Constellations, Configurations and Consumption: Exploring Patterns of Consumer Behaviour Amongst U.K. Shoppers

Margaret K. Hogg, Manchester School of Management, UMIST, UK
Paul C.N.Michell, Manchester Business School, UK
ABSTRACT - This paper extends existing U.S. work on product and consumption constellations and examines the formation of these intermediate patterns of joint consumption amongst U.K. consumers within the framework of symbolic interactionist approaches to understanding self. Following the conceptualization of the forces which influence the formation of consumption constellations and anti constellations, consumption configurations are proposed and a series of models are developed and tested via the application of correspondence analysis to data on shoppers drawn from a large U.K. commercial database. It is argued that retail outlets represent an important service category and that retail constellations can be used alongside other major product constellations such as cars, clothing and electronic goods (identified in earlier research) in exploring intermediate patterns of joint onsumption.
[ to cite ]:
Margaret K. Hogg and Paul C.N.Michell (1997) ,"Constellations, Configurations and Consumption: Exploring Patterns of Consumer Behaviour Amongst U.K. Shoppers", in NA - Advances in Consumer Research Volume 24, eds. Merrie Brucks and Deborah J. MacInnis, Provo, UT : Association for Consumer Research, Pages: 551-558.

Advances in Consumer Research Volume 24, 1997      Pages 551-558


Margaret K. Hogg, Manchester School of Management, UMIST, UK

Paul C.N.Michell, Manchester Business School, UK

[The authors would like to acknowledge the generous support of the British Market Research Bureau/Target Group Index, Ealing, London, for this study.]


This paper extends existing U.S. work on product and consumption constellations and examines the formation of these intermediate patterns of joint consumption amongst U.K. consumers within the framework of symbolic interactionist approaches to understanding self. Following the conceptualization of the forces which influence the formation of consumption constellations and anti constellations, consumption configurations are proposed and a series of models are developed and tested via the application of correspondence analysis to data on shoppers drawn from a large U.K. commercial database. It is argued that retail outlets represent an important service category and that retail constellations can be used alongside other major product constellations such as cars, clothing and electronic goods (identified in earlier research) in exploring intermediate patterns of joint onsumption.


In examining intermediate levels of joint consumption this research lies between studies of consumer behaviour which concentrate on micro issues (i.e. decision making at the product or brand level) and those studies which concentrate on macro issues (such as the factors which influence societal patterns of consumption, Firat 1978). The important argument for this study is that shopping channels, as well as products, are used by consumers to create and interpret cultural meaning; and that patronage choices are inherently part of the process by which consumers construct and constitute their world.

The aggregation of choices across products and services amongst consumers has been explored in a series of U.S. studies (Levy 1964, Kernan and Sommers 1967, Wells 1968, Green, Wind and Jain 1972, Solomon 1983, Solomon and Assael 1987, Kehret-Ward 1987, Solomon 1988, Solomon and Buchanan (1991)). Solomon and Assael (1987) and Solomon (1988) developed role-related consumption constellations: "a cluster of complementary products, specific brands, and/or consumption activities associated with a social role" (Solomon and Assael 1987:191) in examining patterns of joint consumption. This study will argue that constellations and anti constellations together can be taken to constitute consumption configurations which represent a series of joint consumer choices and non choices amongst products and services which can be associated with different groups of U.K. shoppers. Underlying the model building is the assumption that product and consumption constellations and anti constellations can be linked with the creation, maintenance and enhancement of social identities.


The framework for the model building of the forces which influence intermediate patterns of joint consumption was the social psychological perspective of symbolic interactionism (Cooley 1902, Mead 1934, Solomon 1983) which was used to establish the link between consumption, identity and social roles. Symbolic interactionism proposes that individuals make sense of their world in terms of the social realities from which they derive their identities; and that actions and behaviours constitute ways of enacting their social reality and identities. Consumption is seen as one of the means by which individuals create their identities and this is linked with McCracken’s approach to culturally constituted meaning systems:

"They [consumers] use the meaning of consumer goods to express cultural categories and principles, cultivate ideas and sustain lifestyles, construct notions of self, and create (and survive) social change... " (McCracken 1990:xi).

