Political social media marketing: a systematic literature review and agenda for future research

We focus on political marketing and conduct a systematic literature review of journal articles exploring political marketing on social media. The systematic literature review delineates the current state of political social media marketing literature. It spans six databases and comprises sixty-six journal articles published between 2011 and 2020. We identify and categorize the variables studied in the literature and develop an integrative framework that links these variables. We describe the research themes that exist in the literature. The review demonstrates that the field is growing. However, the literature is fragmented, along with being predominantly based in the US context. Conceptual and theoretical shortcomings also exist. Moreover, the literature ignores pertinent contemporary topics such as co-creation, influ-encer marketing, and political advertising on social media. Nevertheless, a nascent domain with growing practical significance, political social media marketing provides various exciting avenues for further research, which we outline in this study.


Introduction
Social media have ushered in an era of significant changes in the political, social, and commercial spheres of life.Consequently, social media attract academics from numerous disciplines, which is especially true for politics and marketing.Social media marketing's role in politics will continue to rise [11].The recent US election cycle (2020) saw political candidates and their PACs spend $1.6 billion on digital marketing [47].This figure was $22 million in 2008 when Barack Obama won the presidential election.Allocation of vast resources necessitates a deeper understanding of politics from the marketing perspective [66].Like the disruptive technologies preceding them, such as the printing press, radio, and TV, social media have changed political marketing [27].Unlike prior communication technologies, social media allow voters to interact with political brands and other citizens.This makes them distinct from earlier media and more consequential.
Political marketing is a dynamic and relatively young field [7,81].It is complex because of its diverse origins in marketing and political science.Political marketing is under-researched, fragmented, and in its nascent stages [89].Despite the high scholarly activity in recent years, political marketing still adheres to frameworks borrowed from other disciplines.Moreover, the discipline requires a stronger effort towards theory building [89].Systematic literature reviews are pertinent in this scenario as they synthesize literature, consolidate knowledge, offer a holistic understanding, and steer a discipline towards theory building [36,84].
Like political marketing, political social media marketing (PSMM) has also seen an expansion in literature.Presently, it is the dominant stream of research in political marketing.Perannagari and Chakrabarti's [89] bibliometric analysis of political marketing  shows that 'social media' is the most common keyword besides 'political marketing' itself.Figure 1 demonstrates a significant rise in publication activity over the last five years, including a special issue dedicated to the topic [122].Moreover, PSMM has crossed the threshold of forty articles that are required for a literature review [84].Hitherto, there is no systematic review article that synthesizes the literature in the domain.
It is important that the literature is revisited periodically due to the everchanging landscape of social media marketing and the interdisciplinary nature of PSMM.PSMM has attracted the attention of twenty-one journals in the period between 2016 and 2020 (Table 3).The resulting literature is scattered and requires organization so that it can be observed in its entirety.PSMM can only advance if prior studies are structured and presented logically [64].Review articles provide a structured approach, which is the need of the hour.This will prevent repetition and facilitate the discovery of research gaps.Moreover, PSMM's profound effect on political campaigning and the financial resources dedicated to it mandate a comprehensive understanding to aid practitioners.We aim to contribute to the advancement of the political marketing discipline by providing future PSMM researchers with a holistic view of the domain.
In summary, the review is motivated by various factors.Our review aims to assist future researchers in theory-building.Political marketing is a young discipline and scholars need to engage in theory-building.We aim to bring together and structure the scattered literature in the field of PSMM [89].By doing so, systematic literature reviews can steer a discipline towards theory building [36].Moreover, our review shall provide a starting point for young researchers who are studying the complex topic of political social media marketing that spans multiple academic domains such as marketing, political sciences, and information systems.We believe that social media marketing holds a pivotal place in the political arena.From American Presidents to leaders across South Asia, many world leaders and politicians owe a great deal of their success to effective social media marketing.Therefore, a broader understanding of PSMM is mandatory.Lastly, there is limited guidance to practitioners.Through our integrative framework and Appendix A, we intend to ease practitioners' access to relevant findings.Our review has the following objectives.
1. Describe the current state of research in political social media marketing.2. Explore the themes that dominate the current literature.3. Identify and categorize the variables investigated in the literature and devise an integrative framework.4. Prescribe avenues for further research.
To achieve our first objective, we rely on content analysis to extract and present relevant information such as theories, methodologies, data sources, contexts, and publication activity associated with the selected studies.Thematic analysis is utilized to identify the themes that exist in the literature.The third objective is achieved through a content analysis of the literature, which identifies and categorizes the numerous variables that are explored in the literature.Subsequently, we construct an integrative framework that links these variables [100].Finally, a research agenda that considers the contemporary focus of social media marketing and analyzes the gaps in the literature is proposed.Overall, the review shows that: • Political social media marketing has attracted significant attention since 2016.
The integration of contemporary marketing concepts, although limited, is underway.• From a methodological perspective, there is limited qualitative research that relies on primary data.Conceptual studies are also scant.The number of comparative studies across multiple countries, voter segments, and platforms is limited, which inhibits our understanding of universal findings.More than two-thirds of the studies are based in the US and the UK, which do not represent the political environment of many countries.
• Conceptually, political social media marketing is lacking a unified direction and the literature is fragmented.Overall, this reflects a weak effort towards systematic theory building.Subsequently, less than a third of the studies are underpinned by established theories and few studies rely on theories that originate in marketing and consumer research.The domain is overtly focused on campaigns and elections, which shows that it is yet to shun the marketing mix paradigm.• Distinct themes are beginning to emerge in the literature.These themes, to a certain degree, align with the recent research in social media marketing.However, several timely and relevant topics such as influencer marketing, customer engagement, value creation, and co-creation remain unchartered in the political context, which provides a great opportunity for future researchers.
Our review identified sixty-six relevant articles and spanned five social media platforms.It contributes to the field of political social media marketing: 1. To the best of our knowledge, the review is the first systematic literature review of the domain.2. It dissects the literature from various angles to provide a comprehensive and indepth analysis of the current state of research in the field.3. It offers a description of the various themes that exist in the literature.4. It provides a framework that integrates the frequently studied variables in PSMM. 5.It identifies academic gaps in the literature and offers an agenda for future research.
The article is organized as follows.First, we describe our conceptual boundaries: social media marketing, political marketing, and political social media marketing.Second, we explain the systematic review process, which is built upon the best practices that are highlighted in the literature (e.g., [97,112].The subsequent section comprises the findings of our first research objective.It includes chronological, contextual, theoretical, and methodological analyses of the literature.This is followed by the findings of the thematic analysis, our second research objective.The subsequent section addresses the third research objective.It identifies and categorizes the variables that are explored in the literature and devises a framework that integrates these categories.The following section comprises the proposed research agenda.The penultimate section highlights the study's theoretical contributions.Limitations are highlighted in the conclusion.

