A consumer engagement systematic review: synthesis and research agenda Una revisi on sistem atica del compromiso del consumidor: síntesis y agenda de investigaci on

Purpose – This paper aims to review the existing literature about consumer engagement, provide an accurate mapping of this research field, propose a consumer engagement typology and a conceptual framework and offer a research agenda for this domain. Design/methodology/approach – A systematic literature review using several quality filters was performed, producing a top-quality pool of 41 papers. After that, a text mining analysis was conducted, and five major research streams emerged. Findings – This paper proposes five distinct research streams based on the text mining analysis, namely, consumer engagement, online brand community engagement, consumer-brand engagement, consumer engagement behaviours and media engagement. Based on this, a consumer engagement typology and a conceptual framework are suggested and a research agenda is proposed. Originality/value – This paper presents scientific value and originality because of the new character of the topic and the research methods used. This research is the first study to perform a systematic review and using a text-mining approach to examine the literature on consumer engagement. Based on this, the authors define consumer engagement typology. A research agenda underlining emerging future research topics for this domain is also proposed.


Introduction
Studies on engagement emerge in scientific research connected to distinct fields of study such as education and learning (Kearsley and Shneiderman, 1998), sociology (Jennings and Stoker, 2004), psychology (Achterberg et al., 2003) or psychology and management (Csikszentmihalyi, 1997). In management, engagement emerges in several studies on individuals' application to workplace tasks and environment, which can be perceived as an experiential state that conveys personal, engaging behaviours (Billett, 2001;Schaufeli and Bakker, 2004). In the marketing field, Algesheimer et al. (2005) develop some seminal work on community engagement, highlighting the role of brand community commitment in consumers' attitudinal loyalty and repurchase intentions, claiming that consumers' motivation to interact with others will positively influence engagement with the brand community itself. Other authors propose engagement as a customer's level of physical, cognitive and emotional presence in their relationship with a service organisation (Patterson et al., 2006).
More extensive literature appears around 2010. This new research suggests innovative perspectives connecting engagement with its behavioural dimensions (van Doorn et al., 2010;Kumar et al., 2010;Verhoef et al., 2010), psychological processes (Bowden, 2009) or even multi-dimensional perspectives of consumer engagement Hollebeek, 2011aHollebeek, , 2011b. From this moment onward, a myriad of research arises about this topic. Calls for further researchas a priority research areaare made by distinct organisations, such as the Marketing Science Institute (MSI, 2014(MSI, , 2016, leading the scientific community to devote their time and attention to this subject. Fast-forward to nowadays, with extensive literature on consumer engagement, which by itself justifies a literature review and serves as the basis for the aim of this research. This paper aims to understand the state-of-the-art of the conceptualisation of consumer engagement in the marketing area and to propose a consumer engagement typology and a conceptual framework. Finally, a research agenda for this domain is suggested. Academics are still struggling to define the influence of the consumer engagement construct despite the valuable contributions of past research van Doorn et al., 2010;Hollebeek et al., 2014). This provides fertile ground for our research question, which is: what are the major research streams of consumer engagement? By answering this question (using a systematic literature review and text mining approach) we claim that it is possible to go further and propose a consumer engagement typology and a conceptual framework, which the authors assume as a relevant output for this paper.
The systematic review follows a methodological and systematic process to identify and critically assess the relevant body of research, providing a top-quality group of 41 papers, from which we generate collective insights of knowledge. Additionally, a text-mining analysis based on the systematic review outputs is performed, and five major research streams emerge: (1) consumer engagement; (2) online brand community engagement; (3) consumer-brand engagement (CBE); (4) consumer engagement behaviours; and (5) and media engagement.
Based on this analysis, we suggest a framework for consumer engagement. The remainder of this paper is organised as follows. The next section describes the method used to collect and analyse the data and proposes a framework. We continue with an overview of the significant research streams on consumer engagement and the conceptual findings are discussed. Finally, we present a section dealing with the emerging research topics and a proposed research agenda, followed by the conclusions of this paper.