Biddle and Thomas’ (1966:11) definition of social role as: "a behavioral repertoire characteristic of a person or a position .. a set of standards, descriptions, norms or concepts held (by anyone) for the behaviours of a person or a position ... a position" has been adopted for exploring the associations between groups of consumers as shoppers and consumption configurations. McCall and Simmons (1982) argued that collections of products and activities are taken by society as defining social roles. Schenk and Holman (1980) and McCracken (1986) argue that roles are culturally constituted:

"Social role performance are learned behaviors that are culturally determined. An individual’s actual role performance cannot be studied by assessing only the requirements of his/her social position since the individual combines the broad culturally defined demands of his/her position with individually-defined goals". (Schenk and Homan 1980:611)

The interdependence between the cultural context and the individual context can be seen when examining the role of parents, mothers and fathers as consumers and shoppers. Their social role performance involves learned behaviours which are culturally determined, and yet these demands are mediated by individually-defined goals.

Self image and self concepts are also central to this examination of the relationship between self and consumption. Rogers’ (1951) theory of individual self-enhancement has been used to propose that as the self concept was of value to the individual then behaviour would be directed towards the protection and enhancement of an individual’s self concept (Grubb and Grathwohl 1967). Despite some concerns about the use of the self concept construct in the consumer behaviour research literature (Sirgy 1982, Lee 1990), situational self concept (Schenk and Holman 1980) will be used here because of the recognition of the central role of the situation or context in determining consumer decision-making, and the emphasis again on the interdependency between the consumer and the cultural context. Studies have confirmed that characteristics of self-image are congruous with characteristics of brand image (Dolich 1969; Green et al 1969; Landon 1974; Ross 1971; Sirgy 1982; Snyder and de Bono 1985). However the idea that a single product or brand can be seen as representative of a consumer’s self concept has been challenged (Douglas and Isherwood 1979, Solomon and Assael 1988). It is proposed here that configurations (Figure 1) consist of consumption constellations and anti constellations, which in turn represent the complementarities and anti complementarities amongst the goods and services used by consumers.



This paper reports constellations which can be associated with groups of U.K. shoppers. Marchand’s ensembles (1985) and McCracken’s Diderot unities (1988) are also linked to the argument (Douglas and Isherwood 1979, Belk 1988) that in order to understand consumption it is necessary to examine the full picture of consumption, and not just the choice of any particular item. Associated with the idea that consumption decisions reflect consumers’ statements about themselves, the model building also includes 'anti constellations’ which represent non consumption which involves both non choice and anti choice. Non choices can include products and services which are not purchased (possibly because they are outside the means of the consumer). Anti choices involve goods which are positively not chosen (Wilk 1994)and relate to Bourdieu’s (1979) concept of the 'refusal of taste’.

It is proposed that there are three forces which can be identified as influencing configurations within consumers’ patterns of consumption: a symbolic-functional force, represented by the nature of the product (understood in its widest sense of goods or services); the physiological-esteem-self actualization force which is represented by the nature of the need(s) to be met (covert behaviour); and the instrumental-expressive force which is represented by the nature of the overt behaviour, that is the actions and activities which are involved (Figure 2). Consumption configurations are seen to be the result of the interaction of these three interdependent forces.


The general research objective proposed that the composition of the intermediate patterns could be established via the quantitative analysis of a large consumer database using correspondence analysis.

Previous studies of joint consumption (Wells 1968, Alpert and Gatty 1969, Solomon and Buchanan 1991) have employed the analysis of large consumer databases in pursuit of evidence of patterns of joint consumption. For this study data was extracted from a large, commercial database: BMRB/TGI which is compiled from an annual consumer survey based on straified sampling using self completion postal questionnaires which are distributed to 40,000 households to generate 25,000 usable responses.

