Consulting against Culture: A Politicized Approach to Segmentation

Usefulness of Segmentation Models

Ideally, a segmentation model is a heuristic for efficient decision-making. Lawrence Gibson (2001) describes the process by which data is shaped into labels: “The creative analyst reviews the complex data patterns and simplifies them with rich names for the segments, such as ‘striver,’ ‘inner-directed,’ ‘foot soldier,’ or ‘socially conscious.’ ” Nation-wide segmentations like VALS and PRIZM market answers on demographics and lifestyle variables in the form of “Survivors” and “Big City Blues.” Because these artifacts reduce the complexity of the market to a manageable set of target segments, they become the main way through which companies interact with the external world of the customer.

Segments can become social facts¹ because they address a variety of organizational needs: responding to the heterogeneity of the market, allocating resources, and developing marketing strategy as well as new products. In marketing literature, segmentations are a response to the fundamental heterogeneity of the market (Wind and Bell 2007). They are assumed to identify needs and wants that have a meaningful relationship with product and purchasing preferences. Cross-cultural research also describes how these models are used to manage complexity (Furrer, Liu, and Sudharshan 2000, Chang and Su 2012). ² Secondly, segments are used to streamline decision-making. A key segment provides a shortcut for micro-level decisions such as packaging architecture and the prioritization of information in marketing. One client described how retailers are used to simplify action in a similar fashion: in response to questions on positioning, “Someone will say, ‘Well, this is a Walmart product, and everyone else is, ‘Oh, ok.’ ” Through well-intentioned efforts to stay true to the customer, these representations have disproportionate impact on strategy as they influence internal perceptions about what people like, want, or will buy.

Limits of Segmentation Models

Limitations emerge from the very usefulness of these models. First, segmentations are seen as an essential part of defining strategy, but workers and teams struggle in how to apply these descriptive documents to predictions about which products will succeed. Their size and formal tone also makes it difficult for decision-makers to connect with them emotionally and intellectually. This is part of the reason why companies turn to consultancies or internal innovation groups to “explore,” “build an understanding,” or “create a context” for segmentation.³ Segmentations also create disadvantages when they become static. The time period during which the segment was authored is treated as eternally true and accurate in regards to the individuals that it describes. Correctives to traditional segmentation have already found a foothold in the business world through consumer culture theory (CCT). CCT is inspired by semiotics and practice theory, in particular Pierre Bourdieu’s work on social and cultural capital. In contrast to what they describe as the “skeletal descriptions” of traditional approaches, proponents argue that products and brands have multiple meanings to consumers based on their social and class position (Ahuvia, Carroll and Yang 2006:35). Segmentations are representations by necessity, but their descriptive and static traits can put companies at a disadvantage when left unquestioned.

Segmentation models tacitly discourage challenges because they take a lot of time and investment to produce. This makes it all the more disturbing when the relationship between a segment and the users it is supposed to represent seems to break down. During a Microsoft project, Flynn et al. write, “Within days in the field, we knew something was terribly wrong – at least half of our participants did not fit into the segments as we had understood them” (2009:84). Later, they discovered that while the names of all the segments remained the same, the cluster analysis their vendor used no longer included usage variables. The experience of these researchers helps challenge the near-sacredness of segments and encourages teams to investigate their origin, in order to avoid errors in strategy that come from relying on the “black box” of the segmentation algorithm. The expense of a segmentation also means it is not constantly updated, so those that originally managed it may have since moved on. One client shared that most people only know a few bits and pieces, or “the general gist of it.” Despite the incomplete knowledge of those expected to work with the market segmentation, the organization as a whole maintains an interest in its legitimacy.

Finally, segments take on a life of their own as they are used and circulated. Lawrence Gibson (2001) believes that these constructs can unproductively push clients away from the data: “The denotative and connotative meanings of the labels will be scrutinized… the labels, not the data, will drive marketing insights and strategies.” The tendency to generalize also subsumes any variation among individuals perceived as being part of a certain segment. This happens in the field, where a reaction like “That’s not our customer!” is framed in the terms of the segment rather than the individual experience. Similarly, Arnould and Cayla (2013) see a fetishizing of the consumer in statements like “We saw three Marys today.” Segmentations create a sense of rightness, comfort, and common sense the more they are self-referenced. Although familiarity provides a common ground of shared beliefs, the danger is in effacing that the ground, or data, is shifting and complex.

