A couple of weeks ago, I attended the MRS Big Qual conference. With passionate practitioners and cool tools rapidly speeding up data collection and analysis it’s an exciting time for qualitative research. The industry feels on the cusp of the sort of disruption we’ve seen in quant over the past decade.
What struck me however was the underlying assumption that research tech will enable quallies to perform their roles more efficiently. I believe we need to reframe this thinking. Rather than imagining how tech will enable ‘qualitative research at scale’, research tech is borrowing from qualitative research, as well as other disciplines, to generate entirely new forms of value from data.
Let’s look at how qualitative research is showing face in a Big Qual world.
Qual is… Responsive questioning
Traditional qualitative research, which I’m going to refer to here as ‘Little Qual’, uses semi-structured interviewing guides which can flex to respondent context to uncover deeper meaning. Big Qual, in contrast, enables us to be more expansive at the start of projects by drawing from existing datasets to develop stronger hypotheses; meaning we need to ask fewer, or even zero questions of humans.
Discover.ai, working on behalf of Unilever Homecare, were able to identify 10 emergent themes at pace through meta-analysis of in-market content and conversation. Whilst Listen & Learn were able to identify 169 graphic art tribes on TikTok for Colart, before narrowing down to 5 tribes for deeper investigation. Neither asked any questions of people.
If Little Qual is about asking people questions to get at meaning, Big Qual uses qual’s smart questioning techniques to create new meaning from datasets. Just as we have adapted to ask Alexa questions more effectively, or refined our search terms on Google, researchers are flexing to the machine.
Qual brings… Context & nuance
The idea that humans are more adept at understanding context and nuance than machines is held onto as something of a totem by qualitative researchers. Even with all the great advances we’ve seen in natural language processing, tech, it was argued, will never be anything but a supporting act.
Yet as Ryan Howard of Malliavin Marketing Sciences pointed out, humans are far from infallible. In his experience, if you get three individuals to sentiment code the same dataset you end up with 65% variance between them. The reality is that we all bring our own biases to the party, whether you are human or machine.
In Little Qual, the researcher and their interpretative input is key to the process. In the world of Big Qual, the definition of qual is reduced to the fundamental practice of making sense of unstructured data using a set of analysis techniques rooted in language and context.
Qual is… Not about numbers
Qual isn’t interested in numbers, but in the in-depth examination of small-scale samples, reporting on meaning rather than incidence rates. Yet the big win of Big Qual is that it provides the speed and scale that Little Qual lacks. The ability to analyze, size and recut data and sheer ease of reporting via dashboards means you inevitably get to measures.
What can be measured will be measured. The question is on the validity of those measures. Relative Insights (who win the award for the most fascinating origin story!) – reported their IPA client data as ‘X times’ more or less likely, whilst Listen & Learn and Discover.ai would not be drawn on measures at all.
Whilst Little Qual can hold the line when it comes to the validity of quantifying qual, as we move to a Big Qual future with automated dashboarding, there will be an inevitable move towards measurement. Raising the question, at what point does qual become quant?
The future of qual in a tech-first world
Being an accomplished guitarist doesn’t give you an advantage at Guitar Hero – if anything it’s a hinderance. Likewise, it’s no use pointing out how AI falls down against traditional qual because the digital game is nothing like the ‘real thing’, it just borrows elements from it. Research tech is borrowing from qualitative research techniques to create entirely new forms of value from unstructured data.
Unlike quant, which lends itself wonderfully to technology, I still can’t see a world without traditional qual. Rather I believe we will see an entirely new breed of researchers, seeking to work in true collaboration with technology - taking inspiration from a range of disciplines including qualitative research, social intelligence, semiotics, digital anthropology and more.
At Verve we are building a research agency of the future. We offer a full range of services, from face-to-face qual, to community panels, alongside in-house expertise in semiotics, social intelligence, behavioural science and anthropology.
If you’d like to talk more about how we could help you, I’d love to chat!