Systematic literature review 2.1 Method
Following previous studies (Ceipek, 2019;Crossan and Apaydin, 2010;Galvagno and Dalli, 2014), we adopt a systematic approach to the literature review to make the results reliable, verifiable and reproducible. The process may be defined as a literature review method performed through an automated search to collect relevant information and identify, substantiate and conceptualise all fundamental research streams on the topic under study (Moher et al., 2009).
A collection of the relevant literature addressing consumer engagement is extracted based on two distinct database aggregators that are Web of Science (WOS) and Scopus online libraries. This procedure ensures the presence of the most relevant papers published in top journals, also comprising the most relevant publishers, such as Emerald, Elsevier, IEEE, Springer, Sage, Taylor and Francis or Wiley. The same search terms are used in these two databases. The selection and definition of the terms to be used is of foremost importance. The authors define the five different research terms from the existing literature to cover, as much of possible, all the relevant topics in this field of research. The terms used are, namely, "consumer engagement"; "customer engagement"; "customer brand engagement"; "brand engagement"; "online engagement". Table 1 shows the final set of search terms.
Regarding quality and credibility assessment, this research intends to use only cuttingedge research, i.e. the best quality papers published in the best journals. Engagement in the marketing field is still in the early stages of development, which may lead to new research Synthesis and research agenda being published first in top-tier journals. To assess quality and credibility, we first apply some generic filters to the databases, these being: full-text papers; peer-reviewed papers; and scholarly journals.
Secondly, we adopt a selection of quality journals. We use the "journal quality list -64 edition", compiled and edited by Professor Anne-Wil Harzing (Anne-Wil, 2019). From this list, we apply the WIE 2008 Rating -WU Wien Journal Rating May 2008 List. The list includes only A and Aþ journals, and from that list, we consider only the journals inside the scope of the study. Thirdly, we select journals devoted to the topic of marketing, general management and strategy and international business. With this option, we intend to cover as many different areas of research as possible that could address this topic. Concerning the period to be examined by the systematic literature review, we search for all publications available until June 2019. On reaching this stage, the methodological procedure evolves to content validation. The first assessment is based on the title, and all papers that demonstrably do not address the topic of engagement are excluded. The second assessment is made by reading the abstract, and each paper is evaluated based on its scope. The researchers carefully analyse what constructs are dealt with by each paper, what scientific research paths they cover and their adherence to the scope of the study. Finally, the last phase of our content analysis involves reading the full text. To measure its relevance, we choose to apply two questions that each paper should address. Only the ones that answer affirmatively to the following questions move forward to the final group in this systematic literature review: Does the paper address our well-defined scope of study? Does the paper contribute to developing the engagement construct (or one of its variations, such as customer engagement, consumer engagement, brand engagement, customer-brand engagement or theory of engagement)? The flow chart presented in Figure 1 illustrates the steps undertaken in this systematic literature review. Figure 2 presents the results of the aggregate outcome of the search strategies, showing the aggregate results of all the screening processes. From the final list of resultsbefore fulltext reading -6 papers are found to be repeated, and therefore, withdrawn, reaching the final number of 61. After the last screening process (full-text reading), a final group of 41 papers emerges ( Figure 2). Finally, to achieve consistency and reduce subjectivity, the final aggregated result of this systematic literature review is checked and rechecked separately by the researchers, as suggested by the relevant literature in this domain (Moher et al., 2009). Table 2 shows the final list of papers by authors, year of publication, journal title and number of citations on WOS and Scopus.

Data extraction form
With the final group of papers, a data extraction form systematises and analyses the relevant information. This process allows us to achieve several purposes: document the selected research; reduce the possibility of human error; decrease the use of subjective-driven data selection criteria; reflect on the differences and compare the similarities of the main ideas of each paper; and support other researchers' work that may be of interest in this research field (Athanasopoulou, 2009).
The data extraction form comprises eight categories. These categories summarise the core information of each study analysed: authorship and year of publication, journal of publication, the base theory of the study or construct(s), adopted methodology and sample (with characteristics), main conclusions, limitations and future research and managerial implications. The systematised data made available in this extraction form allow us not only to clarify the distinct concepts but also show the major research streams of consumer engagement.

Text-mining analysis and clustering
Based on the data extraction form, a text-mining procedure is performed to find the latent topics in the literature review, a widely used procedure to reveal latent topics in scholarly     Synthesis and research agenda literature (Abbas et al., 2014;Loureiro et al., 2019;Moro et al., 2015). A text clustering techniquebased on the MeaningCloud text mining toolhas the potential to process and aggregate a large amount of unstructured text to extract relevant information (Fan et al., 2006). This technique can create groups (clusters) by analysing the text of each paper (Fan et al., 2006;Srivastava and Sahami, 2009). Hence, texts in the same group (clusters) are more similar to each other than to texts in other clusters (Spinakis and Chatzimakri, 2005). The text-mining system classification for cluster sizes is used to select the appropriate number of clusters (Punj and Stewart, 1983), and no new clusters are accepted when their size is too small (Milligan and Cooper, 1985), which leads us to consider the five largest clusters (cut-off value: 200). The clusters are labelled as follows: consumer engagement (score = 497.82), CBE (score = 373.36), online brand community engagement (score = 291.22), consumer engagement behaviours (size = 239.55) and media engagement (score = 216.04). Based on these clusters, a consumer engagement typology and conceptual framework are proposed ( Figure 3).

