Trajectories of consumption across a wide spectrum of urban services and resources in high income societies continue to trend upwards [1,2]. At the same time, attempts to create “wedges” linked to the introduction of social and economic policies, new processes and products, etc. [3,4], capable of underpinning a significant reduction in consumption, underpin much contemporary research. Many prospective sustainability “wedges” are found in the literature and include:

Dematerialisation: a reduction in the resource inputs required to produce consumer products, such as building materials, domestic appliances and automobiles via recycling, re-use and closed loop manufacturing;

They cover the spectrum of research from sustainable production to sustainable consumption—from the supply side to the demand side. A conceptual framework capable of being used to “map” consumption research is presented in Figure 1. It suggests that demand-side studies—the focus of this paper—will centre on better understanding the role of social and demographic, as well as behavioural attributes associated with individuals and households and the contextual settings related to their dwelling, urban location or social milieu. Demand-side research also needs to probe why studies of urban resource use [5,6,7,8] are finding a gap between attitudes and actions; individuals are not behaving in a manner that is congruent with their stated attitudes and intentions. Here, a major question mark continues to surround the issue of whether individuals or households in high income societies are prepared to make voluntary changes to their consumption practices in order that the 21st century has some prospect for a more sustainable and equitable world. This necessitates a drilling down into the structural and behavioural attributes of individuals to explore those factors, both intrinsic and extrinsic, linked to consumption [9]. It can be argued that behaviour change on the part of individuals and the households with which they are associated holds the prospect for a much faster rate of sustainability transformation than supply-side technological innovation of key infrastructures and services (e.g., energy, water, waste or in the redesign of built environments). Radical change is possible in both these arenas, but the timescale required is considerable, hence the increasing interest being shown by governments in behaviour change policies and programs [10,11].

There is both an applied and a theoretical dimension to the study of household resource consumption and behaviour change. It has been argued that no voluntary changes in consumption practices can be achieved if policy makers, business and consumers alike lack an understanding of how and why individuals and households consume and what factors tend to be associated with behaviour change [12]. In recent years, a significant number of behaviour change programs have begun to emerge, each with varying degrees of underpinning evidence base as support. What have been termed transmissive projects—for example, broadcast messages via television, newspapers and the like relating to some aspect of consumption reduction, such as energy or water conservation and healthy eating—are mostly premised on a linear model of behaviour change: information→awareness→concern→action [5,13,14]. Examples of transmissive projects include Earth Hour [15] and Black Balloons [16]. There is no segmentation or targeting involved. The economic rationalist information-based model of human behaviour has proven influential here in attracting government investment The first phase has been termed “cognitive” (awareness of information about a particular issue, e.g., climate change, drought); the second has been termed “affective” and is typically associated with some expression of concern about the issue; and the third is “action”, where individuals have been motivated to make some change in their behaviour that is tangible and measurable. The narrative of the behaviour change process in this conceptual context is therefore portrayed as follows: an environmental citizen is someone who has internalised information about environmental problems, creating a sense of concern, personal responsibility and duty that is then expressed through consumption and community actions [17]. Research suggests, however, that the three facets of behaviour change are not necessarily sequential [5]. Transformative initiatives focus on face-to-face engagement with specific communities in an attempt to achieve the necessary traction for behaviour change. Examples of transformative programs include: Sustainable Consumption Roundtable’s “triangle of change” (government, business and population) that involves engaging, exemplifying, enabling and encouraging [18]; the “kitchen table” processes [19]; the social marketing approach [20]; the “seven doors” model [21]; Green Street [22] and other socially (rather than cognitively)-based behaviour change programs [23,24].

Lifestyle segmentation (LSS) has emerged from the marketing and communications fields of research as a means by which the behaviour change programs of governments might be better targeted, whether transmissive or transformative in nature. LSS represents a marriage of two concepts into a single system [25]. Lifestyle refers to a distinctive pattern of a person’s social life that straddles notions of individual identity, on the one hand, and community/sociality, on the other, “embodying notions of choice and self-actualisation alongside opportunities for collectivity and attachment” [26]. It is a broad based concept that incorporates everyday facets of individual lives, including their attitudes, opinions, values, feelings, intentions, habits and social contexts. Segmentation refers to a division of the population or market into relatively homogeneous clusters capable of being readily differentiated or labelled—via either quantitative or qualitative routes. The premise underpinning LSS is that it identifies groups of people who might be more likely to behave in similar ways as consumers—of particular products, services, information and messages. The attractiveness of LSS and its link with sustainability is the contention that consumer behaviour is likely to vary across different lifestyle groups comprising individuals with specific combinations of environmental attitudes and practices [27]. The research tasks then become those of establishing a valid and differentiated set of lifestyle groups among a population derived from either quantitative (lifestyle segmentation) or qualitative (e.g., consumer culture or social practice) research methods (or some hybrid) and then examining the extent to which lifestyle “type” can explain differences in some aspect of consumer behaviour. There is an extensive body of publication associated with the former, but there is a dearth of studies that proceed to examine the latter.

