1 Introduction

The market concept has been always considered relevant within economics and business discipline, with multiple definitions (Araujo et al., 2008; Vargo, 2007) yet not fully theorised. In sociology economics it has been perceived as a social arena (Fligstein & Dauter, 2007; Granovetter, 1985; Swedberg, 1994), in economics as linked to institutions (Stiglitz, 1989) and to performativity (Callon, 1998) and in entrepreneurship associated with market creation (Sarasvathy, 2003).

Within marketing, scholars have particularly criticised the idea of market as a given and deterministic context, exogenous to the firm (Mele et al., 2015), in favour of socially constructed phenomena which draw upon multiple views (Peñaloza & Venkatesh, 2006). However, there is still a quest for conceptual delimitation (Kjellberg & Helgesson, 2007; Pels et al., 2022). As Vargo et al. (2017, p. 260) note: “a market is one of the most foundational abstractions in marketing and business research”—yet, as Venkatesh et al. (2006) observe, “paradoxically the term market is everywhere and nowhere in our literature” (p. 252). There are a few studies that have conducted literature reviews on how the market has been conceptualised in marketing (e.g. Diaz Ruiz, 2012; Holmqvist & Diaz Ruiz, 2017; Möller et al., 2020). These articles have tried to map the field; the focus has been on classifying and describing the multiple research streams or approaches by identifying similarities and differences between the approaches (Pels et al., 2022). Others suggest a set of underlying dimensions in market conceptualisations, based on entities or processes (Mele et al., 2015). Recent works have looked at the theoretical underpinnings of the process view of market conceptualisation, addressing marketing innovation (Sprong et al., 2021) and market-shaping (Nenonen & Storbacka, 2021). However, none of these papers focused on identifying the emergent themes associated with the understanding of the market. Emergence can be seen as the generation (from pre-existing material) of new entities, structures, definitions or concepts (Bhaskar, 2008). By adopting the conceptual framework provided by the social emergence model (Sawyer, 2005), we will frame the emergence process of the market in terms of collective definitions of shared intentions and joint activities actors are engaged in (Baker & Nenonen, 2020; Taillard et al., 2016). Sawyer’s (2005) social emergence model establishes five ontological levels ranging from the individual to the wider institutional social structure. These levels show a progression from the individual reflection (in our case, isolated publication) to interactional level (in our case, a theme which is starting to gain visibility in journal publications), to ephemeral emergence (in our case, themes that start to increase citation frequency), to stable emergence (in our case, themes that become transversally adopted), to creating social structure (in our case, becoming the ‘new’ mainstream). Conversely, the upper levels of Sawyer’s model exert a downward force, as structure constrains the emerging agency of the collective and the individual. We start a conversation by mapping the field—what MacInnis (2011) has labelled as explicating through delineating. Our outlook follows Lather’s (2006) suggestion to move away from incommensurable stances or the search for the supremacy of one approach over the others, and instead to argue for multiplicities and proliferations. Specifically, our aim is to offer an overview and analysis of how market definitions are being portrayed over time in marketing, by capturing the emerging themes.

Our research questions (RQs) are: What are the emergent themes associated with the understanding of the market? Which themes are reaching emergent stability?

Consistent with the bottom-up process of emergence, we will conduct a bibliometric analysis of the marketing literature, as “it is a highly useful approach for developing an overview of a research field” (Martínez-López, et al., 2020). As Donthu et al. (2021b) explain, scholars can “uncover emerging trends …. and explore the intellectual structure of a specific domain in the extant literature” (p. 285). Our analysis aims to track “the intellectual interactions and structural connections among research constituents.” (p. 290). First, to address which are the main representative terms, we will observe the authors’ keyword distribution in the relevant field of literature in the 1990–2020 period. Next, to highlight which terms seem to be moving from interaction to ephemeral emergence to stable emergence, we will conduct a co-occurrence network analysis that allows us to elicit the associations among terms, as well as a thematic map which shows the development of themes (density) and degree of relevance (centrality). In the marketing discipline, several previous studies have conducted bibliometric studies. For example, Colurcio et al. (2017) analysed how resource integration has been portrayed in marketing and management; Martínez-López et al. (2020) conducted retrospective evaluation of the Industrial Marketing Management Journal; Chabowski et al. (2013) studied the global branding literature and research agenda, while Pohlmann and Kaatermo (2017) looked at the research trajectories associated with the S-D logic. Recently Sprong et al. (2021) have adopted bibliometric mapping within market innovation research.