Consumer engagement typology and conceptual framework
The consumer engagement construct has evolved and has been redefined over time, generating considerably different definitions, concepts and arguments used to define the construct. However, to date no unifying conceptualisation exists in the literature. Consequently, and based on the clusters mentioned above, this study presents a consumer engagement typology. By integrating the distinct concepts, the authors propose a definition for consumer engagement typology, as the aggregation process of different concepts and dimensions of engagement that acts as intra-interaction causes towards a focal object (i.e. product/brand), which all together or separately, lead consumers' interactivity, which will foster different benefits for the focal object. Accordingly, we develop a conceptual framework for consumer engagement and its key consequences for consumer-interactivity and benefits ( Figure 3). The framework is read as follows. Based on a consumer engagement typology, consumers' interactivity efforts, such as buying (direct effort), referring, influencing and giving feedback (indirect effort) experience the influence of the distinct forms of consumer engagement (i.e. consumer engagement, CBE, consumer engagement behaviours, media engagement and online brand community engagement) to create effects, described here as benefits. These benefits can be of two types, tangible, such as a firm's performance (e.g. sales, profits or market share) or intangible (e.g. opt-in. privacy sharing or relevant marketing activities). Additionally, the proposed framework should not be a static representation connecting parts, but rather a dynamic tool with several variations and contexts. For instance, consumers with a specific type of interaction based on a functional motive (such as buying) can find themselves enjoying the content created by the focal engagement object (i.e. product/brand), leading to a higher expression of their behavioural engagement than those who act solely for a functional motive (Kim et al., 2016).
Interacting with the focal object prompts a particular level of consumer engagement (Schamari and Schaefers, 2015). Thus, the consumer engagement typology acts as an intrainteraction process that will, in turn, foster-related benefits. For example, consumers interacting with a specific brand (and brand content) will invest their resources in those interactions (efforts), which, in turn, develops brand-related sense-making (Hollebeek and Macky, 2019). Consumers' brand-related sense-making will then drive the development of consumer engagement consequences of tangible and intangible benefits, as shown in Figure 3.
By analysing the five consumer engagement concepts that are core in our framework, one can argue that consumer engagement can overlay on various shapes. Firms are focussing on the quality of the relationship that they set up with consumers and maximising the benefits beyond purchases that consumers can provide to their brands (Badenes et al., 2019;Prentice and Loureiro, 2018;Rather et al., 2018). Consumers are willing to contribute to the firm when they feel connected to the brand (CBE), which is a central component of the consumer engagement construct (Pansari and Kumar, 2017). This idea notes that during an interaction, partners may create products for each other or they communicate with each other (media engagement). This idea is becoming clear in the online interactive environment where consumers provide feedback to the company, for the development of new products or improvements to existing ones (online brand community engagement) Rather et al., 2018;Rosado-Pinto et al., 2020). The value of the association between consumers and brands depends on the level of satisfaction resulting from the customer's level of emotional connectedness and the motivations towards this relationship van Doorn et al., 2010). When a brand achieves an emotional relationship with the consumer, we may say that the firm and the customer are engaged with each other (consumer engagement behaviours). Therefore, we argue that it is of foremost importance to understand and further explore the construct of consumer engagement, as engaged consumers can create several direct and indirect benefits (Kumar and Pansari, 2016).
Overall, the proposed framework outlines the process by which consumer engagement creates value through a series of consumer-based efforts and effects. While these concepts will sequentially interact, other antecedents may continue to co-exist with their respective consequences. So, in-depth research, including further conceptual development and empirical testing, is required to assess and validate our proposed framework.
Taking all the perspectives provided by past research, we may claim that the concept of engagement is multi-dimensional, having two major theoretical foundations, namely, servicedominance (S-D) logic and relationship marketing. As a unifying conceptualisation of consumer engagement, we argue this is a motivational and relational state (comprising cognitive, emotional and behavioural aspects) that may occur between two agents, one being the consumer(s) and the other a brand, firm or organisation, which can operate online, offline or both.

