How they really think

Editor’s note: Wilson Readinger is research associate of Klein Associates Inc., a Fairborn, Ohio, research firm.

“Most of the time, when I need to go shopping for a new car, I start with the dealer where I bought my current car. I sometimes spend a day or two at the lot, just looking things over - trying to find something that I like. Mostly, I prefer front-wheel-drive, six-cylinder coupes, and there are some options that I look for, like a CD changer and a car alarm. When I find something that I like, I might then compare the price with the invoice figures I can find online for that model and talk to a salesman. I think I’m a really well-prepared consumer, and it’s hard to sell me on something I don’t want to buy; I don’t pay too much attention to advertising. If the salesman can give me a price that’s within my budget - which is about $240 a month - and it’s the car that I want, then I’ll probably buy it...”

This is an excerpt you may hear in a focus group, and there are certainly some measurable and useful pieces of information contained within it, but when your goal is to truly understand the consumer’s decision-making process, these data could be virtually useless. This consumer actually paid $310 a month for an eight-cylinder SUV that was purchased 90 miles from his home from a dealer he never visited before the time of purchase. There are two important issues here:

1) When a participant is speaking in generalities or making broad statements, it is easy to miss the reality of the situation; that is, what the participant actually did.

2) In most domains, there is a need to better understand the decisions that consumers are making and how these decisions are actually being made.

Now consider the following variation:

“The last time I went to buy a car, I started with my current dealer. I was looking for a coupe, probably a V6, with a CD changer and a car alarm. When I was on the lot, though, I saw another model and I remembered that it had been on the news because of a recall for some problem with the steering system. I think they make all those cars on the same line in Japan, though, so I wondered if maybe several models would have a similar problem. Anyway, a friend of mine mentioned that he got really good service at this other dealer, so I took the drive and when I showed up there, I just had a feeling that the salesman was being honest with me. There was something about the look of the SUV that I bought that seemed right to me when I took the test drive, so I wound up paying quite a bit more than I really wanted to, but I’m very happy with the SUV so far...”

Clearly, there are more feature-and-benefit bullet points that can be taken from the first paragraph. However, the second sample contains the framework for a potentially high-payoff interview that can address some of the fundamental issues of consumer decision-making for companies seeking insight into how products are viewed by their customers. The first paragraph contains the easy answers that many survey respondents would be likely to give; the second provides the building blocks of context and the beginning of an incident, or lived experience in the participant’s life, that can be used to ground several hours of a cognitive interview (a structured inquiry into the decision making of a consumer). Within the context of an actual story, the participant cannot hide behind generalizations or tendencies; both interviewer and participant have a common reference point upon which to anchor further discussions of the cognition and decisions that underlie the consumer’s behavior.

Most qualitative research techniques are designed to answer the question: What are my client’s customers thinking with regard to the product in question? Many research practitioners have tools available that can answer this question with vigor. A subtly different question, though, has some very different implications: How are my client’s customers thinking? Addressing that problem effectively requires a different set of tools, a significant investment into mental modeling (uncharted psychological territory for many companies that use primarily focus groups and surveys to collect data), and a different set of expectations regarding the outcome of a research project. However, when executed effectively, the result is a depth of understanding of the consumer that is richer than the client has ever experienced. This new view of the consumer’s thinking and decision-making is a firm foundation upon which to build future product design, marketing and research.

The power of stories

What does it mean for a research method to be incident-based? In short, incident-based methods of qualitative data collection place great emphasis on the actual, lived experience of the consumer. This collection of techniques emphasizes the importance of being in a particular situation with a product (whether that is the purchase situation, the use situation or some other situation that is relevant to the client’s research question) and recounting this incident in an interview with skilled, trained researchers. Some have argued that the story is the fundamental unit of human memory, and that the structure and content of stories essentially define culture and exert a huge influence on behavior, opinions and decisions. Incident-based methods are consistent with that view, and take advantage of a consumer’s own experience and expertise in making decisions about products. Simply put, an incident-based approach to consumer decision-making retains all the context of the original decision, whether it is the context of time, stress, emotion, uncertainty, the environment, or other factors that we may not even realize are affecting a decision until this interview approach is used. Speculation on the part of the participant is actively avoided, and the interview is, to the fullest extent possible, grounded in actual events.

Part of the uniqueness of this type of approach is that the interviewers are - in most cases - not collecting opinions. In fact, the data that participants provide are usually cold, hard facts. Training in incident-based techniques teaches the researcher to steer clear of responses that take the form of “Usually, X happens,” “Some people believe that X,” or “Generally, my opinion on X is...” Rather, incident-based techniques rely on the details of the situation in question. Often, questions get at the progression of actual events during purchase and the behavioral and cognitive tasks associated with product use. These elements are always elicited with an ear for key decisions associated with the product. The interview participant himself gives clues to where the critical decision-making is taking place. Phrases like “I just knew,” “There was something about it that made me think...,” and “I had a gut feeling” indicate that the participant’s intuition was called upon and that this is therefore a potential decision point.

