Does more time equal more insights?

Editor’s note: Terry Grapentine is principal of Grapentine Associates, an Ankeny, Iowa, research firm.

Much debate swirls around the proper use of focus groups, compared to other exploratory research modalities such as individual depth interviews (IDIs), mini-groups, ethnography and so on. Clearly, research objectives should dictate the proper method or mix of methods, and, often, multiple approaches provide the keenest insight. However, if the primary purpose of your research is to identify and understand the drivers of individual purchasing behavior, IDIs - either at a central location, in-store or as part of an ethnographic study - not focus groups should be your primary method.

In my experience, interviewing consumers about the purchase of a product from the time a need state was felt to the time a product was purchased can take from 30 to 45 minutes. This is a worthwhile endeavor because of a central characteristic of causality: if A causes B, then A has to occur before B. Therefore a proper investigation of cause-and-effect requires the respondent to articulate the sequence of events that led up to the purchase. Moreover, the researcher has to take into account all of the confounding circumstances and factors that can hide these cause-and-effect relationships, a topic I will discuss shortly.

The focus group neither allows enough time for this nor is it a proper venue for this kind of investigation. For instance, assume the following: 1) an average focus group lasts approximately 100 minutes; 2) there are 10 respondents in the room; and 3) 10 minutes are taken up by moderator and respondent introductions and the moderator setting the ground rules for the discussion. This leaves 90 minutes for actual discussion or about nine minutes per respondent vs. 30-45 minutes in a typical IDI.

The focus group format does not lend itself to an individual respondent telling a story about how she came to purchase a product. The nature of a focus group is for one person to talk for a few minutes; then another, and so on. In other words, for a given topic, the moderator wants to promote a dynamic discussion among a group of individuals for a given topic and then move on to the next topic. And, as we all know, moderator guides can be quite long!

A single respondent cannot weave his story uninterrupted in a focus group setting, thereby giving the moderator a sense of how events unfolded over time. And it is this recounting of events and their relationships over time that helps us uncover causality. Of course there are other problems with focus groups such as group-think, respondents giving socially acceptable answers to questions or respondents giving answers that are more rational than emotional.

The causality conundrum

A myriad of confounding circumstances can hide the nature of cause-and-effect relationships, making them difficult to untangle in a focus group. Chief among these are INUS conditions, spurious relationships, attribute importance vs. determinance and the relative roles attributes play in the decision process. When one considers all these factors surrounding a product purchase, one can readily appreciate the advantage of IDIs over focus groups in deciphering the causality conundrum.

The INUS condition. This acronym stands for “insufficient-necessary-unnecessary-sufficient.” This means conditions that are insufficient and necessary as well as those that are sufficient but unnecessary are separately capable of bringing about an event. For example, the following conditions - defined as a collection of events or attributes possessed by a product - may precipitate the purchase of X:

Condition 1:
Events A, B, C, D, E = X is purchased

Condition 2:
Events A, D, E, F, G = X is purchased

Condition 3:
Events A, F, G, I , K = X is purchased

None of the events separately brings about the purchase of X, even though one event is necessary - in this instance, A. In other words, there are multiple routes to purchasing Product X - several conditions may be collectively sufficient, even though no one event of the set is singly sufficient.

One illustration of the above example relates to understanding a banking customer. For most bank customers, A is location convenience - typically a location close to home, work or shopping. However, location convenience is often an insufficient condition to drive bank selection. Other attributes of the bank will influence the prospective banking customer such as service variety, community reputation and price competitiveness. Different groups of bank attributes - or conditions, in the above example - reflect different market segments and it is those unique combinations of attributes that describe a market segment, in addition to Attribute A, and drive bank choice.

But perceiving the existence of relevant INUS conditions in a given market will likely be more difficult in a focus group setting vs. a series of IDIs. This is because of the following shortcomings of focus groups: 1) not all respondents express their views on all topics; 2) the tendency for some people in social settings to agree with the group; and/or 3) there is insufficient time in the group session to explore all attributes or events. IDIs do not have these drawbacks.