Major clusters and key research streams
A growing stream of research continues to contribute to the consumer engagement construct (Bowden, 2009;van Doorn et al., 2010;Kumar, 2013;Loureiro et al., 2019;Rather et al., 2018;Vivek et al., 2012), creating a valuable amount of knowledge on this topic but also generating considerably different definitions, concepts and arguments used to define the construct. This paper gives an overview of the engagement concepts (Table 3), providing a clear contribution to this research field.

Consumer engagement
Despite no consensus on defining the concept (Barger et al., 2016), some aspects appear constant throughout the literature. The first is that consumer engagement is as an attempt to distinguish consumer attitudes and behaviours that go beyond purchase behaviours (van Doorn et al., 2010;Vivek et al., 2012). The second relates to the consumer's role in the concept, which is taken to be active, involved and co-creational, as opposed to being a mere recipient of initiatives from firms Gambetti et al., 2016;Kumar et al., 2010).
Researchers also tend to use different perspectives to conceptualise the consumer engagement construct (Schultz and Peltier, 2013). Focussing on an attitudinal perspective,  discuss the role of interactive consumer experience and co-created value as the basis for the construct (Table 4), addressing the nature of consumers' interactive engagement experiences across contexts, the value of consumers' dynamic and interactive cocreation role in service fields and the multi-dimensional essence, through its cognitive, emotional and behavioural dimensions. So, they define consumer engagement as a psychological state that occurs by interactive, co-creative customer experiences with a focal agent/object (e.g. a brand) in focal service relationships .
In the same line, other researchers also propose three dimensions of consumer engagement (Hollebeek et al., 2014), which correspond to the generic cognitive, emotional and behavioural nature of "engagement". The terms that designate each dimension are cognitive processing (consumer's level of relationship with a brand through processing and elaboration in a particular consumer/brand interaction), affection (degree of positive affective-relation with a brand) and activation (level of energy, effort and time spent on a brand). Vivek et al. (2012) draw attention to the importance of individuals' interactions and connections with the brand or product, regardless of whether they are in "buyer-mode" or not. These authors propose a broader vision of multi-dimensional consumer engagement through the cognitive, emotional, behavioural and social dimensions. The cognitive and emotional components incorporate consumers' experiences and feelings, and the behavioural and social elements include consumers' participation in the brand or product. Thus, they define customer engagement as the intensity of an individual's participation in and connection with an organisation's offerings or organisational activities, which either the customer or the organisation initiates (Vivek et al., 2012). This research path has grown in importance in the digital world as customers become increasingly active co-producers or destroyers of value for companies (Beckers et al., 2014;van Doorn et al., 2010;Leeflang et al., 2014;Verhoef et al., 2010).
The research related to customer engagement has evolved into several complementary pathways. Some authors focus on how firms can benefit from customer engagement (Alvarez-Mil an et al., 2018;Pansari and Kumar, 2017), others on how to measure customer engagement (Brodie et al., 2013;Calder et al., 2016;Hollebeek et al., 2014;Kumar et al., 2019) or even on distinct aspects of customer engagement value, such as customer purchasing  Multi-dimensional through two tenets, namely, cognitive bonds that enrich the product experience; psychological ownership (task-based) and self-transformation (experiential) that is beneficial to the firm Online brand community engagement Algesheimer et al. (2005) Positive influences of identifying with the brand community through the consumer's intrinsic motivation to interact/co-operate with community members Multi-dimensional: utilitarian; hedonic; social Raïes et al.
Community engagement as a formative construct. The frequency and duration of participation in the community, as well as the level of activities performed by the members in the community indicate the behavioural engagement of community members Multi-dimensional: affective commitment; calculative commitment; normative commitment; (continued) Online brand community engagement is the compelling, intrinsic motivations to continue interacting with an online brand community

Consumer-brand engagement
CBE emerges in the domain of relationship marketing (Palmatier et al., 2006) and, more firmly established, in consumer-brand relationships (Fournier, 1998). Different authors have made efforts to reach a single definition of CBE (Gambetti and Graffigna, 2010;, but a common and inclusive explanation of CBE still needs to be settled. Some authors argue that the meaning behind the expression represents the brand's effort to get close to the consumer in an expressive and meaningful way (Graffigna and Gambetti, 2015). Different definitions of CBE generally regard engagement as a multi-dimensional concept, typifying a consumer's interactive state based on cognitive, affective and Concept Author(s) Definition Dimensions \ tenets Jaakkola and Alexander (2014) processes that occur in interaction with the focal firm and/or other stakeholders, thereby affecting their respective value processes and outcomes behaviour; influencing behaviour; mobilizing behaviour Groeger et al.
Customer engagement behaviours as a non-paying consumer's motivationallydriven, positively-valenced behavioural manifestation towards a product, brand or firm, which are predicated on consumption of free offerings