Theories of Culture and Representation

The first direction draws on critiques of how ethnography not only represents, but also is itself implicated in creating cultures through written description. This approach is based on the work of anthropologists like Lila Abu-Lughod, James Clifford and members of the Writing Culture group. The reifying, limiting, and isolating effects of the concept of culture have direct application to segmentation. Uncritically viewed, cultures and segments are taken to be heterogeneous in attitudes and behaviors, assign differences in areas such as economic decisions and gender norms to cultural values, view the culture or segment as individual “personality writ large”⁵ and rely on an objective, professional, and nearly anonymous expert presence in creating the representing text – the ethnographic monologue, final presentation, and segmentation algorithm.

In Writing Culture, rhetoric isn’t just a matter of effective communication but a political act through which the ethnographer chooses one particular definition of a culture, eliminating other possibilities and perspectives. In response, Abu-Lughod argues that the neglect of feminist and “halfie” or indigenous anthropologists is a lost opportunity to address how “cultural difference is a problematic construction” (1991:163). Members of the discipline retain an identity as anthropologists by studying the other through the concept of culture (154). She suggests a number of tactics to counter the reifying power of cultural representations. First, the concept of practice:⁶

[Practice theory’s] theoretical approach is built around problems of contradiction, misunderstanding, and misrecognition, and favors strategies, interests, and improvisations over the more static and homogenizing cultural tropes of rules, models, and texts, (1991:159).

Abu-Lughod also suggests that anthropologists should avoid the false neutrality of generalizing through ethnographies of the particular (160). Introducing the language of everyday life into a professional text creates risk by potentially calling one’s own expert identity into question, but ideally creates a document, or a model, that is more dynamic in its representation of group attitudes, values, and behaviors.

Convenience shopping for a multinational energy company – As the company was planning to extend their brand of convenience stores across the U.S., they commissioned an update to their existing multinational segmentation. The study was conducted in two phases, one-on-one and group interviews with consumers in six countries in order to help build the list of consumer attributes for quantitative testing, and a second phase of online questionnaires conducted in 15 countries focusing on attitudes, behaviors, values and needs. In the U.S., approximately 3,500 questionnaires were administered. Five segments were identified, and the two in highest value became the focus of our project – these were individuals described as stopping at convenience stores to buy multiple items at an affordable price, as well as those who shopped there at least once a day for coffee, food, and other items. Although these were considered two distinct segments, the descriptions of behaviors and attitudes made it hard to differentiate which was which. For example, consider that for one group, “Gas stations offer food 24 hours a day, appealing breakfast items, fresh bakery items, and hot food that is available immediately,” versus for another, “Coffee is the key draw. It is quick, easy, and affordable… They want food to be available 24 hours a day.” As part of the expansion, Conifer was hired to look into on-the-go purchases and behaviors for individuals who typed into these segments.

This project illustrates how theories of representation can inspire more dynamic models of user behavior. One of the advantages of an ethnographic approach is that the methodology instills a focus on the thoughts and actions of particular participants. Instead of generalizing about the behavior of a segment, we discuss individuals and how their experiences compare and contrast in order to better understand who has typed into that group. This project also demonstrates a commitment to practice over segment – how people define, contest, and live what might be typically codified as variables in very different ways. Finally, identifying instances when individuals who type into one segment act like those in another draws attention to the constructed character of segmentations as models.

Abu-Lughod argues for ethnographies of the particular over those that generalize about an entire nation or people.⁷ Her argument also applies to the artificial quality of market segmentations and the lived experience these documents subsume in their expert language.

Showing the actual circumstances and detailed histories of individuals and their relationships would suggest that such particulars, which are always present (as we know from our own personal experiences) are also always crucial to the constitution of experience, (1991:162).

For example, while “family” in the existing segmentation was equally important to both types, this seemingly universal value played out differently in purchasing behaviors. One segment was more family-driven and the other more individual when making on-the-go purchases. Individuals within a single segment also shopped differently even though they were supposed to prefer a similar strategy. We identified two sub-types – a set of people who preferred one-stop shopping as the quantitative segmentation identified, and another who “trip-chained,” or made multiple stops on the same trip – creating an awareness of a different mode of operation through which the client could gain share. In a sense, segments of the particular can create a conceptual space for differing perceptions of values and behaviors believed to be central and unitary.

