What are the best practices to document interviews with users - do you just document the conclusions and take aways or do you do a full transcript or both?
|
I would guess that anyone actually working in UX would not have the time to do a full transcript for each interview. I typically get permission to record the interview with my phone and make a few short notes from time to time as i go. I will listen to the full recording later in case I missed something important. I find a simple spreadsheet document with a sheet for each interviewee each containing a column for the questions, another for the corresponding responses works well. In the rare event that I can't get them to let me record, then I simply need to work harder on live note-taking and lose a lot of the 'conversation' element of the interview. When presenting findings to others, I naturally I summarize the important points with some visual data like pie charts etc. alongside some key quotes from the interviews - I sometimes tie them to personas. Hope this helps. |
|||
|
|
A full transcript isn't necessary if the session is recorded - instead, I've always delivered user testing results in the following broad format:
If you feel that isn't quite enough for your stakeholders, something else you can do is invite them to observe a session remotely with the observer. This is a good opportunity for them to converse with the observer and for the observer to hopefully make the case for UX's claims. |
|||||
|
|
There is no best practice :-) You're doing a user interview for a purpose. The output from that interview needs to be something that's fit for that purpose. For example, currently we're doing some user interviews to validate some assumptions we've made about how a particular product idea will be used. All we're doing there is taking some notes during the interview on post-it notes, heaving them up on the wall back at the office, and affinity diagraming the heck out of 'em. Because doing any more than that would be a pointless waste for our context. Always approach deliverables like any other UX problem. What's your working context? What does the end-user of the output from your user interviews need? That'll give you the clues you need to figure out the best deliverable (if any - a conversation might be best - depends on your context). |
|||
|
|