Social media marketing
Social media hold a central position in modern marketing [5,36].Social media are 'a group of internet-based applications that build on the ideological and technical foundations of Web 2.0, and that allow the creation and exchange of user-generated content' [59], p. 61).In our study, we focus on five popular social media platforms: YouTube, Facebook, Instagram, Twitter, and Snapchat.However, the latter does not feature in our review despite its utility as a political marketing tool [21].Social media marketing research covers many topics such as advertising, branding, eWOM, user-generated content, relationship marketing, and customer behavior [5].It is because social media provide diverse benefits to marketers.A significant body of literature establishes the positive effects of social media marketing on desirable business outcomes (e.g., [35,37,30].
Social media marketing is widely used by businesses and non-profit organizations, including public bodies [60].Although it is used to promote products and services, the underlying goal of social media marketing is to cultivate long-term relationships with the consumers and relevant stakeholders [103].Social media facilitate consumer-to-consumer interactions in addition to brand-to-consumer interactions, allowing brands to benefit from e-WOM, which is considered more credible than brand-generated content.In the era of social media, commercial dynamics have undergone substantial changes.Social media have shifted the power to the consumers but have also provided marketers with new opportunities in the form of microtargeting of consumers and listening abilities through data mining.Particularly in the post-covid world, social media has gained a stronger foothold in the life of the consumers [73].

Political marketing
Political marketing is an old tradition.'Codifying political marketing could take the discussion back to Aristotle's writings on Politics and Rhetoric' [26], p. 152).The modern discipline has origins in Kotler and Levy's [63] expansionary view of marketing.Political marketing is 'a set of activities, processes, or institutions used by political organizations, candidates and individuals to create, communicate, deliver, and exchange promises of value with voter consumers, political party stakeholders, and society at large' [48], p. 244).
In simpler terms, political marketing is the use of marketing concepts and techniques to achieve political goals.As a fundamental part of modern politics, political marketing is practiced by politicians, political parties, and political movements.These entities engage in various marketing activities such as market research, segmentation and targeting, branding, internal marketing, marketing communications, advertising, and relational marketing [65].Additionally, political brands need to choose between different political marketing approaches, which are selling, transactional, relational, and experiential.
Notably, political marketing has permeated the popular culture, becoming the topic of several movies and TV shows.Similarly, political marketing has appreciably grown as an academic discipline [89].Therefore, it is important that marketing and political science scholars appreciate the role of political marketing in contemporary politics and its impact on democracy.A sub-discipline that draws on politics and marketing, political marketing needs to continually evolve to reflect changes in its parent disciplines [48].Researchers note that this is not the case, and a second wave of research is needed since political marketing does not reflect the dynamism and evolution of marketing [81,80,45].Political marketing literature is nascent and fragmented [89].Finding common grounds between two distinct academic traditions is not an easy task [65,81].The diversity of political marketing requires a systematic approach to future research to prevent duplication of research, build theory, assist young researchers, and identify research gaps [84].

Political social media marketing
Social media hold a prominent place in politics [22,57].Political social media marketing, or political marketing on social media, refers to the use of social media to create, communicate, and deliver value for stakeholders [114].The academic domain gained traction following Barack Obama's online presidential campaign (2008) [79,76].Along similar lines, Donald J. Trump's presidential campaign had a significant impact on the research in the field, as demonstrated by our systematic literature review.However, it was Howard Dean who was the first to effectively integrate social media during his 2004 nomination bid [65].
Unlike prior media, social media provide political brands with an unfiltered and direct communication channel.They allow political brands to keep the voters updated, interact with their followers and general public, promote their brand via social media ads, mobilize voters, solicit campaign donations, and engage voters through relevant content.Moreover, social media provide political brands with an effective mechanism to co-brand and co-create.Donald Trump's #MAGA challenge and Bernie Sanders' #MyBernieStory are prime examples of such co-creation and co-branding.Importantly, social media marketing allows politicians to reduce their psychological distances with the voters [116].Literature shows that political social media marketing has an impact on voters' attitudes and behaviors [19,49].
Social media are interactive and voters are influential on social media.Consequently, social media require a different approach to political marketing than the one utilized on traditional media.Specifically, political marketing on social media demands an interactive and relational approach [27,44,65].Recent research shows that social media require politicians to adopt a personal and social approach, which is built upon the politician's and constituent's shared values [1].However, a considerable body of literature shows that politicians and political parties, in general, have been unable to embrace the orientation that social media require or are reluctant to do so [83].