In a second paradigmatic shift Abu-Lughod and others prioritize practice over culture as a driving force of analysis in order to avoid the false coherence of the latter. Focusing on practice over the unifying segment descriptions allowed us to recognize what Abu-Lughod describes as the “multiple, shifting, and competing statements” that are part of group life (159). One such instance was the relationship between food freshness, value, and quality. For participants identified with one segment, these attributes were intertwined, while the others thought of them as discrete entities. Many participants described themselves as attune to trends, but one group defined these as culinary and diet related, and the other to seasonal and product trends like the Pumpkin Spice Latte. Differing interpretations of what might be considered the same variable or motivator highlight the fallacy of thinking of segments as the same set of variables, just in different amounts – the same approach to “fresh food” or “on trend” would not account for a supposedly shared interest among segments.

These directions call attention to the heuristic nature of segmentation – artificial and abstracted even if necessarily so. Tracing contrasts and similarities reminded us that these individuals are not isolated, but rather live and consume in a world where they interact with others classified into a separate segment. The translation of reality to representation was also central in moments when we observed participants who typed into different categories behave in a similar fashion. In the morning, people acted most alike, as well as when they stopped for a physical “boost.” For necessities, they used in-and-out attributes to evaluate a potential stop. User language to rationalize purchases, such as a “pick-me-up,” “treat,” or “escape,” was similar. This suggested merchandising and designing spaces that spoke to customers regardless of their segment affiliation. Later on, we were also asked to help plan workshops for each of the product lines – bolstering socialization of the segmentation as well as becoming part of creating the power and fetish of “bringing a segment to life,” (Arnould and Cayla 2013:395). We became implicated in reproducing the segmentation, but in a way that at least made people account for its constructed quality.

Theories of Consumption and Identity

Bourdieu (1977) offers another critique of representations of the social world that are not realities in themselves, yet are treated as such. He deconstructs the oft-cited metaphor of culture as a map: “It is the analogy which occurs to an outsider who has to find his way around in a foreign landscape and who compensates for his lack of practical mastery, the prerogative of the native, by the use of a model of all possible routes” (1977:2). However, there is gulf between this abstraction and the actual journey, just as there is between a market segmentation and everyday practice.

Bourdieu’s formulation of cultural, social, and economic capital is also relevant for our efforts. He argues that practices like language, education, and consumption are not neutral, but implicated in reproducing the interests, influence, and power of one group or individual over another. He explains, “It is in fact impossible to account for the structure and functioning of the social world unless one reintroduces capital in all its forms and not solely in the one form recognized by economic theory,” (1986:241). Cultural capital, in particular, can take multiple forms through personal dispositions,⁸ cultural goods, and institutionalized means like educational credentials. As described above, his critiques have been taken up by consumer culture theory, but have further potential for the political nature of consumption.

The politics of everyday practice are also at the core of Michel de Certeau’s The Practice of Everyday Life. De Certeau redefines consumption as another form of production, albeit silent and dispersed. Theory must grasp how products are utilized, since statistics accounts for the material aspect of everyday practice, “but not their form; it determines the elements used, but not the ‘phrasing’ produced by the bricolage (artisan-like inventiveness) and the discursiveness that combine these elements,” (1988:xviii). His work suggests that we cannot consider consumption as a one-directional narrative, but recognize it as a political act that has implications for how we position brands and products.

Homemade cooking for a food product manufacturer– This case study addresses a food and beverage company that used multiple segmentations at the category and brand level. To coordinate strategy, they sought out an overarching cross-category model. Their stated goal was a holistic understanding of food behaviors and beyond, including key attitudes, beliefs, values, and motivations. To create this segmentation, a sample of 5000 individuals was surveyed through a one-hour online panel. Seven segments emerged; the most valuable segment was described as the highest income, the most interested in homemade cooking, purchasing multiple in-category products, and also spent the most on grocery shopping by over one hundred dollars per month. Despite the apparent fit, offerings targeted to help people more easily prepare homemade meals – sauces and condiments – were not doing well. We were asked to look into how these users’ needs connected with one brand in particular, as well as to come up with a set of design principles for product development.

Further motivations for a political approach to representation emerge from the role of purchasing and usage behaviors in this project. Reconnecting with usage, rather than just the purchase moment, demonstrated how consumption was a choice intimately linked to identity, class, and power for these individuals. The practices we observed and discussed – shopping, deciding on a meal, cooking, eating, and talking about food – instilled meaning to consumption that had been missing from the segmentation. Additionally, dealing with food, cooking, and related knowledge as cultural capital provided an entry to new marketing and product development opportunities for our client.