The BMRB/TGI contingency tables contain aggregated categorical data. Correspondence analysis was chosen as the most appropriate statistical technique for analyzing this type of data, as it is a variant of principal components analysis which operates on categorical rather than continuous data (Greenacre and Hastie 1987) and because its multivariate treatment of data allows it to reveal relationships which would not be detected via a series of pairwise comparisons (Hoffman and Franke 1986:213). Using the 'transition formulae’ produced by correspondence analysis it is possible to see not only which features are clustered together, but also to understand why they are clustered together (Underhill and Peisach 1985:41). However, the element of subjectivity in interpretation has also to be acknowledged (Dittmar 1992): "By its flexibility, correspondence analysis can lead to greater insight into the phenomena being studied because it affords several different views of the same data set. Subjectivity of analysis is part of the price of this flexibility" (Hoffman and Franke 1986:225).


The formation of the consumption combinations: constellations, anti constellations and configurations, was seen as determined by the interaction of three forces: symbolic-functional, physiological-esteem-self actualization, and expressive-instrumental. Drawing on earlier research which had identified a series of drivers of consumption combinations (Solomon 1983, Solomon and Assael 1987, Solomon 1988, Solomon and Buchanan 1991) it was expected that certain product categories would be found within the different combinations of consumption and that certain commonalities would be found across the different combinations of consumption.



In choosing the components to represent joint consumption in the configurations, the aim was to incorporate categories which had been established by earlier studies as being significantly linked to occupation and social roles in exploring joint consumption (see Table 1). Seventeen categories were included in the model building and this was similar in number to earlier studies (e.g. Solomon and Assael(1987)).

Solomon and Assael (1987) had found that three categories: clothing, electronic equipment and cars, had represented almost half of the product elements elicited when subjects were asked to specify the products which they associated with a particular range of occupations and these product categories were therefore included as components for consumption constellations since they could potentially represent significant differences amongst the consumption constellations. Other categories carried forward for the model building included liquor and tobacco, media, food, personal care products, sports equipment and home products (Solomon and Assael 1987). In order to extend the product categories some components from the later studies (Solomon 1988 and Solomon and Buchanan 1991) were also included such as luxury appliances, credit cards and discretionary use of leisure time. Some additional categories, which complemented existing categories, were also incorporated into this model building and included holidays (which linked to the category: 'discretionary use of time’), and health and diet (which were linked to sports and keep fit equipment). The most important new addition to the product/service components in this model building was represented by the inclusion of retailers as a separate and identifiable category. It was expected that retail constellations would form a significant feature in consumption constellations and anti constellations, and would indicate important differences between the groups of shoppers which could be linked to the forces which influence the formation of consumption configurations.


Data was extracted from the BMRB/TGI database on users of eight mail order catalogue titles (representing the eight major U.K. mail order companies (G.U.S., Kays, Littlewoods, Freemans, Grattans, Empire Stores, Next Directory and J.D.Williams)) across a range of categories. The data matrix for the study consisted of over two thousand rows by eight columns. Correspondence analysis was used to identify potential groups and subgroups among the users of U.K. in-home shopping; and to identify combinations of retailers, goods and services which were used and which were not used by the various mail order shoppers.

The relationship between the mail order catalogues (the column profiles) and the dimensions can be seen from Table 2. Inertia indicates how well each column or row is represented by a particular dimension. The inertias for the seven dimensions produced by correspondence analysis of the BMRB/TGI data are shown in Table 3. Over 60% of the inertia is represented by the plot of first two axes.