How many times have you, as a researcher, heard these sorts of statements? They are common, and they should catch your attention. When an individual has been in a particular situation many times, it is not unusual to have some extra insight into how that situation might unfold. This is true of large-scale, life-or-death, critical decisions, such as those faced by firefighters and F-16 pilots, but also true of everyday decisions made by consumers. What is often referred to as intuition or gut instincts about a product may actually be a reflection of the expertise of the consumer - their experience and mental models of a product unconsciously influencing their behavior. Some incident-based techniques are designed to “unpack” these intuition statements that consumers make all the time, and can help researchers to understand where those gut instincts are coming from and how they can be influenced.

Mental models are the key to understanding how consumers make decisions, and are therefore the key to providing product manufacturers with the insight that they are seeking with regard to their customers. An understanding of consumers’ mental models is one outcome of the careful analysis of the stories that are elicited through incident-based techniques. A mental model is really an internal (i.e., residing in the mind of the consumer) representation of the external world; it is the consumer’s understanding of how something works. For instance, why choose satellite television over cable? The answer will probably depend on several factors (including cost and availability of the services). While all of these are potentially important, the critical factor may be the consumer’s understanding of how these services work: Where does the satellite signal come from? What does my television do with the signals? What types of maintenance are required and why? What happens when the service breaks down or something goes wrong? Investigating actual incidents and exploring the decisions behind critical points can paint a picture of the participant’s mental model of the service. This, along with other information taken from the interview, can tell you not just what the participant thinks about satellite TV, to continue the example, but how he thinks about satellite TV when he makes these decisions.

The consumer as expert

One effective technique for a cognitive interview is a knowledge audit. This technique was developed initially to elicit and categorize the major cognitive differences between experts and novices within several categories of military occupations in order to develop training programs that are more efficient in getting novices up to speed. When applied to marketing research, though, the technique excels at exposing the fundamental reasons for purchase, the decision-making associated with purchase and use, changes in the consumer’s mental model of how a product works and how to use it, and brand loyalty.

The knowledge audit is organized around knowledge categories that have been found to characterize expertise. Some of these are contained in Table 1.

The knowledge audit employs a set of probes designed to describe types of domain knowledge or skill and elicit appropriate examples in the form of stories or accounts of incidents. The goal is not simply to find out whether each skill is present in the task, but to find out the nature of these abilities, specific situations in which they were required, strategies that have been used by the consumer, and so forth. The list of probes is the starting point for conducting these interviews. Then, the interviewer asks for specifics about the example in terms of critical cues and strategies of decision-making. This is followed by a discussion of potential errors that a novice or less-experienced person might have made in this situation. Figure 1 represents the process of a knowledge audit.

Cognitive interviews last approximately two hours and require two skilled interviewers for each consumer who participates in the research. The two-on-one interview format allows one interviewer to engage the participant and lead the interview while the other takes notes, maintains awareness of the research questions, and adds questions or explores avenues that the lead interviewer might have missed. This approach is time- and resource-intensive, and a team of interviewers working a 10-hour day can realistically do no more than four interviews. Some clients are nervous about the prospect of relying on a sample of 15-20 (and sometimes many fewer) interviews to inform their expensive marketing and design efforts. While history, and some empirical evidence, shows that the method consistently delivers actionable insights that are valid and reliable, there are some additional advantages.

One can be seen in the design of new products or new features in the future. The understanding of consumer decision-making that is gained from incident-based research now can pay dividends in the future when it comes time to test new ideas. Knowing that one segment of consumers dislikes the styling features of an automobile is only directly useful for that particular model. However, knowing that consumers from that segment tend to think of this model of car as a tool for getting work done - and therefore find racing stripes inconsistent with a serious tool - can definitely inform the styling of any new products that will be marketed as tools. Similarly, some alternative qualitative strategies would struggle to elicit information about the relative importance of, for example, point-of-sale advertising versus direct mail for a particular product. If this distinction is important to consumers, it will be a part of the incidents they recount, and it will be a part of their mental models of the purchase of this item.

When it comes to purchasing and using products, all consumers are experts at their own decision-making. Why not take advantage of this experience and expertise, and preserve the richness of the context in which these decisions were made and these behaviors initiated? What is lost in statistical power and breadth of sampling is gained in the depth of understanding and clarity of rationale that can result from a well-executed incident-based interview. “What does the consumer think of this product or service?” is a question that needs to be answered over and over again, but answering the question of “How does the consumer think about this particular product or service?” can produce deep and lasting insights.