In the midst of a dynamic debate, we argue that our contribution is aligned to Donthu et al. (2021b), who acknowledge the importance of bibliometric review in order to “unpack the evolutionary nuances of a specific field, while shedding light on the emerging areas in that field” p. 285). By mapping the current themes and adopting the social emergence model as a conceptual framework, we make three contributions to the marketing literature’s conversation on the market. First, by identifying two periods (1990–2004 and 2005–2020) we show that in the last fifteen years, conceptualisation of the market has been in a constant evolution. Second, we recognise five thematic clusters: performance, marketing systems, market shaping, market practices and institutions. These clusters differ both in terms of their associated thematic complexes and in terms of their density and centrality. Finally, based on our bibliometric analysis, we suggest the positioning of the five clusters within Sawyer’s social emergence model. We position two of the clusters at the institutional level (performance and market system), two at the ephemeral level (market shaping and institutions) and one at the interaction level (market practices).

The paper’s structure is as follows. The next section introduces the social emergence model. Second, we explain our research process. Third, we present the main descriptive statistics and the results of the thematic analysis, showing the emerging research trends and how they relate to each other. Finally, we discuss the results, adopting the social emergence framework, and we highlight our conceptual contributions as well as our future research suggestions.

2 Conceptual background: the social emergence model

As stated in our research questions, in this paper we are focusing on the emergence and degree of stability of themes associated with the (re)conceptualisation of the market. Our choice of adopting an emergence framework to analyse the emerging themes associated with the understanding of the market is aligned with Bhaskar’s (2008) definition of emergence as the generation (from pre-existing material) of new entities, structures, totalities, or concepts; as well as Bunge’s (2003) classification of possible emergent systems into: material systems (further subdivided into natural, social and technical systems), conceptual systems (e.g. a scientific theory), and semiotic systems (defined as one composed of signs (e.g. a language).

The study of the emergence of social structures has a long history within sociological literature, and its review is beyond the scope of this section and paper. We draw on Sawyer (2005) for two reasons. First, because he provides a thorough review of the broad literature on emergence—the contrasting positions of the structure paradigm (e.g. Giddens, 1984) and the interaction paradigm (e.g., Bourdieu, 1977)—guaranteeing that all approaches have been considered. Second, and more importantly: in his analysis of both approaches, he finds valuable insight but also limitations. To overcome the limitations, he suggests a novel social emergence paradigm. Sawyer (2005) advocates for the existence of five ontological levels (see Fig. 1): individual (where individual preference and judgements are made), interaction (such as discourse, collaboration, and negotiation among individuals), ephemeral emergent (includes properties such as topic, status, relative role), stable emergent (such as group subcultures, catchphrases, conversational routines), and the social structure level (includes procedures, laws and regulations). Sawyer (2005) argues that emergence moves upwards between the five levels but, at the same time, he recognises that institutions exert downward forces; we shall not discuss this aspect as we will argue that the data shows that the domain of (re)conceptualising the market is still at an early stage. However, the downward effect, in the emergence of new approaches, has been discussed within the marketing literature; for example, the gatekeeping role of journals with regards to qualitative research (Gummesson, 2005).

Fig. 1
figure 1

Adapted from Sawyer’s (2005). Emergence model (text in brackets refers to the application in the context of consolidation of emergent concepts)

Full size image

The adoption of Sawyer’s (2005) five stage social emergence model is not new in the field of marketing. It has been applied with a focus on agency in the form of shared intentions (Taillard et al., 2016) or in collective market work (Baker & Nenonen, 2020, to explore the emergence of service ecosystems and market shaping respectively. In this paper, we will adopt the emergent processes to help to visualise the bottom-up effect of new themes moving from isolated efforts toward gaining robustness and support.

3 Research process

In this section we explain our research process and how the data analysis is conducted. We follow the workflow suggested by Cobo et al. (2011b), Aria and Cuccurullo (2017) and Donthu et al. (2021a): methodological choices, data collection, data analysis, mapping and interpretation (understanding themes).

3.1 Methodological choices

Bibliometric analysis, as the quantitative analysis of the bibliographic features of a body of literature (Hawkins, 2001), comprises a set of tools that apply mathematical or statistical methods to written text, including articles or books (Pritchard, 1969). In other words, bibliometric methodology encompasses the application of quantitative techniques (e.g., co-citation analysis) on bibliometric data (e.g., publication and citation units) (Pritchard, 1969). According to recent studies (Donthu et al., 2021a, 2021b; Galvagno, 2022; Pattnaik et al., 2022) bibliometric analysis has become valued increasingly over the past few years in business research due to its usefulness for (a) handling large volumes of scientific data and (b) facilitating impactful research. The two main approaches to bibliometric procedures are performance analysis and science mapping (Bornmann & Mutz, 2015; Cobo et al., 2011a; Noyons et al., 1999). The former involves the evaluation of the output and impact of groups of scientific actors, the most eminent authors in a research area, their affiliations, countries, and journals where they have been published; using bibliographic data. The latter complements performance analysis using co-citation techniques and keyword co-occurrence, which are used to analyse the current or evolving conceptual structure of a research field (Donthu et al., 2021a). Some scholars (Fortuna et al., 2020; Verma & Gustaffson, 2020) assess that although bibliometrics is popular mostly for quantifying scientific output and measuring its quality and impact, it is also helpful for visualising and analysing the intellectual, conceptual and social structures of research, as well as their evolution and dynamic aspects (Aria & Cuccurullo, 2017). In detail, science mapping could identify (a) the conceptual structure (i.e., examining the research front) (b) the intellectual structure, (i.e., the roots and knowledge base of a research topic or field), and (c) the structure of the social network (Aria & Cuccurullo, 2017).