Returning to Bourdieu’s contrast of the map versus the journey, the description of the target customer had become hard to apply to strategic decision-making because it did not connect with how people that typed into a segment experienced the meaning and role of food in day-to-day life. For example, our client believed they enjoyed cooking homemade food, but the particular meaning of homemade was not used to position their products – their “just add chicken” approach was falling flat. Participants defined homemade cooking as the result of transforming food with knowledge, effort, and affection, rather than assembling components of a meal on a plate. Craftsmanship was also important, because for them, a homemade meal was intentionally thoughtful in its variety of texture and flavor. This led them to avoid the perceived uniformity of more processed food whenever possible. Without reflecting their individual practices through the physical qualities of the product, the packaging, and the positioning, the company limited its own success with the individuals they had identified as their most valuable segment.

Purchasing and use behaviors begin linking to the exercise of social and cultural capital, adding context to how products are used to perform individual and group identity. Like the original description of the segment, we generated a long list of products these cooks had purchased. Unlike the online questionnaire, we were able to observe them select and use these products in person,⁹ highlighting the polysemic meaning of products and brands as inspired by Bourdieu and de Certeau. These participants took pride in the cultural experiences and media consumption their class position and what they perceived as their personal interests engendered. Their culinary knowledge represented a form of cultural capital, so it was no surprise that they did not respond to products that did not reflect or respect this knowledge. Purchasing and usage behaviors further illustrated this insight. Participants gravitated toward prepped items that saved labor and time, but not processed as to remove demonstrations of their skill in interacting with ingredients. Cultural capital reframes the problem of homemade cooking from one of time and availability to one of performance and identity.

The Client Context and a Reflexive Approach

Reflexivity refers to the inclusion of the role of the researcher and the methodology as project data. As Writing Culture and Abu-Lughod suggest, it is an approach that counters the reifying effects of abstractions like culture and segments by situating them in their historical and political context. In the corporate world, this tactic can be repurposed through analytical attention to the research and client context. Clients should be considered as participants and co-creators of project deliverables, most evidently because the firm’s expertise will impact which customers will prove most valuable to them. But this is also important because the socialization of a model is just as relevant to its success as its form or content. Since we work with a variety of groups with different goals, such as marketing, user experience, innovation, and consumer insight, responding to the group and corporate context can address anxieties and assumptions through the structure of the research outcomes.

As described above, ethnography is constrained when framed as an addendum to segmentation, but a reflexive approach provides opportunities for teams to dissect the tacit assumptions underlying representations. For example, Flynn et al. (2009) assumed they all shared the same idea of who their target segments were – it was only when they seemed to have issues with recruiting that the team looked into how the segmentation algorithm eliminated current usage variables. New team members or situations create moments for questions that people may usually be reluctant to ask since they assume the answer to be generally understood and shared. Witnessing the research interaction can itself be an impactful moment. The principle of reflexivity suggests periodic forays into the field to reconnect with the practices a segment is supposed to represent.

Deliverables and Representative Practices

Ideally, our work becomes a part of producing belief and action within companies. In order to address the issues of segmentation as the dominant genre of representation, we can use a variety of models that champion the dynamic nature of user behavior. In a project about media consumption, we identified modes in which content was primary or secondary to the social interaction, along with which household members were involved in selecting the content. Media usage behaviors informed the design of new interfaces to support people as they moved from mode to mode. Tracing the connections between behavioral or user types and how individuals move between categories can help companies innovate around platforms or systems that consider the entire lifecycle of the user. In the convenience shopping case study described above, we indicated situations when individuals in one segment would begin to act like another. In contrast to the original segmentation, this allowed us to ideate concepts with an awareness of the dialectic relationship between segments. No single framework is appropriate to every design challenge, but a critical stance towards the boundaries, connections, and migrations between categories of a model can uncover opportunities for innovation around change and contrast.

A second evolution of representative practices is based on the choices we make in portraying everyday life. As we know, rhetoric and texts are not neutral, but assert assumptions regarding the subject at hand. For example, one possible choice is using photos of study participants rather than stock photographs, as well as video whenever possible. This approach is also limited in how it creates the consumer fetish, but it reminds the reader that the typology is supposed to be connected to behaviors in the real world.¹² We can also whenever possible state insights and category descriptions in user language. Using the internal logic of emic categories can also inspire the factors on which the typology is based. Representative choices of models and how they are depicted can introduce small, incremental changes in how companies connect typologies to real world practice.