Methodology
Systematic literature reviews (SLRs) are appropriate to synthesize or provide an overview of an academic domain, develop themes, create conceptual models or integrative frameworks, and propose a research agenda for the future [84,108,100].The methodology is frequently utilized in marketing and management research [64,94,112].SLRs are especially valuable to social media marketing.The interdisciplinary perspectives, wide-ranging research questions, variety of theories, diversity of research methods, and the rapidly changing social media landscape mandate regular SLRs in social media marketing [5,36].Similar concerns hold for political marketing.However, SLRs are infrequent in political marketing (e.g., [89,123].Specifically, our SLR is a domain-based review that can be further classified as a structured theme-based review [84].These reviews document the various theories, constructs, methods, contexts, and research themes that exist in the literature, along with offering conceptual frameworks and future research agendas (e.g., [54,96].
Our review process follows the best practices that are highlighted in the literature (e.g., [95,84,97,112].In the planning stage, the need for such a review was ascertained.Paul et al. [85] recommend that a research domain warrants a systematic literature review if forty or more articles are dedicated to that domain.PSMM, therefore, meets this condition as there are more than sixty articles that fall in the domain of PSMM.Furthermore, the review is valuable since political social media marketing will continue to increase in significance and consequence [11].Following this, the research aim, criteria, and guidelines were established.These included decisions concerning the timeframe, search terms, databases, inclusion and exclusion criteria, and social media platforms [84,97].We limited the review to articles from academic journals that have an impact factor, choosing to exclude conference papers and book chapters (e.g., [62,100].Table 1 presents the inclusion and exclusion criteria. The selected period (2011-2020) reflects the evolution of social media in politics.The time frame is appropriate since social media were not a significant medium fifteen years ago.Facebook was not open to the general public and Twitter and Instagram did not exist.Ten years is an acceptable time frame for SLRs [84].One generic and five platform-specific search terms were included to extract relevant articles.We used the following Boolean search term: "political marketing" AND ("social media" OR Facebook OR Twitter OR YouTube OR Instagram OR Snapchat).Using predefined keywords to extract literature is a common practice in SLRs [97].The selected platforms are widely used and are relevant to politics.The selection of databases was driven by prior literature in the field of marketing and management (e.g., [64,108].The six databases (ProQuest, EBSCO, Web of Science, Emerald, Science Direct, and Scopus) represent a significant fraction of the marketing literature.
In the second stage, the search was conducted [85,62]).The search results are presented in Table 2.In the initial phase, the titles, abstracts, and keywords were read to ascertain the study's relevance.Full-text screening was undertaken when needed.Since our aim was to identify political marketing articles, we ensured that the articles from non-marketing journals adopted a marketing perspective.Subsequently, studies from journals having an impact factor were isolated.These articles were read and articles from academic domains of political campaigning and political communication were excluded (e.g., [13].The final count after applying our inclusion and exclusion criteria and removing duplicates was sixty-six.These studies were read and coded for the country, theory (or literature review), context, publication information, focus, methodology, and data among other categories.Additionally, a thematic analysis was undertaken.Notably, we excluded a guest editorial (e.g., [122].Table 3 lists the sources of the selected articles.

Field of research
The research is embedded in the political context A complete or partial focus on social media (including comparative studies) The study utilizes a marketing perspective (such as, appearing in a marketing journal or the use of marketing or political marketing in the keywords or abstract) All other studies

Journal quality
The journal has an impact factor

All other journals
Other Full text available  Thematic analysis reveals the patterns or themes that are present in the literature [23].An inductive approach was utilized to identify broad themes.An inductive approach reflects the fundamental concepts and topics that are present in the domain [56].Since the present study is the first review on the topic, it is important that it highlights the actual state of the literature.Two authors were involved in the process.Articles were read and coded independently before checking for consistency.The inter-coder agreement was high.The few differences in coding judgments were resolved through discussion.Further discussion was undertaken to finalize the number and scope of themes [93,112].For example, certain themes (social media's role in protests) were merged into others (social media's effect on voters' behavior) due to the limited number of articles investigating the former theme.
We did not use a pre-determined conceptual framework to identify themes [97], however, a marketing dominant approach meant that some of our themes reflect broad marketing concepts such as branding, marketer-generated content, voter behavior, user-generated content, and relationship marketing.Similar themes are reported in prior systematic reviews of social media marketing [5].Several articles researched multiple topics and thus featured more than one theme.In such cases, the article was categorized into more than one theme.For example, Buccoliero et al. [25] touched upon three themes since it explores the differences in the social media orientation of politicians during the US presidential race in 2016, along with a discussion on political marketer-generated content and followers' responses.Similarly, Abid et al. [2] was included in both political marketer-generated content (MGC) and political relationship marketing since the article examines the impact of political marketer-generated content on online relationship quality.
In the third stage, we present the findings from the review.The literature is dissected from various angles, a standard practice in systematic reviews.Tables and figures are included to assist in the presentation of results [85].In line with prior literature, we propose an integrative framework and a research agenda [97,100].Appendix A provides a brief overview of the studies included in the review.

Objective 1: the current state of research in political social media marketing
To achieve our first objective, extensive codification of the studies was undertaken.We review the chronological evolution of the domain.Subsequently, we discuss the geographic and political settings, social media platforms, voter segments, theories, sources of data, and research methods associated with our selected corpus.

