Age of Empires II called Right Methodology?