In this study, we aim to map the conceptual structure of the market research field by developing a co-occurrence network analysis and a thematic map. The combined use of these techniques allows us to illustrate how concepts relate to each other (co-occurrence) and the main themes (centrality) within a given research field and how they evolve (density) (Colurcio et al., 2017; Triunfo & Rossi, 2021). These procedures are explained below.

By performing a co-occurrence network analysis (Aria & Cuccurullo, 2017), a research field can be represented as a set of terms (authors’ keywords, in our case).

The most important and recent topics in this paper are depicted through a network representation and inserted into a matrix that considers the number of co-occurrences of two terms in the articles. At this point, we adopted the index proposed by Van Eck and Waltman (2009) to normalise the co-occurrences between terms. In the weighted graphs, each term is a node, and the edge represents the association between linked terms as well as subgroups of terms that frequently co-occur. We used Louvain’s algorithm to community detection (Blondel et al., 2008) and thus identify subsets of strongly linked terms that are our specific focus of interest.

According to Cobo et al. (2011a) a thematic map plots themes in a bi-dimensional matrix where axes are functions of the centrality of themes (i.e., a theme’s importance) and their density (i.e., the theme’s development), respectively (Callon et al., 1983). Four typologies of themes can be positioned (Cahlik, 2000). By having high centrality and high density, the topics in the upper right quadrant are called motor subjects. They are both developed and important to the research area. The subjects in the upper left quadrant are referred to as highly developed and isolated topics or niche topics. Their low centrality means limited importance for the field, although they have strong internal connections (high density). Themes in the lower left quadrant are referred to as emerging or declining topics. Their position, with low centrality and low density, denotes that they are only slightly developed or marginal. Topics in the lower right quadrant are fundamental and transversal themes with high centrality and low density.

3.2 Data collection

Analysis of the textual content available in bibliographic databases (e.g. keywords, titles or abstracts of the publications) allows addressing the conceptual structure of a research field. In this paper, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) method was used for the selection process of the publications (Liberati et al., 2009).

We retrieved the publications from the Thomson Reuters Web of Science (formerly ISI Web of Knowledge) (WoS), which best met our study's requirements. First, during May 2020, we conducted a topical query in two citation indexes: Science Citation Index Expanded (SCI-EXPANDED) and Conference Proceedings Citation Index-Science (CPCI-S), from the WoS core collections, for inquiry, and the timespan was set from 1985 to 2020, because WoS indexes only publications from 1985.

Second, we defined our domain of study as the conceptualisation of the market, and our unit of analysis, the keywords adopted by the author(s). Our focus was on understanding how the concept of the market has evolved and the main views of what is taking place in the marketplace. Third, we looked at market related keywords. There are other keywords that have been adopted in combination with market definition such as ‘market orientation’ which we have not included in our list. These keywords are not directed to describing or modifying the understanding of the market as a concept but used to understand how firms need to gather market intelligence, internalise it and respond (Kohli & Jaworski, 1993; Slater & Naver, 1994). We also recognise that scholars could have used other terms such as networks or ecosystems to study the same phenomenon. However, our interest is limited to authors working from a market perspective in the marketing literature. To retrieve articles related to our domain of study, we used the following twenty-three keywords: ‘market theor*’, ‘market formation’, ‘market shaping’, ‘market configuration’, ‘marketization’, ‘market practice*’, ‘market system*’, ‘market ecosystem*’, ‘market scripting’, ‘market co-creation’, ‘market representation’, ‘market network’, ‘market concept*’, ‘market innovation’, ‘market perform*’ ‘market sense-making’, ‘market entities’, ‘market plasticity’, ‘market interaction*’, ‘market exchange*’, ‘market relationship*’, ‘market institution*’ and ‘marketing system*’ (* = wildcard plural).

Next, we selected (a), online articles in the Business and Management WoS categories, (b) written in English, and (c) published in fifteen marketing journals (see “Appendix”). The initial search resulted in a total of 423 articles. A manual process of reading the abstracts was then performed and 182 articles that were not related to the discussion of the market were removed. We then conducted our bibliometric analysis on the remaining 242 articles, as shown in Fig. 2.