A chronological review
Figure 1 highlights the number of articles published yearly.A special issue on the topic explains the high article frequency in 2017 [122].2011-2015 (13 studies): PSMM gained prominence following Barack Obama's presidential bid in 2008.Consequently, Obama's campaign featured in several studies during this time (e.g., [31,76].Studies primarily relied on the case-study methodology to explore the role of social media in elections or campaigns (e.g., [44,43,107].Few studies tested relationships between variables during this time [40,75,109].Trust and political participation interested scholars during these years (e.g., [40,105,112]. 2016-2020 (53 studies): Sophisticated methodologies, reliance on big data (e.g., [17], dependence on theory (e.g., [33], and integration of marketing concepts (e.g., [91] has increased over the last five years.Publication outlets have increased from two journals in 2015 to twenty-two journals in 2020 (Table 3).Donald Trump and the US presidential election (2016) have revived the interest in the domain.

Geographic distribution
Figure 2 exhibits the geographic settings of the studies.The focal point of the literature is the US, which features in thirty-six studies.This is followed by studies based in the EU (9), UK (7), Asia (4), Australia/New Zealand (3), and Africa (2).Two studies explore the US in relation to the EU and Egypt, whereas one study compares the behavior of the UK and US voters.The distribution shows that there is room to explore PSMM across diverse geographic contexts.The political environment varies across countries.Consequently, the concentration of research in the US limits the applicability of PSMM.For instance, the personalization of politics associated with the US and the two-party presidential system might not be relevant elsewhere.

Social media platforms
Table 4 presents the frequency with which various platforms appeared in the literature.Most studies discussed social media in general.Specifically, Twitter is the preferred platform among PSMM researchers, particularly since 2016.Although this might signify Twitter's rising influence in the political arena and the US, the ease of access to data is another possible explanation for Twitter's preference among researchers.Social media platforms that are comparatively new are yet to be explored.Surprisingly, research exploring YouTube and Instagram, popular mediums for political marketing, is also limited.Although a few studies explore multiple platforms, these are not comparative in nature (Table 5).

Comparative perspective
Several studies in our selection adopt a comparative perspective.These studies focus on various comparisons such as that between different media [8,102], young and old voters [50], American and British voters, [71], political and commercial brands [20], political candidates and campaigns (e.g., [25,90,117], brand communities of candidates [69], social media followers of politicians and regular citizens [19], characteristics of user-generated content during and after political events [17], and the US and EU laws governing political marketing on social media [16].

Political and electoral contexts
Roughly three-fourths of the articles are embedded in campaigns, lead-up to elections, and elections.Particularly, the US presidential elections of 2016 (14), 2008 (7), 2012 (5), and the UK general election of 2010 (3) are researched frequently.Most studies focus on the national or presidential level and an understanding of how parliamentary, state, and local candidates utilize social media marketing is limited (see Table 6).Similarly, Table 5 shows that candidates are explored more frequently, with the US research almost exclusively focusing on candidates barring odd exceptions (e.g., [2].

Voter segments studied in the literature
Young voters are the most studied subjects.As per the literature, they have low trust in government and politicians [102,109] but hold favorable perceptions of minor political parties' and local politicians' use of social media marketing [3].Political marketing on social media leads to stronger relationships with young voters and increases their political efficacy [7,49].They prefer brief political eWOM [50] and desire personal and social content from politicians [1].They are generally less engaged in traditional elections [91].Other segments investigated in the literature include minority voters [38,40], aged voters [50,109], female voters [43], bloggers [87], and followers of politicians [19].

Research methods and data collection
Quantitative studies dominate the literature (number of quantitative studies = 44; Appendix A lists the quantitative studies).However, only twenty-nine studies test relationships between variables (see Appendix A).Content analysis of social media pages is the most frequently utilized research method in PSMM (see Table 7).The use of case-study methodology, which dominated literature till 2015 (e.g., [44], is declining, whereas surveys, social network analysis, online content analysis, and experiments are being used more frequently (e.g., [15,20,49,69].Remarkably, the number of studies relying on interviews and focus groups comprise a small fraction of the literature.Longitudinal research is also absent in the literature.Similarly, conceptual papers are scarce in the literature.Expectedly, social media pages are the favored source of data in the literature (see Table 8).Less than a third of the literature relies on primary data that was collected from the voters.Similarly, data from party officials and political candidates constitute a small fraction of the literature.Few studies integrate offline sources of data.These are Berman et al. [17] and Peres et al. [90], which integrate data from presidential debates and press articles in their studies of usergenerated content and world leaders' use of social media respectively.Various forms of secondary data are utilized in the literature.These include the use of publicly available information such as news, electoral results, existing survey data, and author's personal data (political consultant), among others (e.g., [31,40,76].

Secondary data 12
User tweets/comments/networks/groups 10 Politicians and party officials 7 Others (press articles/debates) 2 Published journal articles 1

Theoretical foundations of PSMM
Less than a third of the studies are underpinned by established theories and models.
Theories from the field of psychology, particularly social psychology, are used most frequently.Among these, social identity theory is the most utilized theory, featuring in three studies.Other theories originating in psychology like theories of Self-Concept, Planned Behavior, Self-Presentation, and the Elaboration-Likelihood Model are also engaged in the literature, with each appearing in two studies.Additionally, theories from the fields of communication and mass communication form the basis of several studies.These include the Two-Step Flow of Information Theory (n = 2), Symbolic Convergence Theory, Framing Theory, and Transmission Model of Communication.Besides these two disciplinary underpinnings, few studies rely on theories from the fields of media (e.g., Uses and Gratification Theory and Connective Action (n = 2)), information systems (e.g., Technology Acceptance Model), consumer research (e.g., Reference-group influence and Persuasion Knowledge Model), and behavioral economics (e.g., Reference Dependence Theory).The theoretical underpinnings of the studies are highlighted in Appendix A.

