Qualitative market research has a strong role to play in today’s data economy. As brands continue to collect increasingly vast sets of consumer data, often with some uncertainty as to what it all means, qualitative research is as valuable now than ever before, allowing researchers and marketers to go beyond the numbers and understand the deeper narrative behind consumer behaviour.
Over the last decade, new technologies have emerged which make qualitative research easier than ever. While in the past, researchers would often have to methodically trawl through focus group recordings or transcripts for their meaning, new technologies reduce the need for such manual handling, while also rendering them less susceptible to human subjectivity or error. This has enabled qualitative analysis to be done faster and to greater depths than previously possible.
While it is great to see new technology facilitating qualitative research, there remains an untapped area of potential when it comes to the analysis, and that is emotion detection.
Before we discuss the how, let’s discuss why analysing participants’ emotions might be valuable to a qualitative researcher. Emotions typically occur instantly and subconsciously, long before our brains can process the words to express them. As a result, emotions can be an effective way of gauging how a person is feeling without needing to expressly ask them. This can be particularly useful when broaching sensitive or uncomfortable topics or situations, but it can also be a valuable way of accessing a person’s unfettered, true thoughts about an idea or situation, irrespective of what might be ‘right’ or ‘wrong’ to say in any given context.
Until recently, analysis of participants’ emotions required a trained manual observer, limiting its applicability outside of academic research. However, a number of recent technology platforms such as Affectiva and Emotient have begun introducing machine-learning-based algorithms trained to detect human facial expressions with 90% accuracy. These technologies, broadly termed ‘Facial Action Coding Systems’ (FACS), work by using 3D facial mapping, assigning changes in facial expressions to distinct emotions based on large sets of facial data. These changes may be obvious and simple to detect, such as a furrowed brow to indicate anger or disgust, but they may also be undetectable to the human eye, based on thousands of small facial changes which together create a complex emotion like envy, or pride. Importantly, these platforms can analyse facial expressions without the need for invasive camera equipment, assuring conditions remain as natural as possible.
Such technologies have already been used to great success by brands such as PepsiCo, who track customers’ emotions while browsing products to understand their shopping experience. Similarly, other companies (i.e., Spotify) have begun to leverage this as part of customised brand experiences, tailored to a customers’ particular mood at any given time.
Altogether, there appears to be strong potential in this technology for focus group research. Aside from the content of a discussion, analysis of emotions data can provide an indication of what’s happening behind, or in the absence of, the conversation.
Combining this data with more traditional written qualitative analysis could be used to achieve even greater depth and nuance than possible through analysing the words of a discussion alone. In the case where emotional data supports the content of the discussion, researchers and marketers can know that the research manipulation has had its intended outcome. However, situations where a disconnect between emotions data and discussions emerges may in fact be more interesting to a researcher, as this can suggest there may be additional factors at play worth exploring within the topic. Why the difference? What’s preventing participants to express their thoughts freely? These are all valuable insights which can emerge through analysing emotions data as part of qualitative research.
The role of qualitative research in market research is only set to grow into 2024, assisted by the introduction of new technology enabling easier, faster, and more varied forms of qualitative data collection. Prime among these is participant’s emotion data, which can be useful for understanding implicit moods and cognition, as well as providing an additional layer of research credibility.