Fig. 2
figure 2

Data retrieval process

Full size image

In Table 1 reports the main statistics about the collection analysed. Despite the timespan (1985–2020) during the retrieval process, the reference period after the selection process was from 1994 to 2020. The overall number of citations is 13096 with an average value of 44.24 citations per document and 3.85 citations per year per document. Our collection is described by 721 different Keywords Plus (ID) and 925 Keywords assigned by Authors (DE). The latter terms are as useful as Keywords Plus in exploring the knowledge structure of scientific fields but also provide a more comprehensive representation of the article’s content (Zhang et al., 2016). In this study we carried out all of the analyses on DE. In order to preserve the authors’ choices in defining keywords for their articles, we decided to perform a soft pre-processing. We normalised only hyphens (i.e., market shaping/market-shaping) but we did not lemmatise keywords. Therefore, in the following analysis we do not consider as a single unit, terms expressed in singular/plural (i.e., network/networks, market/markets). By looking at the frequency distribution of keywords we select as representative keywords, those occurring more than four times in the collection.

Table 1 Main information about the collection
Full size table

4 Data analysis

All of the analyses were carried out with the open-source R package bibliometrix, an open-source tool for quantitative research in scientometrics and bibliometrics that includes all of the main methods for performance analysis and science mapping (Aria & Cuccurullo, 2017).

In this section we present the results of the bibliometric analysis. First, we show the main keywords; next, we conduct a two-period analysis (1990–2004 and 2005–2020) of the number of articles published and address that in the first fifteen years there were very few publications. For each period we look at the most quoted authors to identify the main themes. As a result of the small number of publications in the first period, all of our subsequent analysis are conducted over the whole time period. Specifically, to see how keywords are related as well as their importance, we construct a keywords co-occurrence network, and to identify both the clusters and their density and centrality, we develop a thematic diagram (see Fig. 7).

4.1 Keywords analysis

A word cloud is a unique way of visualizing text in which the words that appear more often are highlighted effectively by appearing more prominently in the illustration. The more frequently the keyword occurs in the text being analysed, the larger it appears in the graphic being created. Considering the whole period, we observed that the main keywords are marketing systems (29), macromarketing (16), market practices (15), market shaping (14), performance (13) and innovation (13), as highlighted in the word cloud (see Fig. 3). As we will explain in greater depth in the following sections, some marketing scholars have linked the market concept to a system view in a macro perspective (Layton, 2009a, 2009b). For example, such a view informs how to shape or innovate the market (Storbacka, 2019) and to develop practices (Mele & Russo Spena, 2015).

Fig. 3
figure 3

Word cloud of the twenty most frequent terms

Full size image

4.2 Time analysis

To detect the thematic evolution related to how the debate evolved over time, we analyse the articles and the most quoted authors published in two periods: 1990–2004 and 2005–2020. Figure 4 shows that there were sixteen market-related articles from 1990 to 2004, while the number greatly increased in the 2005–2020 time period, reaching 226 articles. Although the increase in the number of articles is not proof, it appears to support our rationale that the concept is still in a quest towards conceptual delimitation. Scholars recognise that what the market is and how it operates have yet to be fully resolved (Stewart & Zinkhan, 2006): “paradoxically the term market is everywhere and nowhere in our literature” (Venkatesh et al., 2006, p. 252).

Fig. 4
figure 4

Year-wise distribution of the 242 articles

Full size image

During the first fifteen years of our timeframe, we identify the most quoted authors to detect their main themes (see Fig. 5). Initially, scholars focus on market orientation culture (e.g. Homburg & Pflesser, 2000) and behaviour (e.g. Sundbo, 1997) as well as market-relating capabilities (e.g. Day, 2000) and market knowledge competence (e.g. Li & Calantone, 1998; Sinkula, 1994). Competitive advantage and market performance (e.g. Hooley et al., 2005) are also a strong concern. Later, an interest in a network perspective develops (Achrol, 1996), with an understanding of interactions, relationships, and stakeholders (Palmer & Quinn, 2005).

Fig. 5
figure 5

The ten most cited articles in the sub-period 1990–2004

Full size image

In the second period (2004–2020), the number of studies grows. Figure 6 shows the most quoted authors. Main terms become market practices (Kjellberg & Helgesson, 2006; Azimont & Arajuo, 2007) and market shaping (Baker et al., 2019). Studies seek to develop an understanding of market participants or actors (Hagberg & Kjellberg, 2015), value co-creation (Chandler & Vargo, 2011), institutions (Vargo et al., 2015), resources (Storbacka, 2019), systems (Domegan et al., 2019; Layton, 2009a, 2009b), value co-creation (Pohlmann & Kaartemo, 2017) and market propositions (Storbacka & Nenonen, 2011b).