Synthesis and research agenda behavioural components.
Hollebeek's definition includes those components when she defines customer-brand engagement as the level of a customer's cognitive, emotional and behavioural investment in specific brand interactions (Hollebeek, 2011b). Nevertheless, not all authors consider the three components. Some of them (Goldsmith et al. (2010) and Wang (2006)) focus only on the cognitive components of CBE. Indeed, Wang proposes that engagement in the advertising environment may be a measure of contextual relevance, where a brand message is enclosed and presented based on its surrounding context. Other studies underline the affective component of CBE. Heath (2009) suggests CBE as the amount of subconscious feeling going on when an advertisement is being processed. This author presents an emotional engagement model that highlights how strong brands can be built via emotional creativity rather than using rational messages, and without the high levels of attention that advertising usually needs.
Various academics contribute to this discussion with their perspectives of behavioural dimensions, which appear particularly diversified in their extent. Kumar et al. (2010), for instance, have drawn three fundamental behavioural components of customer engagement with a firm, namely, customer's purchase behaviour, customer referral and customer's influencing behaviour towards other customers. Verhoef et al. (2010), for their part, have defined customer-brand engagement as a consumer's behavioural manifestation towards a brand or firm that goes beyond transactions. In a more diversified approach, other authors conceptualise CBE dimensions from a more dynamic viewpoint, underlying the processbased nature of the concept. Thus, Bijmolt et al. (2010) suggest that several expressions of customer engagement may be generated in different stages of the customer lifecycle. Bowden (2009) defines customer engagement as a psychological process that forms the subjacent mechanisms by which customer loyalty is created in new customers and the mechanisms by which loyalty may be maintained for consumers' repurchase of a service brand. Bowden (2009) also proposes a conceptual model of the customer engagement process, which highlights the sequential development of loyalty, as customers evolve from being new consumers to re-purchasers.
Considering the articles published until 2010, we find that they propose conceptual models and conceptualisations (deductive approach) based on previous literature, and little empirical research has been conducted to gain new insights leading to supporting or rejecting the proposed theories and conceptualisations. Even so, this theoretical discussion has been essential in starting to describe a more realistic conceptual framework. The study by Gambetti et al. (2012) is one of the first to go further in trying to shed light on the meaning of engagement based on the practitioner perspective. They use the grounded theory approach to understand the CBE construct in an exploratory qualitative study. Gambetti et al. (2012) interview practitioners to understand how they conceptualise and perform this construct in day-to-day marketing activities. From this first study, they find out that practitioners see CBE as a dynamic, process-based concept, growing in intensity according to the brand's ability to increasingly understand consumers' desires and expectations, using all possible physical and virtual points of contact between brands and consumers. They also introduce this expression as a comprehensive marketing concept containing several consumer decision-making dimensions, from brand preference to brand purchase.
CBE is now regarded as a multi-dimensional constructcomposed of cognitive, emotional and/or behavioural dimensionsrepresenting a psychological state that occurs through the features of interactive and co-creative consumer involvement with a specific brand Machado et al., 2019;Vivek et al., 2012). The CBE construct is also broadly recognised as a dynamic relational process presenting multiple phases (Bowden, 2009;Hollebeek, 2011a), covering physical proximity and based on the progressive value of the brand for its consumers (Gambetti et al., 2012).
In this evolutionary process, several other studies have taken academic research on CBE to a new level, leading to unexplored territories. The research of Hollebeek, Glynn and Brodie (2014) regard customer-brand engagement as a consumer's positively valenced brandrelated cognitive, emotional and behavioural activity during or related to focal consumer/ brand interactions (Hollebeek et al., 2014). These authors point out that the CBE construct has relevance in a theoretical, interactive, consumer-brand relationship (which includes consumer culture theory, S-D logic and relationship marketing) ( Table 4). They contribute to the development of enhanced insight into consumers' specific cognitions, emotions and behaviours during brand interactions, which may be used to rethink or redesign the nature of specific consumer/brand interfaces to increase effectiveness. They also highlight that this concept appears to be very useful for both scholars and practitioners when looking to predict specific consumer behaviour results and that CBE represents a promising and underexplored concept in the literature.
Other types of contributions also arise. De Villiers (2015) proposes a new perspective on CBE by offering an expanded typology of brand fans or advocates, which includes negatively engaged, disenfranchised or even alienated consumers. In this study, the author presents a scale of CBE to measure a multi-dimensional model of engagement. He suggests a model, which contributes with an asymmetrical, configurational view of CBE that considers multiple paths to consumer enmeshment (more confluent to reality). Indeed, this work proposes the development of a more predictive (and at the same time more comprehensive) model, for configuration of the necessary or sufficient antecedent conditions that may cause high levels of CBE (and enmeshment). Gambetti et al. (2016), in turn, find that much of this engagement process is being experienced in a different dynamic relationship, with less focus on consumers and far more focus on brands. Indeed, these authors encourage firms and brands to reverse their perspective, adopting a pure logic of co-creation in the consumer-brand dyad and replacing the univocal perspective that still exists, which treats consumers as the ultimate static receiver of their initiatives. Thus, we can underline the engagement construct as a flexible, interactive space where consumers are recognised as central in the consumer-brand encounter. Kim et al. (2016) develop the theory and practice of advertising effectiveness through surfacing and decoding brand-consumer engagement over narrative transportation. These authors argue that consumers from different countries and cultures create similar narrative themes for the same luxury brands and are just as likely to engage in co-creation of advertisement narratives. Consequently, they suggest that powerful messages to consumers can provide a valid route for co-creation and brand engagement, mainly if these messages can project consumers into the storylines.
4.2.1 Brand engagement in self-concept. Sprott et al. (2009) focus their research on brand engagement in the self-concept (BESC), trying to demonstrate the nature and importance of the concept and suggesting that consumers can create associations between brands and their selfconcepts. Their empirical findings show BESC affects brand-related consumer constructs, including brand knowledge, attention, preference and loyalty. The authors also claim: a substantial variation among consumers in their level of engagement with their favourite brands; that firms may benefit from understanding whetherin addition to other relevant criteriaa brand's current target segments are lower or higher in BESC; and that managers should adapt their brand communication strategies accordingly. Baldus et al. (2015) attempt to conceptualise online brand community engagement as the compelling, intrinsic motivations to continue interacting with an online brand community. They propose a scale, which includes 11 dimensions based on previous studies, namely, brand influence, brand passion, connecting, helping, like-minded discussion, rewards (hedonic), rewards (utilitarian), seeking assistance, self-expression, up-to-date information and validation. However, the manifestation of particular cognitive, emotional and behavioural dimensions depends, to a great extent, on the engagement actorsengagement subjects/objectsand contexts , including media contexts (new online media in contrast to traditional advertising media) (Calder et al., 2009). Raïes et al. (2015) claim that intensive engagement in the community does not lead directly to brand loyalty behaviour and that higher levels of commitment combined with active engagement in the community are sufficient but not necessary for brand-loyal intentions. The commitment patterns of highly brand-loyal newcomers and longstanding members are also distinguished (calculative, affective and normative commitments differ in the length of membership) and a strong commitment to a consumption community can lead to high brand loyalty behaviour with or without intensive engagement in community activities. Raïes et al. (2015) also highlight the differences between own-brand communities and online brand communities created by consumers.