A two stage procedure was adopted in constructing and interpreting the constellations. The tables of transition formulae were examined to identify constellations amongst the row profiles from which to elicit product constellations and anti constellations. Those row variables which could be clearly associated with one predominant axis were included as part of the 'set of consumables’ which were relevant to the users of the catalogue which could be linked to that dimension or axis. One difficulty was deciding what the measure of 'relevance’ should be, particularly in the absence of any preceding research using correspondence analysis for identifying combinations in the data set. Table 3 had shown that nearly 61% of the inertia in the data had been accounted for on the first two axes, and that the remaining 40% inertia had been accounted for by the remaining five dimensions. Within the context of these inertia figures it was decided to use COR values of over 500 for the shoppers linked to dimension 1; COR values of over 400 for shoppers linked to dimension 2; and COR values which fell between 300 and 400 for the remaining five dimensions, when identifying the consumption combinations. The other important research objective of the study was to identify the potential anti complementarity of goods and services. Anti constellations have not been operationalized in any of the existing studies of constellations and therefore theory building about this aspect of consumption choices remains largely undeveloped. For the purposes of this study it was decided that absence could be treated either explicitly or implicitly. Interpretation of explicit non consumption could be based on a study of the details of the 'non usage’ of categories by groups of shoppers, and some of this information was extracted from the BMRB/TGI database and was included in the correspondence analysis of the cross tabulated data. Implicit non consumption could be constructed from the absence of specific categories from the constellations. There is evidence from earlier studies on the complementarity of goods that certain choices are seen as more feasible than others, e.g. choices made by consumers amongst different menu listings (Green, Wind and Jain 1972).







The second stage of this procedure, the interpretation of the constellations, anti constellations and configurations was undertaken within the framework of the model building of the forces which influence intermediate patterns of joint consumption. It was expected that different product constellations and anti constellations would be found and these consumption combinations would reflect the differing responses amongst consumers to the various forces which influence decisions on consumption. Where the influences were predominantly symbolic and expressive then consumption constellations and configurations would be expected to show a concern with symbolic consumptionof goods and services, notably by the use of particular brands (e.g. product brands such as Levis; or retail brands such as Harrods). Where the influences were predominantly functional and instrumental then it would be expected that different consumption configurations would be found, consisting of different constellations of goods and services (with the emphasis on factors such as value for money; and the convenience of shopping). Consumers’ configurations of consumption would also be influenced by their needs (physiological-esteem-self actualization) which would be reflected in the consumer choices. Certain product fields could be associated with the consumption combinations and could therefore be taken as proxies of consumer behaviour.


[Details of the tables of transition formulae, and most importantly of the COR figures for the product categories, are not included in the paper because of lack of space.]

Correspondence analysis of the behavioural data showed that constellations and anti-constellations of consumption could be identified from which consumption configurations could be derived. These consumption configurations could be linked with three different groups of users of mail order catalogues: two groups were associated with focused catalogue offerings, and one group with the mass market mail order catalogue offering. It was possible to associate combinations of product and service usage and non usage with distinct groups within the mail order shopping population. For the purposes of this paper, only the three main groupings will be reported.

The first important consumption constellation could be associated with Next Directory shoppers(an up market niche catalogue). The consumption constellation on the positive axis of the first dimension was characterized by the presence of a range of branded goods (e.g. Levi 501 jeans, L.A.Gear trainers); strong retail brands (e.g. Benetton, Laura Ashley and Harrods); high status foreign cars (e.g. BMW, Saab and Volkswagen Golf); electronic products which were linked both to status items (e.g. cordless telephones) and to goods from the top end of the price range (e.g. expensive cameras); and upmarket branded toiletry products (e.g. Chanel, Paco Rabanne and Yves St. Laurent aftershave). The consumption anti constellation included the absence of the tabloid press; British cars; package holidays; tobacco products; tinned food products; and mainstream mail order catalogues. A concern with health could be seen in the absence of tobacco and tinned food from this configuration. The features of this consumption configuration were indicative of the importance of the symbolic and expressive forces in influencing the formation of this consumption combination.