Summary of findings: objective 1
The key findings from the preceding analysis are stated below.
• The research on PSMM is on the rise.• The US and the UK dominate PSMM research.
• There are limited studies that offer a cross-country comparison between voters.
• Candidates are studied more frequently than political parties.
• Presidential and national politics dominate the literature.
• Young voters are the most frequently explored segment.
• Sophisticated research methods such as experimentation, big data analysis, and social network analysis are replacing the case-study methodology.• Social media pages are the preferred source of data.• PSMM's theoretical foundations need to be strengthened.

Objective 2: What are the themes that exist in the PSMM literature?
To identify the themes that exist in the literature, we conducted a thematic analysis.Our analysis revealed ten themes.The themes are presented in Table 9.The themes vary in their presence over the last decade.For instance, the last five years have seen diminished interest in the adoption of social media for political activity and an increased focus on branding, relationship marketing, and social media content.Similarly, certain themes like 'political marketer-generated content', 'political brands' approach towards PSMM', and 'adoption of social media for political activities' are more popular among researchers.This is also evidenced in Tables 10 and 11, which reveal the high number of variables associated with these themes.Consequently, a few themes can also be traced in the integrative framework, which is derived from the variables that are studied in the literature.Besides identifying themes, we include a brief synopsis of these themes.

Political social media marketing during elections/campaigns
Studies in this theme dissect the social media strategies of political parties and candidates or the extent and manner of their social media use during elections.Several articles explore Barack Obama's groundbreaking use of new media in 2008 [31,79,110] and Donald Trump's Twitter-savvy strategy [25,33,53,101].While Obama executed a grassroots movement that connected like-minded voters via social media and his website, Trump was able to use social media to bypass party elites and cultivate an authentic brand.Studies outside the US explore social media marketing during the UK [44,43,107] and Indian elections [6,52], with social media playing a consequential role in the Indian context where Prime Minister Modi and his Bhartiya Janta Party were able to use social media effectively.The UK general election (2010), however, showed that Obama's campaign did not trigger an instant adoption of PSMM in the UK.

Approach towards political social media marketing
Our review shows that politicians do not adopt an interactive or relational approach to social media and utilize them in a traditional manner, i.e., for political broadcasting, self-promotion, and self-advocacy [41,44,61,98].Politicians tend to inform and mobilize through political marketing [117].Their posting frequencies increase or decrease based on campaign needs [98].Moreover, the distinctions between politicians are diminishing as social media mature, with a vast majority of politicians adopting similar communication styles, self-presentation strategies, visual framing, and emotional appeals [41,78,86,90].2. Political brands' approach towards PSMM [78] 3. Adoption of social media for political activities [75] 4. Social media's effect on voters' behaviors [34] 5. Social media's ability to predict elections [29] 6. Political marketer-generated content [120] 7. Social media and political relationship marketing [44] 8. Social media and political branding [18] 9. User-generated content [17] 10.Digitalization and professionalization [42]

Adoption of social media for political activities
Candidates: Factors such as the candidate's age, the target market's age, and the type of election dictate the adoption and usefulness of social media marketing [76].Further, a candidate's adaptation, relationship-building, leadership, and innovation capabilities influence their adoption of social media [9,10].Other factors that are pertinent to the adoption of social media include the digitalization of national politics and the candidate's education and understanding of social media [42,99].Besides politicians' general adoption of social media, their adoption of social media for eliciting inbound communications is also investigated in the literature [75].
Voters: Political interest, gender, race, and party identification determine the general political use of social media [40,111].Specific social media activities, however, vary in the factors driving them.Following political brands, for instance, is driven by gender (male), higher income, race, and college education, whereas political tweeting is driven by low education, disagreements, political ideology, and political motivations [19].Sharing political videos is primarily driven by personal motivations, political motivations, and political ideology [87,121].Finally, the research shows that some factors (e.g., visibility of likes) inhibit engagement with political content [71].

Mediators
Emotional reaction to message [72] Candidate image [49] Message involvement [50] Persuasion knowledge [20] Political efficacy [7] Liking the communication [32] * Variables from qualitative studies are specified.The remaining variables are extracted from quantitative studies

Social media's effect on voters' behavior
Research demonstrates that PSMM has an impact on online and offline political participation (e.g., [7,34,40,111].Although, the effect is more pronounced when voters are active followers (liking and sharing) rather than passive followers [34].Social media also facilitate grassroots activism and political protests, which was witnessed during the Arab Spring [58].

Social media's ability to predict elections
The predictive capabilities of social media have also received attention [104,29,68,99].These studies show that various social media-based indicators such as the number of Facebook friends, pre-election changes in the numbers of Facebook friends, retweets, account type, and verification badge can be used to predict electoral results [104,29,68].

Political marketer-generated content
The content posted by politicians and political parties is the subject of several studies in our review.Whereas some studies attempt to understand the characteristics of the content posted by politicians or political parties, such as the level of personalization, production techniques, emotional appeals, themes, word count, communication styles, and credibility cues used in the content [90,41,86,87,82,98,78], other studies explore the effects of various content characteristics on content virality and voter outcomes.The virality of tweets, for instance, is dependent upon various structural elements of the tweet, source characteristics, sentiment of the tweet, and its content [120].Emotional content drives favorable attitudes and behavioral intentions [72].Lastly, certain factors affect the composition of political marketergenerated content.Gender [86], country or culture [90], and party or individual characteristics are a few variables that are discussed in the literature [2,25].