Online brand community engagement
Other studies underline the interactive, experiential process, based on consumers' engagement with specific products or brands and other brand community members (Baldus et al., 2015;Brodie et al., 2013;Gambetti et al., 2016;Hanson et al., 2019;Hollebeek and Macky, 2019;Monferrer et al., 2019). These authors also discuss how members engage in specific online brand communities. Schivinski et al. (2016) demonstrate that consumers engage in higher levels of online brand-related activities as a result of the learning process. They identify the consuming dimension as an antecedent of consumers' engagement in the contribution of brand-related content on social media, and this contributing dimension acts as an antecedent of user-generated content. For non-communities' online engagement, Hollebeek et al. (2014) devote attention to the social media context, underlining the need to rethink or redesign the nature of specific consumer/brand interfaces to increase online engagement and not online brand communities.

Consumer engagement behaviours
Some authors consider the consumer engagement concept from a more behavioural perspective (van Doorn et al., 2010;Kumar and Pansari, 2014;Kumar and Reinartz, 2016). Indeed, van Doorn et al. (2010) introduce the consumer engagement behaviour concept, defined as the customer's behavioural manifestation towards a brand or firm, beyond purchase, resulting from motivational drivers. These authors highlight that consumer engagement behaviours result from motivational drivers, including word-of-mouth activity, customer-to-customer interactions (C2C) and/or blogging activity. They propose five dimensions for customer engagement behaviour, namely, valence, the form of modality, scope, nature of its impact and customer goals.
In a firm value perspective, studies continue to suggest that consumers can generate value for firms through several behavioural contributions besides purchase (Kumar et al., 2010), such as customer lifetime, referral and influencer value, which lead to the customer engagement value construct (Kumar and Pansari, 2014;Kumar and Reinartz, 2016). Further research has also emphasised behavioural dimensions. Verhoef et al. (2010), for instance, regard consumer brand engagement as a consumer's behavioural manifestation towards a focal object (i.e. brand or company) that goes beyond transactions. This perspective is more aligned with the practitioners' point of view about engagement, which pays increasing attention to aspects such as C2C interactions (i.e. word-of-mouth, e-word-of-mouth; recommendations, reviews, etc), in an interactive environment . Calder and Malthouse (2008) discuss the concept of media engagement. These authors focus on the consumer's psychological experience while consuming media, and distinguish media engagement from mere liking, implying that engagement is a stronger state of connectedness between the customer and the media than merely liking. Afterwards, the same authors (Calder et al., 2009) discuss the engagement concept linked to the online environment. They suggest two distinct types of engagement, namely, personal engagement and social-interactive engagement, which can be described as taking part in discussions and socialising with peers through a website. In this perspective, online media cover a distinct form of engagement with an impact on advertising effectiveness, and they propose that the principle of engagement can be extended to other media, such as mobile media or social media (Bilro et al., 2018;Brandão et al., 2019;Gligor et al., 2019;Manser Payne et al., 2017). Moreover, engagement is considered as a second-order construct when considering measurement models, the prevailing experience being the first-order construct. This position will be refuted later by several other authors in subsequent research (Baldus et al., 2015;.

Media engagement
Other advancements in knowledge about engagement in the online environment are suggested by Bijmolt et al. (2010). These authors conduct an overview of available data for consumer analytics and propose distinct models for studying consumer engagement based on three phases, namely, customer acquisition, customer development and customer retention. They also claim that an increasing part of consumer behaviour towards brands or companies occurs in the online environment, which results in new sources of data for consumer engagement research (Bijmolt et al., 2010). In this study, the researchers also underline six barriers to the implementation of customer analytics, related to data - (1) data size, data quality, new types of data; (2) data ownership, to models -; (3) complexity of models; (4) ownership of modelling tools -; (5) usability of the results; and (6) integration of customer analytics in organisations' processes.

Emerging future research topics and research agenda