As intimated above, two contrasting approaches to LSS have emerged in research literature: the longer established and more widely implemented “traditional” approach, which is characterized by quantitative analysis and data reduction (and which is adopted for this paper—see section on Methods); and the cultural theory or social practices approach, which is qualitative in nature, drawing on a different set of disciplines for its execution. Quantitative LSS obtains responses to self-rating questionnaire items encompassing attitudes to a wide spectrum of statements that are deemed to cover the breadth of domains of interest to contemporary living. They are then subjected to some form of cluster analysis to identify groups that have similar response patterns, after which descriptive labels are assigned capable of evoking the distinctiveness of each group. The strength of this approach is its ability to synthesise a large volume of complex information into meaningful insights that are valued by clients in both private and public sectors. LSS studies of this type have been part of market and consumer research since the 1960s and 1970s for products, such as cars, alcohol, tourism, clothing and furniture, and there are many proprietary lifestyle typologies that have been developed. These include GfK Roper’s broad, multi-product lifestyle segments that comprise: dreamers, homebodies, settled, adventurers, rational-realists, open-minded organics and demanding [28]. Other segmentations [27,29] have been developed for specific applications, such as:

energy saving behaviour: economically modest, open-minded value pluralists, hedonists, conservative environment conscious, alternative environment-conscious, disinterested materialists and potentially sustainable;

The weaknesses of this class of LSS studies have generally been the lack of validation or replication and lack of proof as to their capacity to discriminate or predict actual behaviour.

The second, qualitative approach to LSS research to emerge more recently has been termed the social practice or cultural theory approach [30,31,32]. It focuses on the common contexts of everyday life, such as homes and workplaces, and the associated research seeks to understand how observed consumption behaviours reflect daily routines and practices, i.e., habits. It views the traditional LSS as failing to understand the “complex socio-cultural processes and situations that underlie human behaviour” [33]. By undertaking in-depth interviews often combined with observational and participative (ethnographic) research, it provides the basis for fleshing out lifestyle segments as part of a process for better understanding and characterizing contemporary society and the value this holds for developing more successful strategies for behaviour change. Its prime weaknesses have been identified as restriction to a small number of consumption settings and activities, such as water and energy use in the home [34], and a focus on overly complex, in-depth, but small-sample, qualitative studies, uncovering experiences that are often seen as unique to individuals and their spatial and temporal settings and are difficult to generalise or ascribe to a larger population [35].

More recently, LSS research has been directed towards the identification of an evidence base for behaviour change policies and programs associated with the environment and climate change [36,37,38]. In a UK study [39], seven lifestyle clusters were identified: positive greens, waste watchers, concerned consumers, sideline supporters, cautious participants, stalled starters and honestly disengaged. In a rare replication of such studies (albeit with less than half the indicator variables) in New Zealand [40], significant variation was found in the proportions of the population aligned with each of the seven lifestyle segments. In a similar Australian study [41], three segments were derived from self-reported environmental behaviours: committeds, privates and reluctants, “naturally forming groups of people who exhibit similar behavioural patterns regarding the environment”. A United States study [42], identifies “six Americas”, where each population segment holds different attitudes to the issue of climate change: the alarmed (18%), the concerned (33%), the cautious (19%), the disengaged (12%), the doubtful (11%) and the dismissive (7%). In later surveys [43], the alarmed and concerned proportions had declined somewhat and shifted to “cautious”. What all studies reveal is the existence of common environmental lifestyle “types” ranging from the green/engaged to the sceptic/disengaged. None of these studies, however, assessed alignment of segments with target actions or behaviours.