On the second dimension (positive axis) a significant consumption configuration could be associated with the J.D.Williams catalogue shoppers (mainly larger, older women shoppers using a niche catalogue) and this consumption constellation displayed an emphasis on large, particularly foreign, cars (e.g. Austin Rover, Audi, Mercedes, Mitsubishi); a very detailed food constellation (which was seen by the variety of detail about bread); a strong association with a series of mail order catalogues (e.g. Ambrose Wilson, Fashion Extra, Fashion Plus, Heather Valley, Oxendales, Personal Selection and J.D.Williams); and a rather different combination of leisure activities, relating to watching sports, compared with the dimension discussed above. This consumption configuration included watching bridge, cricket, golf and show jumping in its consumption constellation. The consumption anti constellation included electronic products (as explored in this data set); branded clothing products and also branded toiletries; holiday destinations such as Canada and the U.S.A.; eating out in the evenings (apart from at steak houses); and fashionable high street stores such as Next Retail, Principles, River Island and Wallis Shops. This consumption configuration suggested the importance of the instrumental rather than the expressive force in the choice of retail outlet; and also suggested a concern with needs from the physiological end f the spectrum. The presence of large, often foreign, cars suggested the influence of the symbolic force on this aspect of consumption decisionmaking.



The consumption configurations on the remaining dimensions could be associated with the six mass market catalogues, and were representative of a series of subgroups within this mass market group of catalogue shoppers. For the purposes of illustration the details for one of the major catalogue in this mass market group, Kays, will be described. On the third dimension (negative axis) the consumption configuration which could be linked to Kays shoppers consisted of small and medium sized cars (e.g. Cavalier, Citroen, Nissan/Datsun Sunny); some electronic products from the medium price range; limited branded clothing products (Champion trainers); and an association with the two largest mainstream mail order catalogues (Kays and Littlewoods). The discretionary use of time included watching cricket, horse racing and bridge. The anti constellation included most notably high street retail brands (e.g. Ravel, Russell and Bromley, Laura Ashley, Jaegar, Richards and Solo). This consumption configuration seemed to have been influenced by instrumental and functional rather than expressive and symbolic forces in the choice of retail brands. The minor presence of branded clothing and electronic products suggested a compromise between symbolic and expressive needs and budgetary pressures in this combination. In Table 4, the main points are summarized for each of the major groups of catalogue shoppers.

Table 5 indicates the presence of the product categories in the various consumption constellations and confirmed earlier research about the importance of clothing, electronic goods and cars as important distinguishing characteristics of consumption combinations. Discretionary time and retailers also emerged as important distinguishing characteristics of consumption choices.

From the consideration of the COR figures for the row points it was clear that some of the product category items overlapped the various dimensions (e.g. some of the cars and some of the branding clothing products). This can be understood in terms of 'fuzzy sets’; especially where certain product categories carry image-laden messages and have symbolic meanings which are pursued by more than one group of consumers.




The specific objectives of this research were firstly to identify the composition of constellations of products, services and activities from the behavioural data and from which configurations could be derived; secondly to elicit constellations of retailers, as representative of a service category of consumption, hitherto not explored in detail in earlier studies of patterns of joint consumption; thirdly to identify potential anti complementarity constellations; and fourthly to see whether consumption configurations could be formed from consumption constellations and anti constellations, and could be associated with different groups of shoppers. The most important product categories elicited by earlier studies proved to be significant components of consumption constellations in this research and the constellations which emerged were close to those of previous studies (Solomon et al).

These findings contribute to our understanding of the interaction between the patronage and product decisions made by consumers, and would also suggest that shopping is more central to the creation and maintenance of social identity for some consumers than for others. It is important to recognize both the importance of retail patronage decisions for some consumers, and that retail constellations will vary considerably across different dimensions. Retail constellations will contribute to the picture of the enactment and maintenance of social identity; and thus to our understanding of different 'consumption communities_olomon 1987:210). However, retail constellations would probably not, by themselves, be sufficient to distinguish amongst all the dimensions. The value of the retail constellations will lie in the contribution which they can make to attempts to refine understanding of the different intermediate patterns of joint consumption.