Fig. 6
figure 6

The ten most cited articles in the sub-period 2004–2020

Full size image

4.3 Thematic analysis

In this section, we focus first on the relationships between keywords, unpacking the conceptual structure latent in the extant literature (Callon et al., 1983; Ravikumar et al., 2015). To this end, we plot a co-occurrence network of keywords to highlight both the importance of each and how they are related. Indeed, according to Kumar et al. (2021) “keyword co-occurrence [analysis can] indicate several noteworthy trends” (p. 282). Each research field or topic can be represented as a set of terms and “keywords converging into a cluster representing a common theme” (Mukherjee et al., 2022 p. 105). We perform this analysis on the 242 articles using the authors’ keywords. In Fig. 7, each node is a keyword, and its size is proportional to the frequencies of keywords. The thicker is the edge between two keywords, and the higher is the number of times that the two keywords co-occur. This means that a strong conceptual relationship exists. We highlight in different colours the topics (clusters of highly related keywords) identified by the community detection process.

Fig. 7
figure 7

Co-occurrence network of the most representative terms

Full size image

The co-occurrence network analysis was complemented with a thematic diagram.

Each theme, detected by means of community detection, can be visualised as a bubble in the thematic diagram (see Fig. 8). The terms in the bubble are terms with the highest number of occurrences within a cluster and the bubble size is proportional to the occurrences of terms in the cluster. Its position is based on Callon’s centrality and density, which express the theme’s importance and the theme’s development respectively. Here we can visualise five clusters together with their current developments: innovation/performance, marketing systems/macromarketing, institutions/markets, market/market shaping, market practices/market dynamics.

Fig. 8
figure 8

The thematic diagram of the cluster related to market

Full size image

The first cluster, Performance, is in the extreme right of the thematic map, with a very high density and centrality. This means that it is a motor theme in the literature. Terms belonging to this cluster are performance, market performance, competitive advantage and co-creation. Marketing system also is a motor theme, with high centrality and high density. This cluster contains contributions mainly from the research area of macromarketing. Main descriptors are value creation, consumer and consumption, market formation and entrepreneurship. The cluster Market shaping is a theme with low centrality and medium density. Its main words are market innovation, performativity and practice. Institutions is a smaller though emerging cluster with medium centrality and low density. Co-current words are markets, service dominant logic, marketisation, relationships and networks. The Market practices cluster relates to an emerging theme with quite low centrality and low density. Co-current words are market dynamics and market studies.

5 A taxonomy of market clusters

Having identified, in the co-occurrence network of keywords, the themes most likely to enrich the conceptualisation of the market; in this section, we describe each cluster qualitatively with regard to its understanding of the market. Each cluster can be seen as a “the conceptual subdomain of a research field (Pohlmann & Kaartemo, 2017, p. 54).

5.1 Market as performance cluster

The first cluster addresses the paramount role of innovation in achieving higher market performance (Achrol, 1996; Durmusoglu & Barczak, 2009; Kim et al., 2006). Markets are complex systems aimed at creating and coordinating knowledge as they evolve over time (Bleda & Chicot, 2020). Studies draw from the resource-based view (Gregory et al., 2019) to analyse markets with companies’ internal resources as well as “external resources embedded in a firm’s business network” (Zhang & Wu, 2017 p.3). In other words, market as performance is strongly linked to marketing capabilities (Day, 2000; Morgan et al., 2004 p.90), market knowledge competence (Li & Calantone, 1998) and marketing resources in contributing to the creation of a competitive advantage (Hooley et al., 2005). Specifically, marketing capabilities assume a key role in market competitive intensity (Kaleka & Morgan, 2019) as they allow “a firm to create superior customer value” (Guenzi & Troilo, 2007, p. 98) and to implement market strategie (O’Cass & Julian, 2003). Alternatively, market knowledge competence is “the process that generates and integrates market knowledge” (Li & Calantone, 1998, p.14), sclosely linked to organizational learning (Sinkula, 1994).

An example of this cluster's understanding of the market is the company Coca Cola (or most mass consumer goods companies). These firms develop innovation through the application of their competitive advantage (i.e., their market capabilities, resources and knowledge competencies) with the goal of achieving improved performance.

5.2 Market as systems cluster

The ‘market as systems’ approach is quite broad and does not have a unified perspective. A first stream, the macromarketing perspective, understands markets as social arenas (Layton, 2015) with “the understandings, languages and logics that people use in everyday life” (Layton, 2019, p.1). In such socially constructed arenas “embedded actors compete for material and status rewards” (Fligstein & McAdam, 2012, p.5). The different combinations of self-interest, morality and mutuality gives rise to divergences (Domegan et al., 2019). The conflicting and competing interests of actors are evaluated, manipulated and resolved, with the identification of threats and opportunities (Layton, 2015, 2019). In the macromarketing perspective, the marketing system has a key role as a consequence of “the historical framing of choices made by all participants, generating, through self-organization and emergence” (Layton & Duffy, 2018, p. 400).