Many lifestyle typologies, such as those listed above, are made up of objective, as well as subjective, attributes. This would appear to be valid for descriptive analyses of a population or sample. To the extent that socio-demographic factors have been found to be particularly significant determinants in their own right of consumer behaviour [44], combining socio-demographic with attitude, opinion and intention attributes in a segmentation for use in explanatory or predictive modelling of consumption risks redundancy. As Gust [27] argues:

Socio-demographic variables, such as age, income, family size and lifecycle stage were found to limit any attitude or lifestyle variables. For example, although young mothers are quite environmentally conscious consumers, this consciousness may not go as far as replacing their car with public transport, since the car might be very important in the daily family organisation. Here, the fact that there is a small child in the household will have a greater effect of the likeliness of sustainable consumer behaviour than the environmental orientations of the mother.

Other weaknesses of the lifestyle approach identified by this author [27] are that most typologies are conceived for specific applications and, as such, lack the prospect for comparative assessment or replication and that attitudes and behaviours are highly dynamic and can change at any point in time, especially as needs and contexts change (e.g., environmental versus economic priorities, before compared to during the global financial crisis of 2008 and 2009).

The consensus is, however, that lifestyle orientation and segmentation can provide additional insight into consumer behaviour that may not be attributable to socio-demographic factors alone. The LSS approach has also been embraced as a tool with a capacity to tailor and better communicate sustainable behaviour diffusion strategies, as well as to market sustainable products and services to specific groups more likely to be receptive to the differentiated message [36].

Against this background, the aim of this paper is four-fold: (1)

To establish an environmental lifestyle segmentation based on the cluster analysis of data from a representative sample of Melbourne household’s responses to a range of questions that probed environmental attitudes, opinions and intentions;

Given that the research objective of this paper was to explore the nature of any link between “environmental lifestyle” and consumption, it was deemed necessary to establish a lifestyle typology that derived from an interplay of the behavioural attitudes, opinions, concerns and intentions of a population—not their structural (socio-demographic) attributes, nor their actual patterns of consumption (see Figure 1). In this, it aligns with the traditional LSS approach, sometimes characterized as AIO (activities, interests, opinions), since it was part of a much larger study of the determinants of urban consumption (see Acknowledgements; [44]). The objective of the overarching study was to quantify the respective contributions of individual (behavioural, structural) versus contextual (dwelling, urban location) to an explanation of energy and water use, housing and domestic appliance consumption and carbon emitted on city travel. In this way, the survey provided data enabling an examination of the extent to which there is or is not a gap between the subjective indicators of sustainable living (the lifestyle groupings) and actual consumption behaviour (see Figure 2). This is a “model” for studies that seek to explore the disparity between awareness, concern and attitudes among the population for some significant issue, such as climate change and consumption and the level of action displayed by that population [45]. Literature suggests an explanation for the identified gap can be attributed to a set of barriers, constraints or situations that include the psychological, social and institutional, as well as the informational.

How does actual consumption vary for these three clusters? A multivariate general linear model analysis performed using the five per capita consumption measures defined previously suggested some small significant differences between the clusters (F(10,1430) = 2.99, p = 0.001, partial eta-squared = 0.021), but only in the case of the carbon intensity of travel and per capita appliance consumption (see Table 13). Appliance consumption was highest for the “enviro-sceptics” and lowest for the “material” greens, while carbon intensity of travel was highest for the “material” greens and lowest for the “committed” greens. However, a univariate general linear model analysis for the total of these consumption measures (after standardisation) showed no significant difference between the clusters (see Table 13). Lifestyles characterised by pro-environmental values, attitudes and intentions did not reflect actual low consumption behaviour.

Next, the clusters were compared in terms of urban location. Table 14 suggests that there is some geographical segregation (Chi-Sq = 35.7, df = 4, p < 0.001). Half of the “committed” greens live in the Inner City suburbs (where, in recent years, the Greens Party has become politically dominant), while the “material” greens tend to live in greenfields and outer suburban areas—the group that political commentators have termed as “aspirational” in terms of their ambition to “get ahead” in life. The “enviro-sceptics” cluster is dispersed across the city.

Are there significant socio-demographic differences for these three clusters of people? Indeed, there are. Table 14 suggests significant differences for age, gender, level of education, household income, family structure and suburb location. However, there was no significant difference for country of birth (Australia versus overseas).

The “committed” greens cluster contained more university graduates and households with higher incomes. They know what behaviours are likely to be required in a climate- and resource-constrained future and have a financial capacity to embark on that transition. The “material” greens households, by comparison, had the lowest proportion of university graduates, were the youngest category of households and also tended to be on lower household incomes. The “enviro-sceptics” cluster contained more men and those aged 45 and over than either of the other clusters. Although the “enviro-sceptics” and “material” greens clusters tend to have similar incomes, the latter cluster is more likely to consist of households with children, which could have had some influence on their pro-environment attitude.