This paper proposes an integrated overview of consumer engagement that can be useful for both practitioners and academics in future endeavours. We now discuss the areas that appear to be gaining attention in relevant literature so far and highlight key areas for future research. A summary of these areas is proposed, providing a research agenda towards better understanding of consumer engagement. Firstly, future research needs to get new insights and developments concerning consumer engagement knowledge. Many studies in this research field are devoted to theoretical analyses. This fact renders a need for future empirical research to explore this subject further. For instance, qualitative research may reveal distinct triggers and outcomes of the benefits of consumer engagement, as suggested by our proposed framework (Figure 3). Additionally, quantitative research could focus on the tangible, intangible or both types of benefits.
Secondly, and still building upon the proposed framework, consumers' interactivity can take distinct roles in the engagement process, such as buying, referring, influencing or giving feedback. These distinct roles may result in successful engagement that firms can transform into an advantage. For example, further understanding of consumer interactivity motivations and the driving of those motivations towards consumer engagement with the focal object (i.e. product/brand) can be explored. Thirdly, research about consumer engagement processes needs to consider the existence of several types of antecedents of a different nature. These antecedents involve consumers, firms and contextual factors that may influence consumer engagement directly. However, some antecedents can also have a moderating effect on other antecedents. Additionally, other factors can ease interactions, such as personal affinity between stakeholders and should be considered. So, further research endeavours should also focus on these latter domains.
Fourthly, the interactive nature of consumer engagement recommends adopting new approaches to future research. We recommend that authors should devote their efforts to studying consumers interacting with several distinct products/services from the same product category simultaneously. Could a firm's engagement efforts directly/indirectly influence the level of consumer engagement with a competing product? Further developments are needed to provide a more accurate vision of the engagement processes in a competitive atomised market.
Fifthly, a possible line of research can be to perform longitudinal studies on consumer engagement. The engagement concept in Marketing is relatively new and academic research on this topic is still in its early stages, meaning that academics lack understanding of how the construct evolves. So, future research could focus on assessing consumer engagement evolution: do consumers react to distinct marketing efforts over time? Do these efforts influence the level of engagement? Does consumer engagement also have a life cycle? Future research could also address the dynamic aspects of consumer engagement that assess the potential impact of initiatives on the consumer lifecycle relationship, how these changes depending on the consumer's life and the degree to which engagement effects decrease over time.
Sixthly, the literature devotes considerable attention to the positive views of brandrelated engagement constructs. However, consumers can also have negative emotional states towards a brand. Future research should address negative engagement not only to have new insights into the unengaged consumer but also to contribute tools and frameworks that managers can use. Seventhly, another fertile area for research could be to explore the role and effectiveness of online brand community engagement platforms that enable consumers to contribute to the focal object (i.e. the product or service). This is a less explored research area and investigation can target either firm-created platforms or existing thirdparty platforms such as YouTube, Instagram or Facebook.
Based on the research developed so far, several research questions are still unanswered, leaving fertile ground for further discussion and investigation (Flavi an et al., 2019(Flavi an et al., , 2020Grewal et al., 2017;Harmeling et al., 2017;Harrigan et al., 2018;Lim and Kumar, 2019;Pansari and Kumar, 2017;Venkatesan, 2017). Based on the aforementioned key areas for future research, this paper suggests a set of research questions for further development: RQ1. How can consumers' interactivity endeavours, such as buying, referring, influencing or giving feedback, interact to produce focal consumer engagement and firm-based benefits?
RQ2. What is the relative importance of consumer-based interactivity efforts in particular consumer engagement contexts?
RQ3. Will consumer and firm-based engagement always develop through the sequential process outlined in the framework or may relevant framework-based concepts have less (or heightened) importance in some contexts?
RQ4. Under what conditions will consumer engagement generate optimal consumer behaviours?
RQ5. How can consumer engagement's contribution to consumer/firm relationship development be expedited?
RQ6. How can consumer engagement be leveraged to increase existing customers' brand loyalty and attract new consumers?
RQ7. What are the benefits of consumer engagement for brand equity and brand attitude and how can these relate to one another?
RQ8. Does the negative impact of consumer engagement vary according to the product category and prior relationship experience that is according to whether consumer and brands/firms' engagement efforts are applied?
RQ9. How does firm-initiated content through firm-created social networks drive usergenerated-content?

