Should we still use qualitative research in an age of big data?

John O'Neill
Conversations with the right people are priceless

In the year that marks the 80th anniversary of the invention of the focus group, and the establishment of the eleven trillionth ‘big data’ startup since the term was coined in 2005, one could be forgiven for wondering about the place of qualitative research in the modern economy. 

The focus group in particular has been hauled over the coals pretty regularly. A Harvard Business School professor pointed out the tenuous link between stated opinion and behaviour, which is the most common criticism of qualitative work. How can you rely on information that depends on honesty, when so few of us are really honest about what we think and feel? Just consider how someone’s Google search history differs from their social feeds. That Venn diagram barely touches. 

It’s not always clear, in 2021, what differentiates qualitative and quantitative. It used to be that any information not numerical was by definition qualitative, because it could not be statistically measured. Now, though, AWS can easily take your company’s terabytes of long-form textual customer feedback and transform it into binary data to conduct a sentiment analysis. So although a piece of information might start out as qualitative, making sense of it is a quantitative exercise, especially when there is a lot of data.

I would like to humbly submit that qualitative research is not only still relevant, it is an essential step to direct quantitative work. It works best in tandem with quant, and in the pursuit of better customer experience, it is the best way to tap into human emotion. That’s especially true for companies without the resources or the Amazon/Google-sized data pool from which to profile customers down to the year and composition of their first tooth filling. 

The strength of quant is that it is not subjective – when done well, it can be replicated and can reveal improvements to all sorts of things, from price to supply chain to where the ‘purchase’ button should be placed. The strength of qual is that it is subjective because it’s dealing with human emotion. A paradox. The point of qualitative research is to understand the emotional context of decisions – the famous ‘why’ that, in the end, drives all the decisions human beings make. Statistics too should be taken with a grain of salt. Nobel Prize-winning physicist Lord Rutherford: “If your experiment needs statistics, you ought to have done a better experiment”. 

With vast amounts of quantitative data about what people click on and what they search for, sometimes it is possible to get close to emotional information about an individual. For the vast majority of organisations that don’t have this, a skilled interviewer or workshop moderator can do it, and accurately, in an hour.

Qualitative work has its origins in psychology, biology and anthropology, not statistical science. Ethnographic research like observational in situ studies and mystery shopping reveal and articulate what people actually do rather than what they say they do. In-depth interviews uncover powerful insights – you might know that a lot of individuals on your website are veering away from a particular user journey you’d like them to make, but do you know why? Do you know how you could turn that emotional reaction around, or even turn it to your advantage? Chances are, not unless you actually speak to your users. The Nielsen Norman Group has done extensive work on how many of them to speak to since their original 2000 paper, and it’s not as many as you’d think.

Qualitative and quantitative belong together and are far less effective alone. I couldn’t say it better than Professor Mohanbir Sawhney of Northwestern University’s Kellogg School of Management:

A consumer insight never comes from quantitative research, so don’t look for it in your surveys. Qualitative research and quantitative research need to go hand in hand for true customer understanding. Qualitative research generates insights, while quantitative research validates insights. You need both research traditions, and in the right sequence.