Beyond the big patterns: The value of qualitative research
In market research, we need to simplify. We group people into segments, turn responses into percentages, and identify patterns that allow us to describe a market and make decisions about it.
This is an essential part of our work, but every simplification comes at a cost. The more we summarise the behaviour of a group, the easier it becomes to lose sight of the differences that explain why seemingly similar people ultimately make different decisions.
A finding such as “7 out of 10 consumers value sustainability” may seem to offer a fairly clear conclusion. Yet behind that statement there may be people concerned about reducing their environmental impact, consumers looking for more durable products, people interested in where products come from, and others for whom sustainability is simply a positive attribute, but one that remains secondary to price.
They may all have selected the same option in a questionnaire. The question is whether that is enough to conclude that they are expressing the same underlying need.
The same answer can conceal different needs
This is something we encounter constantly in qualitative research. Imagine asking several people what they expect from a bank, and they all mention “security”. One may be thinking about fraud prevention; another about the financial stability of the bank; another about receiving an alert when something unusual happens; and someone else may simply mean being able to speak to a real person when they have a problem.
The same applies to apparently straightforward concepts such as convenience, quality, trust, innovation or value for money. Two consumers may use exactly the same word to describe quite different needs. Before turning a recurring response into a conclusion, it is therefore worth understanding what experiences and expectations we are grouping together under the same label.
Qualitative research is particularly valuable at this point. An in-depth interview, focus group or contextual observation allows us to explore what a response means to each person and the specific situations in which it becomes important. This enables us to move from identifying a pattern to gaining a better understanding of what is producing it.
Aggregated data can conceal very different markets
The opposite can also happen. Two groups may provide very different responses while actually reacting to a shared concern that is not immediately apparent.
Consider, for example, a new AI-powered service. Some consumers may welcome it enthusiastically because it saves them time, while others may reject it because they feel they are losing control over the decision-making process. If we limit our analysis to measuring acceptance and rejection, we end up with two clearly differentiated positions.
Looking more closely, however, we might discover that both reactions are connected to the same underlying need for control. For some people, automating certain tasks means regaining control over their time; for others, it means giving up control over a process they would rather manage themselves.
This second interpretation opens up different possibilities for a business. It is no longer simply a matter of identifying who accepts or rejects the service, but of understanding which aspects of the proposition are driving each reaction and under what conditions it could become relevant to different types of consumers.
Contradictions are information too
In research, we constantly encounter differences between what people say they value and what they ultimately do. A consumer may claim to compare options extensively before making a purchase and then choose one of the first products they encounter. They may consider privacy essential while automatically accepting an app’s terms and conditions, or say they prioritise quality and ultimately choose the cheaper alternative.
These differences between stated attitudes and actual behaviour do not necessarily invalidate what someone has told us. Real-world decisions are shaped by available time, price, information, context, habits and immediate priorities. This is precisely why examining where these contradictions emerge can provide particularly valuable information about the factors that ultimately determine a choice.
Qualitative research allows us to explore these areas of apparent inconsistency in greater depth. Rather than trying to decide which response is the “true” one, we are interested in understanding which factors come into conflict and which ultimately carry more weight when a person faces a specific situation.
Research is not just about finding majorities
In any study, there is an understandable tendency to pay more attention to what occurs repeatedly. If nine participants mention an issue and only one raises something different, the first issue automatically seems more important — and in many cases, it probably is.
However, frequency and relevance are not exactly the same thing. A minority observation may point to a barrier that currently affects only a small number of users, reveal a different way of engaging with a category, or anticipate an expectation that is not yet widespread enough to appear clearly in aggregated data.
This is particularly important in innovation, service design and the exploration of new value propositions. If we focus exclusively on what is already established, we may develop a very accurate description of today’s market while paying too little attention to behaviours that are beginning to challenge it.
Qualitative research allows us to explore these signals without assuming that they are representative. We can understand where they come from, analyse what differentiates them from mainstream behaviour and, if they prove relevant, subsequently use quantitative research to determine how widespread they are within the market.
Where does artificial intelligence fit in?
Artificial intelligence is dramatically expanding our ability to work with information in market research. Today, we can use it to analyse large volumes of interviews, process open-ended responses, identify recurring themes, challenge hypotheses or synthesise documentation in a fraction of the time some of these processes required only a few years ago.
For those of us working in research, these tools offer exciting possibilities. But their adoption also makes it particularly important to distinguish between identifying a pattern and determining what that pattern means in relation to the problem we are trying to solve.
If we use AI only to identify what occurs most frequently, we risk reinforcing a tendency that existed in research long before generative models arrived: confusing frequency with relevance. An exceptional response, a case that does not fit our segmentation, or an apparently minor contradiction may contain important information precisely because it challenges the dominant interpretation.
This does not necessarily create an opposition between artificial intelligence and research with real people. The opportunity lies in using new tools to expand our analytical capabilities without losing sight of the context, exceptions and differences that can help us build a more accurate interpretation.
From pattern to insight
A pattern allows us to observe that certain behaviours, opinions or needs appear repeatedly. An insight requires an additional step: building an explanation of how those elements relate to one another and what that relationship means for the problem we are investigating.
Within this process, qualitative and quantitative research can play complementary roles. Quantitative research allows us to understand the scale of a phenomenon, how it is distributed and how different groups compare. Qualitative research can help us better understand the experiences, needs and mechanisms behind those results, while generating hypotheses that can subsequently be quantified.
Understand before simplifying
Reducing complexity is inevitable. A research report cannot reproduce the experiences and decisions of hundreds of consumers in their entirety, nor would there be much value in doing so. Our job is precisely to transform that complexity into knowledge that can be used to make decisions.
The relevant question is when we simplify. If we group people too early, we may end up building segments around individuals who resemble one another only superficially, treating responses with different meanings as equivalent, or dismissing as exceptions behaviours that deserved further investigation.
One of the most valuable contributions of qualitative research occurs precisely before that simplification takes place. It allows us to separate apparently similar responses, identify connections between different behaviours, and understand which dimensions are genuinely relevant to explaining a decision.
From there, we can measure them, compare them and determine their prevalence within the market. But to do that properly, we first need to understand what is worth measuring and what the things we are measuring actually mean.