Spurious relationships. Some relationships are coincidental, or spurious, rather than causal. IDIs can help differentiate between the two. Consider Figure 1, in which the purchase of a product, P, and V1 are both caused by V2. Further, V1 is correlated with but does not cause P, thus the dotted line from V1 to P.

In an example using dry dog food, product quality, V2, is a determinant of brand purchase, P. Product quality is also a determinant of package quality, V1, because manufacturers use higher-quality packaging for higher-quality dry dog foods. However, product quality is the driver of brand purchase, not the perceived quality of the food’s packaging (given that the packaging quality is within an acceptable range), based on this researcher’s category experience.

Admittedly, the above example might easily be investigated in a focus group; however, some product categories are not so transparent. For example, in my experience, separating a spurious from causal relationship between a) the torque vs. horsepower of an outboard engine and b) product preference is difficult and can only be understood by talking to an individual boat-owner respondent for several minutes about this topic - an undertaking that would take too long in a focus group.

Attribute importance vs. determinance. As social scientists, we differentiate between an attribute that is valued - that is, it is important - versus an attribute that is both valued and determines choice a determinant attribute. Respondents, however, do not often make this distinction. They do not differentiate between the two concepts when asked questions such as, “What was important in your selection of X?” or “What influenced you to purchased X?”

Consequently, the moderator has to ask a series of questions to uncover the true causes of brand choice. Consider an example from a focus group on the purchase of lawn mowers. Respondents reported that a trigger for purchasing a new lawn mower was the increasing difficulty in starting their old lawn mower. Thus, the attribute “easy start” was an important product attribute guiding their most recent lawn mower purchase. Upon additional discussion, however, the moderator learned that the “easy start” attribute did not differentiate brands in these respondents’ consideration sets; therefore, ease of starting did not influence brand choice. Other factors such as ease of lawn mower maneuverability or price played a more influential role.

True, this insight was discovered in a focus group. A problem arises, however, when there are many such issues surrounding a purchase decision and, because of time limits, to parse each issue with respect to what is important from what is determinant becomes problematic.

Relative roles of attributes. A central focus of research is to understand the relative role individual attributes play in the decision process; however, understanding the decision process itself is more complicated than simply understanding the relative role particular attributes play in it. You need to understand the way in which consumers use attribute information to make a decision. In this process there are two general models to keep in mind: compensatory and non-compensatory models (adapted from Consumer Behavior: A Perspective by John C. Mowen and Michael Minor).

Compensatory models: Low perceived performance on one attribute can be compensated by high performance on another. For example, a certain dog food may be difficult to purchase because of limited distribution (you have to go to your veterinarian vs. buying it at a grocery store) but this negative is compensated for by the product’s perceived quality.

Non-compensatory models: High ratings on some attributes will not make up for low ratings on others. This is often found with a low-involvement product. For example, having an attractive package may not allow a grocery store to charge more for simple table salt.

To optimize your perspective of the various decision processes that consumers use to purchase products, one needs to have the respondent recount the process s/he went through from the time the need arose to the time the purchase was made. In addition, the moderator needs to probe to uncover other issues discussed previously. For example, do consumers segment into different INUS conditions? Which relationships uncovered in the interview are merely spurious versus reflecting a cause-and-effect relationship? What factors are merely important versus determinant? And, how is all this information processed by consumers? This last issue is of particular importance when designing a follow-up quantitative study. For example, you don’t want to conduct a conjoint study if the attributes consumers consider when making a purchase are not compensatory in nature.

Understand the true protagonists

In summary, identifying causal relationships in consumer behavior is complicated. IDIs, not focus groups, offer a more appropriate approach in which to investigate these issues. Moreover, to the extent we are familiar with how spurious or hidden relationships can mask true causal ones, the better equipped we are to identify and understand the true protagonists motivating consumers toward their purchases.