Today started off reviewing our homework. We had to write an objective and 3 to 5 tasks to review the inmod web site. We spent the first half hour reviewing the tasks in small groups. I am always surprised when working in small groups how easy it is for people to take the group off track.
The big topic for session 2 was about conducting the user test. You need to make sure that not only you are prepared, but everyone involved is prepared. The steps of a user test session are Introductions, Run the test session, Debrief the user, and Prepare for the next session. You should prepare the participant for what to expect, and make them comfortable. Stay as neutral as possible during the session. Get any final feedback you may need from the user, and answer any questions they may have. Then reset the computer and get your notes ready for the next test.
After all the sessions, it is time to analyze all of your new data. You organize your data into Findings with supporting details, assign each of the findings a Priority, make Recommendations that are based on the findings, and then cycle your changes into the next development cycle. All of this information should be included into a Report of your Findings. You should try to provide a short, informal report within the first 24 hours, and a more detailed, formal report within two weeks. Your reports should focus on the positive (what worked well) as well as the negative (what needs improvement). You should also formally present your findings to your client. Keep your meetings short. Leverage the video you took and include quotes from the users.
Jakob presented variants to the user testing methodology. You can test more than one user at a time, if it makes sense (like husband and wife, co-workers, etc). When you cant go to the user, and the user can’t come to you, remote testing is one of the last possibilities. You can also test more than one site (either two separate designs, or competing sites). Sometimes you may want to follow users over an extended period of time, so diaries or videos can be used. Eye tracking is a new technology that is very useful with video recording.
Based on the studies that NNG have implemented, tehy have found that the ROI per user tested is maximized at 4 users. They recommend on testing 5 users to be sure that you are safe in your results. But, anything beyond 5 users, the number of findings flattens out but the cost continues to increase incrementally.
You also need to be very careful when testing users with special needs. Disabled users and low literacy users should be tested with simpler tasks and shorter sessions. Senior citizens love teh attention and are over polite, but should also be given less tasks, and expect more time for the introduction and wrap-up of the session. Testing childres is also very different than testing adults. Testing in schools is ideal, if you can get permission. Shorter sessions and co-discovery methods make testing easier. International testing can be much more expensive, as it requires translators or much more travel. Hardware testing works similar to software testing, but you need the real product to test.
Jakob closed this session with a discussion of the ethics of user research. You need to remember that these are people that you are testing. You need to treat them with respect and dignity. The rule of thumb is that you should treat them as you would want to be treated.
After today’s session, a few of us went to a sports bar to watch the North Carolina / Michigan State game. The three of us that went all had ties to Michigan. Shawn and his wife live there now, Rebecca has bounced back and forth between Detroit and Chicago, and I used to live there when I was very young. Naturally, we were all cheering for Michigan State. After 3 minutes of the game, Michigan State was down 15 points, and they never recovered.