A second research stream is related to investigating the role of consumer and consumption (Diaz Ruiz et al., 2020). These studies address how new markets formed as a consequence of consumption that from being mundane and insignificant has mutated into something socially identified (Ulver, 2019). Consumers, incidentally, modify a market, by extending its boundaries in relation to their social context (Ulver, 2019). Consumers' identity work helps to shape the market by giving new discursive and material interpretations (Ulver, 2019). Meaning and value are interwoven in social and market domains, characterising contemporary market creation (Peñaloza & Mish, 2011).

An example of this cluster's understanding of the market are firms such as Masterchef. These firms see consumption as a social identification, providing a new interpretation of meanings (in the MasterChef example, the meaning of food and eating goes beyond nutrition) shared by a set of participants within the larger marketing system.

5.3 Market as market-shaping cluster

Markets are complex systems that evolve through time, rather than as pre-existing stable structures (Mele et al., 2015; Nenonen et al., 2019). Markets are configurations of market actors (Storbacka & Nenonen, 2011a). These configurations are malleable (Kinsdstrom et al., 2018) or plastic (Nenonen et al., 2014) as new actors enter the context and bring new ideas and new business model elements to shape the market. Markets are what actors make them to be. They are “socially constructed artefacts, created by the actors who populate a specific context and link resources within this context” (Storbacka & Nenonen, 2012, p. 184). Subjective market definitions arise in terms of “market entities (referring to which actors are involved), market representation (regarding how markets are signified), market performing (referring to what actions are carried out), and market sense-making (concerning how markets emerge and evolve)” (Mele et al., 2015, p. 100). Market representation assumes a key role in depicting “what a market is and how it works” (Diaz Ruiz, 2013). Alternatively, market shaping activities (Baker et al., 2019; Kinsdtrom et al., 2018) and market shaping capabilities (Windahl et al., 2020) address the importance of “generat[ing] new intra- and inter-stakeholder resource linkages by directly influencing various characteristics of the market” (Nenonen et al., 2019, p. 617). The performative idiom directs attention to the social–material practices (Orlikowski, 2007) that agencies engage in to innovate markets. From this perspective, market innovation is “a social process of construction by a group of actors” (Mele & Russo Spena, 2015, p. 43) fuelled by a market learning process (Storbacka & Nenonen, 2015). Innovation is no longer the making of novel units of output, but the design and creation of new markets.

An example of this cluster's understanding of the market is Netflix. These firms design new value proposal representations to carry out social-material practices involving actors who increase their learning by being engaged.

5.4 Market as institutions cluster

This cluster has two different research streams. First, Service-Dominant logic (S-D logic) sees markets in terms of the benefits they provide (Vargo, 2011b). Markets and society are seen as a whole socially constructed process of creating conditions of “mutual value through mutual service provision” (Vargo & Lusch, 2011, p. 184). A “theory of the market is one of value co-creation through mutual service provision, made possible by resource integration” (Vargo & Lusch, 2011, p.186). S-D logic conceptualisations move towards a systems perspective of the market; that is, a market made up of complex, dynamic service ecosystems “characterized by mutual value propositions and service provision, governed by socially constructed institutions” (Vargo, 2011a, p. 220). This perspective clearly distinguishes markets as entities of “simultaneous, continuous exchange” from markets that “emerge from simultaneous, continuous processes at different levels and layers of context” (Chandler and Vargo, 2011b, p. 12). The identification of context allows framing of not only markets but also the interactions and relationships among actors in those markets (Chandler & Vargo, 2011). There are thus no a priori markets. Rather, markets are a function of combinatorial evolutions of value proposition as well as of actions and interactions ‘performed’ by actors engaged in practices that become institutionalised solutions (Vargo et al., 2015). A second group of studies focus on how formal and informal institutions affect the process of marketisation (Bajde, 2013; Wang et al., 2019). They see “the market as a socially embedded institution in which community ties are formed and sustained” (Varman & Costa, 2008, p. 141). Market discourse and social norms affect social cohesions within market-embedded networks (Wang et al., 2019).

An example of this cluster's understanding of ‘the market is the network’ are Uber or AirBnB. For these firms, customers' interactions and relationships affect value co-creation through changes to institutionalised solutions; for example, changes in the payment methods (i.e., through mobile Apps), or through the adoption of referral systems.