Overall, it seems that the “committed” greens cluster can afford to have “green” attitudes and “green” behaviours. Low consumption in the case of carbon intensity of personal travel suggests that this cluster is making an effort to reduce consumption in some areas—by virtue of where they live in the city—but in other areas appears to be doing no better than the other clusters. The “material” greens cluster also has sympathy with the environment, but these households feel that their limited budget does not allow them to pay higher rates or taxes in order to help the environment. As indicated in Table 13, they have the lowest per capita consumption of appliances (reflecting the “averaging out” effect of families with children identified earlier), but in other areas, their consumption is no lower than the other clusters; in fact, their carbon intensity of travel is particularly high—reflecting the car dependency of households in the outer suburbs. Finally, the older, male-dominated “enviro-sceptics” cluster tends not to care about the environment as much as the other clusters, but, surprisingly, the total consumption score does not appear to be significantly higher for these households, although their mean total consumption score is highest overall.

Changing behaviour, which involves winding back currently unsustainable levels of resource consumption, has the prospect of making an impact much more rapidly than most, if not all, supply-side responses, although both clearly need to be operating in tandem if the 21st century is to deliver a sustainable platform of urban living. It is the relative speed with which individuals and households can potentially decide to stop or significantly alter a range of domestic practices—related to their choice of housing, mode and frequency of travel, energy and water use habits, appliance purchases (number and environmental performance)—that makes “social marketing” so attractive to governments. An ability to get households to change aspects of their behaviour voluntarily via engagement processes that seek to provide the social, environmental and economic narrative for change and the processes (“steps”) by which this can be achieved would obviate governments having to pursue less politically attractive and costly paths (such as taxation, regulation and incentives) to secure more sustainable patterns of consumption.

This paper has found that voluntary change will be no easy task. Actions speak louder than words.

The research has demonstrated that there is a gap between intentions and actions. Segmentation of the Living in Melbourne sample of households into lifestyle groups based on their responses to a battery of questions probing environmental sustainability-related values, attitudes and intentions produced three archetypal clusters: “committed” greens, “material” greens and “enviro-sceptics”. Each cluster was also found to be distinctive on a range of socio-demographic variables that further validated the typology. Commonalities with similar typologies elicited for the UK, USA and NSW (Australia) are strong. On the basis of the Living in Melbourne survey, only one-third of households (the “committed” greens) would currently appear to be prepared to voluntarily change their consumption behaviour and bear the direct economic consequences. However, when it came to examining the extent to which actual consumption behaviour varied across the three clusters, there was little or no differentiation evident.

The reasons why there is currently a gap between intentions, on the one hand, and action, on the other, are to be found in the range of current barriers to sustainable consumption that need to be overcome. These include: problems of lack of relevant information or how/where to find it; organisational challenges, i.e., how to get it done; time constraints; and financial constraints—determining whether the benefits reward the financial outlay. At a pragmatic level, there remains a lack of information on what can be done and how best to get it done. For individuals and households, comfort, convenience and cost factors seem to underpin many of the habits and practices that currently promote consumption of urban resources in Australia. The issue of sustainable consumption would appear to be more deeply rooted within 21st century society, however. High income societies are now over-reliant on consumption as an engine for growth—and developing economies are being encouraged to follow suit. Social norms relating to sustainable consumption are yet to materialise in high income societies, such as Australia, that would constitute an important influence on the voluntary behaviour of individuals and households. A culture of unsustainable consumption is reflected in a dominant set of behaviours at present. And it appears to be embracing all segments of the population, including those who espouse green attitudes, opinions and intentions. It is here that social practices research can add value to our understanding by probing more deeply into these lifestyle segments to explore whether consumption related habits and activities are also held in common. Whether imminent system failure will be required to trigger a “tipping point” in societal values associated with environment and consumption remains an open question (witness the increasing number of well-credentialed reports emerging of a future “4° world”). This is a major reason why supply-side urban sustainability initiatives need to proceed apace, why governments need to remain actively involved in regulation, pricing and incentive programs and why research that spans the cognitive-social spectrum of consumption must continue to search for triggers for effective behaviour change. Understanding urban consumption is clearly a more complex challenge than a market segmentation approach alone can address.