Social media and political relationship marketing
Relationship marketing is the only feasible orientation towards political marketing on social media [44].It increases political participation among citizens [7].Qualitative studies in the domain explore the extent to which politicians practice relationship marketing on social media [3,44], the factors hampering the application of relationship marketing [83], and the nature of social media-enabled voter relationships [1].The latter study finds that young voters desire a personal and social relationship with political brands, whereas the former studies conclude that politicians and political parties do not adopt a relational orientation to social media.However, local politicians and minor parties enjoy a positive perception [3].Quantitative studies confirm that certain social media marketing activities and content cues influence relationship equity and relationship quality respectively [2,49].

Social media and political branding
Donald Trump's branding strategy has received substantial attention [18,79,101].Studies highlight that social media are transforming political branding, which is becoming a co-created, technology-driven phenomenon as demonstrated by the rise of Donald Trump and 'cyber political brands' in the EU [18,70].The research on political brand communities shows that communities devoted to lower-tier candidates are dense and exhibit greater reciprocity [70].Further, brand communities play an important role in political co-branding [18].Other studies demonstrate that the behavior towards political brands on social media is different from that towards commercial brands [20,71].

User-generated content
Like marketer-generated content, user-generated content (UGC) is also examined in the literature.The virality of user-generated content is determined by certain content characteristics such as the tweet's surface features, linguistic style, emotion, and topic [17].Studies show that young voters prefer brief user-generated content on social networking sites [50].
The nature of user-generated content is affected by a politician's gender, party, and the nature of marketer-generated content itself [53,84].For instance, Congresswomen receive more comments with joy, whereas Republican candidates prompt more disgust and anger.Emotional MGC leads to emotional UGC [87].Studies examining political brand's responses to user-generated content recommend proactive behavior [15,27].Besides user-generated content, the user's roles as citizen marketer [87] and gatekeeper of citizen-led Facebook newsgroups [6] are also discussed in the literature.

Digitalization and professionalization
Lastly, a few studies highlight that modern politics is becoming professionalized and digitalized.This is the case in the US [27] and Europe [42].Political consultancy is becoming tech-driven, creating an immense need for specialist consultants.In India, political parties are creating IT cells and hiring advertising agencies to train and assist politicians or post content on their behalf [99].

Objective 3: Variables investigated in the literature and framework development
To achieve our third objective, we identified and categorized the variables that are studied in the literature (see Table 10 and Table 11).The first author read the articles and listed the variables studied in these articles.The categorization of variables was undertaken by the first author.The variables are categorized as per the antecedents-consequences model.Mediators and moderators are also incorporated [119].Tables 10 and 11 show that official channels of political brands interest scholars.Scholars are particularly interested in voter, profile, content, and situational variables that amplify the effect of political MGC that is generated by the official channels of political brands.Similarly, the drivers of PSMM adoption and variations in politicians' use of social media are topics of significant interest.Beyond this, the literature is lacking.For instance, there is limited literature devoted to brand communities, political advertisements, and user-generated content (e.g., [17,20,88], the other mechanisms through which voters are influenced besides official channels.Considering this skew in the literature, we devise an integrative framework that focuses on the topics of scholarly interest and not the entire literature.Integrative frameworks are valued by practitioners and academics alike [119,74].The framework elaborates on the manner in which official social media channels of political brands influence voters.It allows readers to understand the linkages between different categories of variables that prevail in the literature.To ensure objectivity in our framework, we only include variables that are studied quantitatively.The absence of an integrative, discipline-specific framework is a shortcoming of the current research that our framework rectifies.Figure 3 illustrates this framework, which adheres to the antecedents and consequences model, with the inclusion of moderators, mediators, and contextual factors, as done by Vrontis et al. [119].The framework is derived from Tables 10 and 11.
Our framework highlights two types of antecedents.These are factors that drive political brands and voters to adopt and consume PSMM respectively.The former is motivated by macro trends like digitalization and personalization, along with politicians' individual characteristics (e.g., [42,9].Voters, on the other hand, are driven

Drivers of PSMM (Candidate):
Digitalization, personalization, ability to lead, innovate, adapt, and build relationships, satisfaction with social media, ease of use, subjective norms, etc.

Drivers of PSMM (Voters):
Perceived social influence, political motivation and disagreements, age, gender, race, party identification, trust in media, education, political interest, knowledge, etc.

Moderators (how politicians use social media):
Ethnicity, party, gender, incumbency, competitiveness of race, country, culture, candidate tier, campaign v/s non campaign periods, etc.

Mediators:
Emotional reaction to message, candidate image, message involvement, persuasion knowledge, political efficacy, liking the communication

Offline outcomes:
Political attitudes, source attitudes, political participation, source trustworthiness, interest in party, candidate favourability, relationship equity, intention to vote, etc.

Online outcomes:
Online behavioural intentions, UGC creation, online political participation, online relationship quality, content virality, intention to like, likes, etc.

Consequences
Fig. 3 Integrative framework by demographic factors and personal motivations and knowledge (e.g., [40,111].The framework highlights two consequences of PSMM, which are categorized as online and offline outcomes.The factors that are categorized as antecedents and consequences are presented in Table 10.
Moreover, we identify two sets of moderators.The first set comprises variables that influence the way politicians use PSMM and explain the variations between different politicians that occur due to their strategy, gender, tier, or country [53,69,87,98].The second set of moderators constitutes variables that influence PSMM's impact on online and offline voter outcomes.These include political MGC cues, profile cues, and voter characteristics [72,71,90].Besides identifying the moderators, we identify the variables that mediate the relationship between PSMM and voter outcomes.These mediators include factors such as candidate image, emotional reactions, and persuasion knowledge [20,49,72].The moderators and mediators are presented in Table 11.Although the framework is not comprehensive, since it ignores certain topics that have yet to attract substantial academic interest, political marketers can benefit from it since it highlights the factors that magnify the effect of PSMM.