The relationship between social roles, social identities and consumption constellations has been explored here using quantitative data analysis techniques and these findings confirm that the relationship between social roles, social identities and consumption configurations would reward more detailed study in searching for further understanding of the patterns in the stream of consumer decision making.


A complete list of references is available on request

Alpert,Lewis and Gatty,Ronald (1969): "Product Positioning by behavioral life styles", Journal of Marketing 33 (April), 65-69

Dittmar, Helga (1992): The Social Psychology of Material Possessions: To Have Is To Be. Harvester Wheatsheaf, Hemel Hempstead.

Green, Paul E; Wind,Yoram and Jain, Arun K. (1972): "Preference measurement of item collections", Journal of Marketing Research 9 (November), 371-377.

Greenacre, M. J. and Hastie, T. J. (1987): "The Geometric Interpretation of Correspondence Analysis", J. of the Am. Statistical Assoc 82(June),437-447.

Hoffman, D. L. & Franke, G. R. (1986): "Correspondence Analysis: Graphical Representation of Categorical Data in Marketing Research", Journal of Marketing Research (August), 213-227.

Kehret-Ward,Trudy (1987): "Combining Products in Use: How the Syntax of Product Use Affects Marketing Decisions". In: Marketing and Semiotics: New Directions in the Study of Signs for Sale. (Ed: Umiker-Sebeok,Jean) Mouton de Gruyter, Berlin, 219-238.

Kernan, Jerome B. and Sommers, Montrose S. (1967): "Meaning, Value and Theory of Promotion", Journal of Communication, 17 (2) 109-135

Lee, D. H. (1990): "Symbolic Interactionism: Some Implications for Consumer Self-Concept and Product Symbolism Research", in Advances in Consumer Research: Vol. 17. (Eds: Goldberg; Gorn; Pollay) ACR Provo, UT, 386-393.

Levy,Sidney J. (1964): "Symbolism and life style", in: Towards Scientific Marketing, Proceedings of the American Marketing Association Conference. (Ed: Greyser,Stephen) American Marketing Association, Chicago, .

McCall,George J. and Simmons J.C.(1982): Social Psychology: A Sociological Approach. The Free Press, New York.

McCracken,Grant (1990): Culture and Consumption. Indiana University Press, Indianapolis.

Schenk, Carolyn Turner and Holman, Rebecca H. (1980): "A Sociological Approach to Brand Choice: The Concept of Situational Self Image", in: Advances in Consumer Research. Vol. VII. (Ed: Olson,Jerry C) ACR, Ann Arbor, 610-614.

Solomon, Michael R. (1983): "The Role of Products as Social Stimuli: A Symbolic Interactionism Perspective", Journal of Consumer Research 10(December), 319-329.

Solomon, Michael R. (1988): "Mapping Product Constellations: A Social Categorization Approach to Consumption Symbolism", Psychology and Marketing 5(3, Fall), 233-258.

Solomon, Michael R. and Assael, Henry (1987): "The Forest or the Trees? A Gestalt Approach to Symbolic Consumption". In: Marketing and Semiotics:New Directions in the Study of Signs for Sale. (Ed: Umiker-SebeokJean) Mouton de Gruyter, Berlin, 189-217.

Solomon, Michael R. and Buchanan, Bruce (1991): "A Role-Theoretic Approach to Product Symbolism: Mapping a Consumption Constellation", Journal of Business Research 22, 95-109.

Wells, William D. (1968): "Backward segmentation", in: Insights into Consumer Behavior. (Ed: Arndt,Johan) Allyn and Bacon, Boston, 85-100.

Wilk, Richard R. (1994): "'I hate Pizza’: Distaste and Dislike ..." American Anthropological Association meeting, Atlanta, November.