Conclusion
This paper reviews a decade of research on consumer engagement, regarding how the literature has evolved and the major research streams on this topic. Our aims are to review the existing literature on consumer engagement, provide an accurate mapping of this research field, propose a framework for consumer engagement and offer a future research agenda. To achieve our aims, we conduct a systematic literature review (Table 2) leading to an integrated overview of prior research findings that can help to summarise and discuss the major research streams in this field (Table 3). This approach means we can answer our research question and provide different perspectives on the conceptualisation of consumer engagement (Table 4). Five research streams are highlighted, with both convergent and divergent viewpoints, exploring the construct through different concepts: (1) Consumer engagement (Brodie et al., , 2013Sashi, 2012).
Based on these, we propose a consumer engagement typology and a conceptual framework (Figure 3) that intends to organise these concepts visually. Further conceptual development and empirical testing are required to assess and validate this proposal, which is highlighted in the emerging future research topics and research agenda.
In the diverse research streams highlighted, several dimensions are suggested depending on the context. Although some research points to a uni-dimensional perspective (Sprott et al., 2009), most studies tend to suggest consumer engagement as a multi-dimensional construct (Baldus et al., 2015;Brodie et al., , 2013Hollebeek et al., 2014;Jaakkola and Alexander, 2014),

Synthesis and research agenda
where cognitive Harmeling et al., 2017;Hollebeek, 2011b;Hollebeek et al., 2014;de Villiers, 2015;Vivek et al., 2012) and behavioural dimensions (Brodie et al., 2013;Groeger et al., 2016;Hollebeek, 2011aHollebeek, , 2011bHollebeek et al., 2014;Jaakkola and Alexander, 2014;Vivek et al., 2012) seem to be two of the most important. This paper also underlines the emerging future research topics and research agenda, discussing the areas that appear to be gaining attention in relevant literature and highlighting key areas for future research. Several research streams are underlined, some of them based on the proposed framework. Here we suggest that researchers should focus their attention on empirical research on the benefits of consumer engagement, on consumer interactivity motivations and the drivers of those motivations towards consumer engagement with the focal object or on the antecedents of consumer engagement that involve consumers, firms and contextual factors that may influence consumer engagement directly, etc. Several research questions are also suggested for further development, with some emerging from our framework. Finally, we outline the main research streams, building on important existing conceptual and empirical work and providing support for the advancement of research on the topic of consumer engagement.