Objective 4: future research agenda
Through the first three objectives, we offer researchers a structured and comprehensive view of the domain and the extent of current knowledge, which can help avoid duplication of research and facilitate the discovery of research gaps.The review identified various academic gaps in the literature, along with conceptual and methodological shortcomings.In light of these inadequacies and contemporary social media marketing literature, we propose an agenda for further research in the field.

Methodological directions
From a methodological perspective, PSMM requires exploration across diverse geographic contexts, with an emphasis on how differences in the political climate, democratic forms, and voter participation levels influence PSMM.For instance, is political content with negative valence less effective in countries where polarization levels are low?Similarly, comparative studies like Marder et al. [72], which found variations in the social media responses of American and British voters, are rare.These studies shed light on how political and cultural factors influence voters' engagement with political brands on social media.
Another methodological issue is the limited voter perspective.Reliance on primary data that is collected from the voters is limited.Moreover, barring young voters, few studies explore PSMM in relation to a specific segment (e.g., [43,111].Therefore, future researchers are advised to investigate how different electoral segments (undecided voters, women, and minority voters) engage with political brands on social media.Researchers can also explore how the level of politics (local, state, national) influences PSMM and voters' behavior.Furthermore, the literature is predominantly focused on Twitter and Facebook.Future researchers should study other platforms such as Instagram, YouTube, and Snapchat [21].With Gen Z becoming a sizeable voter segment, TikTok merits investigation since its political use is on the rise [39].Particularly, the differences in PSMM practices and voter behavior across social media platforms merit investigation.For example, the visibility of a user's 'likes' has a negative impact on Facebook [71], but does this hold for Twitter?Is one platform better than others for achieving specific goals such as building voter relationships?The effect of platform characteristics on users is a relevant and timely topic in social media marketing [36].
Finally, studies relying on interviews and focus groups are scant, which means that PSMM is missing a rich, in-depth understanding.Qualitative studies are advocated in social media marketing literature also, which primarily relies on a quantitative approach [5,36].Conceptual papers have a limited presence in the literature.Conceptual articles are valuable as they propose new and relevant constructs and relationships [51].Therefore, future researchers are encouraged to devise rigorous and novel conceptual studies.Additionally, there is a need to use longitudinal analysis to understand PSMM's effect on voters.

Conceptual directions
The review highlighted a domain that is lacking a concerted approach to systematic theory development.First, the number of studies embedded in sound theories comprises less than a third of the literature.Second, the majority of the research is geared towards problem-solving rather than theory-building and can be classified as practice-oriented [63].Third, most of the theories utilized in the literature do not originate from consumer research or marketing literature.Therefore, we encourage future researchers to focus on theory-building, which is essential for the discipline's academic growth and standing.Specifically, concepts and theories originating in marketing should be preferred.Importantly, PSMM researchers are encouraged to utilize concepts and paradigms that reflect the contemporary focus of marketing and social media marketing.Finally, the literature is almost explicitly focused on the short-term (elections and campaigns), which is contrary to contemporary marketing thought.Thus, researchers should focus on embedding their studies outside the contexts of campaigns and elections.