5.5 Market as practices cluster

This approach implies a shift towards practices (Azimont & Araujo, 2007; Finch & Geiger, 2011; Kjellberg & Helgesson, 2006; Rinallo & Golfetto, 2006). The market emerges from a network of transactions that link exchange practices (i.e. individual transactions), normalising practices (i.e. rules and forms for market behaviour), and representational practices (i.e. the structure and the functioning) (Kjellberg & Helgesson, 2006). It is the specific combination of practices that defines the exchange mode operating, with a specific focus on the retailing context. Accordingly, “markets should be studied as sites of multiple and often conflicting sets of practices, and attention should be directed towards the forms markets take as a result of efforts to shape them” (Araujo et al., 2008, p.6). The focus on the market dynamics stresses “the socially constructed nature of the link between the representation and its referent” (Rinallo & Golfetto, 2006, p. 857). A process of framing allows distinct actors to interact and agree on a price for the market exchange of goods and money (Callon, 1998). Markets are spaces of calculability (Azimont & Araujo, 2010) requiring activities that embed each exchange in a specific context (Araujo, 2007). In essence, market framing relies on “a variety of practices that construct a space of calculability” (Araujo et al., 2010, p. 7).

An example of this cluster's understanding of the market is Virgin Music. In this type of firm, actors moved from classical buyer–seller roles to downloading and listening practices. The exchange mode is completely modified and, subsequently, the rules and norms are also altered, allowing new actors to enter (e.g., iTunes, Spotify).

6 Discussion

The co-occurrence network and thematic diagram show five clusters, and our description in the previous section highlights that, at times, within a cluster, there is not always homogeneity. In this section, we bring together all of our analysis and descriptions to answer our two research questions. Which are the emergent themes associated with the understanding of the market? Which themes are reaching emergent stability?

As stated, we chose Sawyer’s (2005) social emergence model to help visualise the bottom-up effect of new themes associated with the (re)conceptualisation of the market, as these themes move from isolated efforts towards gaining robustness and support. When applied to the emergence of a conceptual system (Bunge, 2003) we argue that the co-occurrence network provides the means to elicit the themes while the thematic diagram provides a proxy to position the themes in the ontological levels of the emergence.

To answer our research questions, we start by pointing to the low number of articles prior to 2005, which shows that, as stated in the introduction, there was a long period in which the neo-classical economic theory’s view of the market was not problematised.

Next, we want to argue that two of the five clusters, the Performance and the Marketing system clusters, although different in their understanding of the market and set of keywords, are positioned, in the thematic diagram, in the high density/high centrality quadrant, allowing us to argue that they are at Sawyer’s institutionalised level. The next two clusters we will discuss follow different trajectories but, we will argue, are at the same emergence level. The Market shaping cluster has medium density and low centrality, meaning that although it has not fully developed its internal links, it is achieving a discrete level of echo within the marketing community; we will argue that this status positions it at the ephemeral emergence level. The Institutions cluster is in the opposite situation: with its medium centrality it is still in the process of developing its internal links fully, and its low density shows that it has not yet been adopted by the broader marketing community; thus, we will argue that it is also at the ephemeral emergence level. Finally, the Market practices cluster is in the low/low quadrant, suggesting that it is still at the interaction level as it is still seeking to gain both recognition beyond the core group and to develop itself fully. Figure 9 helps to visualise how the clusters are positioned within Sawyer’s ontological levels of emergence.

Fig. 9
figure 9

Clusters positioned within Sawyer’s ontological levels of emergence

Full size image

To summarise, our bibliometric study answers our research questions by introducing three new themes, none of which has yet reached the emergent stability level as, for this level, both robustness (density) and transversal adoption (centrality) are required. We will argue that these results are consistent with the fact that the problematisation of the conceptualisation of the market, within marketing, is only about fifteen years old. The increase in the number of articles seems to indicate that this is a very dynamic situation and that the three new clusters will evolve on the emergent path.

7 Theoretical contribution

We agree with Harrison and Kjellberg’s (2014, p. 2) argument that theories of markets matter, because definitions, conceptualisations and theories-in-use shape actors’ understandings and scope for influence (Geiger & Finch, 2010; Henneberg et al., 2010; Johanson & Mattsson, 1992), and because market theories form the basis for regulatory efforts and other forms of interventions in markets (MacKenzie, 2006; Callon, 2007; Pels et al., 2022).

Our analysis provides three contributions to the marketing literature conversation on markets. First, our time analysis of the number of articles confirms the recent dynamism (i.e., × 14) in our domain of study, the (re)conceptualising of the market as well as a shift in main authors and keywords adopted to discuss the market. Second, our co-occurrence network of keywords enabled us to elicit five thematic clusters: performance, marketing systems, market shaping, market practices and institutions, while the thematic mapping offered the means to visualise their positioning in terms of their density and centrality. Both the clustering and the mapping enables the marketing discipline to situate the current state-of-the-art of the diverse understandings of the market; we shall come back to this aspect in the future research. As Donthu et al. (2021b) note, identifying network clusters offers “understanding [of] how a research field manifests and develops (p. 291).

Third, by adopting Sawyer’s (2005) social emergence model, we offer a criterion to position the five clusters in terms of their level of emergence; as stated, two clusters (Performance, Marketing systems) are situated at the institutional level; two clusters, though following different trajectories, are positioned at the ephemeral emergent level (Market shaping and Institutions), and the fifth cluster (Market practices) is still at the interaction level.

