Tom Woodnutt
Tom Woodnutt is Founder of Feeling Mutual, the multi-award winning agile online qualitative research specialists. He helps clients and agencies run global studies and offers training in the space. Tom has been a Digital Skills trainer for the Association of Qualitative Researchers (AQR) and is a regular speaker at industry conferences, including the MRS, MRMW and IleX.
I am a big fan of asynchronous online qual methods, by which I mean online text based methods like diaries, forums and communities. One thing I like about them is the way they give you so much detail and instant high quality data to work with - this meets the need for depth and enables rapid analysis and reporting within an agile project workflow.
Accelerating online qual with AI [video]
4 ways AI works better with asynchronous online qual
In many ways AI tools are currently better placed to add value to diaries, forums and communities than they are in real time methods like depth interviews or groups.
Greater depth of data
Asynchronous methods need AI summary tools more than other methods because they can create a lot more content. That's because in a focus group for example, only one person can speak at once. So when you invest in 2hrs of feedback per person in a 6 person focus group, you actually end up with less than a sixth of that (ie. less then 20 minutes per person). Whereas in asynchronous online qual they can all speak in parallel. So you get 6 times more feedback and a full 2hrs per person. This higher volume of data makes the need for AI summaries all the more significant.
More structured data
Another reason is that asynchronous online qual organises the data by question or concept if applicable. So this means it’s easier for the AI summary to target the relevant text. Whereas if it's working on an entire transcript or multiple transcripts - it sometimes struggles to focus on just the feedback associated with a particular concept or question. Some AI summary tools have work arounds for this (as you can break the content down by question or concept and run summary analysis on that isolated set of text). But this can be laborious.
Higher quality verbatim
The data from text based asynchronous online qual is of such high quality - it doesn’t have all the typos and misinterpretations that you get with AI transcripts from face to face or webcam discussions (and it doesn't take as long or cost as much as human transcription does). AI transcripts are impressive, fast and they save money but they do not give you the full, rich, high quality verbatim and you end up having to go back to the video source to fill in the blanks.
Avoids delays in automated probing
Automated probing using AI is fairly impressive (although not as good as a human researcher). It can offer thanks, ask for elaboration and some platforms even let you train the AI moderator to ask particular probes triggered by particular responses. However there can be a frustrating delay if this is in real time as the AI works out its probe after the participant has answered. I think this creates some friction and reduces the flow of a discussion. So AI automated probes in asynchronous methods (i.e. discussions that are not in real time) could evade this issue.
AI adds more value to asynchronous online qual
So overall, I think AI tools are going to add more value to asynchronous methods like diaries, forums and communities than they will traditional qual research methods like real time focus groups.
