Like we said before, doing voice research is very much different from usual forms of user research. Last time (read pt. 1 of this article), we shared tips and tricks on the pretest and the observation of your voice research. Today, we pick up where we left off and look at the posttest and reporting of your insights.
- Plan more time for the evaluation.
You shouldn’t ask questions to the participant while he or she is interacting with the voice assistant — this sounds very logical, but it’s counterintuitive for most researchers. That means you have to ask all your questions afterwards, during the posttest. So you simply need to reserve more time for that.
- Remember that people have a bad memory.
Right after people interact with a computer or voice assistant, they can still explain their thoughts and what they tried to accomplish pretty well. But only a few minutes later, it’ll be harder for them to explain why they did what they did. Because you ask your questions afterwards, it might be difficult to find the real reason behind certain choices and actions when doing voice research.
- Emphasize that the test material doesn’t represent the ‘state of voice’.
If a participant doesn’t have experience with VUI’s and the test was hard, it probably influences their attitude towards voice solutions. Even though it’s only one individual, let him or her know that the test material is exactly that: test material. You don’t want them to leave your building with a voice-anxiety :)
- Realize that it’s relatively hard to show the context of an insight.
With a desktop or mobile solution it’s pretty easy to link insights to specific parts of the flow, by for instance using screenshots. With voice interactions this is much harder, but it’s still possible. For instance, by describing the context clearly, by typing out a part of the dialog between participant and VUI, or by adding video footage of the interviews to specific insights.