The multiplicity of emergent themes as well as their positioning in the lower levels in Sawyer’s (2005) emergence model, validates our initial argument that we are at the initial stage of the process of (re)conceptualising the market and remain far from a Kuhnean shift in paradigms. However, the ascending number of articles confirms that a process to contest the traditional understanding of the market is in action. As stated, we agree with Lather’s (2006) support for a proliferation of approaches and contend that it is beneficial to have a multiplicity of emerging themes; particularly, in these early phases where the creative tension of a rich conversation can only improve the outcome.

8 Limitations and future research

This work contributes to the marketing discipline by providing an overview of the leading trends in studies focusing on markets. However, it is worth noting several limitations which invite expanding on this initial effort.

First, in relation to the source of our keywords: we focus on marketing’s fifteen top journals and a more exhaustive list of journals could be contemplated. Our dataset consists of journal articles, based on the assumption that these amount to the frontier of research. However, further additional materials such as books and conference papers may also provide evidence of knowledge transfer and are worthy of future investigation. Further, with regard to the choice of WoS as bibliographic database: Scopus or Google Scholar can be considered other equally valid alternatives to WoS and in the last few years there has been an intense debate concerning which source is better (e.g., Bar-Ilan et al., 2007; Harzing & Alakangas, 2016). Scopus, for instance, considers a wider timespan than WoS. Retrieving articles published before 1985, and then merging the different sources, is an attractive possibility to extend the timespan and enrich the number of articles focusing on markets. In this case, the analysis of textual contents would have to be performed on titles or on the overall content of the articles, because keywords and abstracts are generally not indexed in documents published before the nineties (Liu et al., 2015).

Second, with respect to the keywords employed, we chose twenty-three market related keywords. As stated in our data collection sub-section, keywords independently of their association to the term ‘market’ could have been employed, such as ‘networks’, ‘service ecosystems’ or ‘systems’, just to mention a few. Different keyword searches when undertaking the database searches might also have affected the results. These are some suggestions for possible expansions of the scope of the research.

Third, in this paper, we focused on highlighting the conceptual structure of the scientific literature related to the market by considering author’s keywords as a unit of analysis. However, a deeper investigation could be done by analysing other textual components of bibliographic databases, such as the titles of articles or abstracts. For this study we decided to perform a soft pretreatment on keywords by preserving the authors' original choices, but different preprocessing procedures could be considered (i.e. stemming, lemmatisation) (Vijayarani et al., 2015).

Fourth, as stated, our unit of analysis was the keywords, but an alternative option could have been developed. By means of science mapping and performance analysis techniques, it is also possible to investigate both the social and the intellectual structure of a research field (Noyons et al., 1999). The first aim can be reached by analysing collaboration networks among authors, institutions and countries, highlighting how they interact with each other. The intellectual structure aims at visualising the connections among sources, publications or authors, through bibliographic or co-citation coupling networks to suit the aim of the analysis (Marshakova, 1981). Future research can be addressed to these issues to obtain a comprehensive map of the actors involved in the re-conceptualisation of the market. For example, Pohlmann and Kaartemo (2017) identified two seminal papers and then studied their research trajectory. On a similar basis to our own study, for each cluster the seminal papers could be identified and their research trajectories explored. Another example of alternative units of analysis is the study by Jia et al. (2018) who identify key authors and their collaboration networks. This approach could be an interesting complement to this study as it would facilitate understanding the causes of the low and medium–low centrality in the three emerging themes. Is there high endogeneity?

Fifth, leaving aside the research opportunities opened by adopting different methodological choices, there are future research questions that could be addressed as a complement to or a result of this study. For example, (1) our focus was on plotting the field by identifying emergent themes and their degree of robustness and transversality. Future papers could focus on the interconnectedness of disciplines—marketing is nurturing itself from other disciplines such as sociology and economics—but are the concepts developed by articles in our discipline being adopted in other fields of research? (2) Another line of study could be related to the emergence model. Sawyer (2005) highlights the downward effect between all levels. An interesting research question could be: are the authors or concepts in the established clusters acting as gatekeepers to the detriment of the emerging ones? When looking at the time analysis and the two periods studied, new keywords began to be adopted (e.g., from market-related capabilities to market shaping). (3) A further stream of study could be related to the types of keywords. It could be argued that the emerging keywords and themes support a more subjective understanding of the market. It could be interesting to follow up our work with an investigation of the ontological nature of the keywords adopted to understand the market. (4) Finally—but this is by no means the last possible stream of study—a comparison between approaches could be conducted. As we saw, there are keywords as well as concepts that are present in more than one cluster (e.g., innovation, practices, networks) but they have developed with a different twist. It would be interesting to explore similarities and differences between the five clusters.