Thematic directions
The impact of PSMM on voters and voting behavior: The review shows that scholars want to know how PSMM is used by practitioners (political advisors and politicians), with an emphasis on its impact in the short-term, i.e., elections.Therefore, future research could further investigate PSMM in relation to voter behavior.Various outcomes have been explored (e.g., [7,34,49], but pertinent outcomes such as voting, volunteering, and financial contributions remain under-researched.The latter is important given seventy-five percent of Facebook ad spending in the US election cycle of 2020 aimed to raise funds, solicit contributions, or sell merchandise [47].Future researchers could further refine our understanding of social media's predictive capabilities (e.g., [30,68] and identify indicators of an electoral win across different platforms.Identifying such metrics or analytics will help political marketing managers understand the ROI of PSMM and that of each platform.A comparison of the predictive capabilities of various platforms also merits further investigation (e.g., [29]. General approach to PSMM: Candidates are becoming homogenous in their PSMM, which indicates an ideal approach (e.g., [90].However, Donald Trump's unique approach and his success negate this view.This presents an interesting dilemma for future researchers to resolve.Politicians and political parties offer limited interactivity and engagement opportunities on social media (e.g., [1], but there is inadequate guidance as to what these interactive and engagement opportunities entail, and whether they have a positive impact on desired outcomes.For instance, is it feasible for a politician to engage in a dialogue with voters on social media?Personalization is a growing trend in politics and few studies explore its' effectiveness (e.g., [33].Is it more effective than an issue-dominant approach?Should candidates post personal content frequently?Do politicians who manage their social media themselves (Donald Trump) fare better?Such research has practical significance.Barack Obama showed that PSMM is important beyond elections and campaigns, i.e., once politicians have been elected and are in government [31].However, researchers have ignored this aspect of PSMM.Are distinct strategies, content, and orientations required when in government and opposition?Such questions demand attention from scholars of political marketing.
PSMM and branding: PSMM's effects on brand loyalty, brand personality, brand awareness, brand knowledge, and brand image remain unexplored and present viable directions for future research.Since behavior towards political brands differs from commercial brands [20,71], research is needed to understand when, how, and why these deviations occur.Current research on brand communities, both official and unofficial, offers limited insights into how these communities operate.
Social media and political relationship marketing: Relationship marketing is the advocated approach to political marketing [81] and PSMM [44].There remains a need to understand what a relationship marketing approach towards PSMM entails.Future studies should identify effective examples of political brands that have used social media for relationship marketing.Additionally, quantitative research is yet to establish if a relationship orientation is more effective than a traditional approach to political marketing.Further, how PSMM can facilitate inter-voter relationships is an important question [105].Finally, political marketing literature highlights a relational approach towards society and various stakeholders [48].Future researchers can add value by exploring PSMM beyond voters.For example, Donald Trump regularly communicated with stakeholders like Fox News and National Rifle Association via social media.
Political user-generated content and eWOM: Berman et al.'s [17] study on Twitter is the only direct attempt to understand the effect of UGC cues on content virality.Future research can attempt to understand the virality of UGC using different cues .Further research could understand how and why voters create political content (e.g., [87].Limited research explores the effect of UGC/eWOM on voters' attitudes and behaviors (e.g., [50], which is worthy of examination since the effect of UGC is different from MGC [77].Future researchers can also explore the effects of different types of UGC (e.g., influencer-generated, celebrity-generated, and citizen-generated).
PSMM and political MGC: Politicians need to provide content that is relevant, valuable, and enriching to the voter experience [105].Future researchers can use various content classifications and characterizations highlighted in marketing literature to understand the effect of various MGC cues and characteristics (e.g., [14,106].Importantly, the effect of political MGC has mostly been studied via content analysis, which does not allow for an understanding of the interplay between political MGC and source, situational, or user characteristics.How source characteristics, situational variables, and voters' personality traits impact political MGC's reception are topics that demand attention.Experimental studies can add value here (e.g., [33,20].
PSMM and political advertising: Only two studies in our selection investigate political social media advertising [20,117].Evidence suggests that promoted tweets have a counteractive effect [20], which is surprising since political advertising on social media, particularly Facebook, constitutes the largest portion of most campaigns' digital marketing budget.Therefore, the effectiveness of political ads merits further investigation.
PSMM and value creation: Value creation and co-creation, prominent themes in marketing and social media marketing [12,60], are rarely explored in PSMM.Value creation is a critical element of political marketing [48].How can social media facilitate the co-creation of value between political brands and voters?What are the antecedents and consequences of this co-creation?Such questions remain unanswered.
PSMM and voter engagement: Engagement, an important concept in contemporary marketing [24,46], is pertinent to social media [4,12].However, Pich et al. [91] is the only study that engages this paradigm and studies voter engagement using the customer engagement framework.
PSMM and influencer marketing: Another important topic in social media marketing, influencer marketing [11,119], has received negligible attention in PSMM (e.g., [104].The motivations driving political influencers, characteristics and types of influencers, and their impact on voters are valid areas of research.
PSMM and overall media mix: There is a need to understand PSMM's role within the overall media mix and digital marketing strategy.Social media do not function in isolation and are impacted by or impact other media [36].Therefore, an understanding of PSMM as a component of a holistic political marketing strategy is beneficial.This perspective is highlighted in the recent social media marketing literature [11,36,118].
PSMM and ethics: Despite ethical concerns around PSMM [11], we have a limited understanding of ethical issues related to PSMM.The topic remains unexplored, which provides researchers with a meaningful avenue for further research.Table 12 provides a summary of the directions of research that may be undertaken in the future.

Theoretical contribution
Review articles advance theory in several ways [92].We contribute to the theory by describing the ten themes that exist in the relevant literature.It is important for researchers to understand the dominant streams of research "for developing and strengthening the theoretical positioning of research" [67], p. 1148).It helps future researchers position their research in relation to existing literature, as well as uncover gaps within these streams of research.Rather than summarizing the literature, we adopt a critical perspective and dissect the literature from various angles to identify the shortcomings in the literature.For instance, from a contextual perspective, most of the studies are embedded in the American context and focus on presidential politics, which limits the generalizability of the current literature.Similarly, methodological limitations are also highlighted in our analysis such as the infrequent utilization of qualitative techniques.The theoretical limitations of the relevant literature are also revealed in the study.For instance, the domain is primarily built upon theories that originate from the fields of psychology, communication, information systems, and media.The review demonstrates that there is a lack of integration of contemporary marketing concepts in the current literature.Political social media marketing as a domain has traditionally been unreceptive to modern marketing paradigms.For instance, engagement, co-creation, and servicedominant logic are yet to gain prominence in the relevant literature.Relationship marketing, however, is an emerging perspective within the PSMM literature and presents an opportunity to unify and update the theoretical foundations of the PSMM literature.
The review makes another contribution to the theory.This is in the form of our integrative framework and identification of variables that have been studied in the literature.The integrative framework identifies the antecedents and consequences of political social media marketing, along with the various factors that mediate and moderate the effects of this marketing.By identifying the constructs studied in the literature and the contexts in which they were studied, we help future researchers in theory building as they can choose between studying new constructs, investigating constructs in a different context, or jointly examining multiple constructs from the same category of variables, given most of these constructs were studied independently.

Conclusion
Our systematic review synthesizes and presents an overview of the literature in the field of PSMM.Like similar domains, such as political campaigning and political participation [22,57], PSMM can also benefit from a systematic arrangement of the literature.The review illustrates that PSMM is gaining traction globally, particularly among marketing scholars.Research is gradually starting to assimilate contemporary marketing concepts.Similarly, the prevalent themes, which are emphasized in our review, reflect growing synchronization with the social media marketing literature [5].The growing number of publication outlets, a nascent domain, high practical significance, and the many promising areas of research offer an opportune time to undertake research in PSMM.
Despite a thorough and systematic approach, the review has limitations.The search term, social media, is not the only relevant term.The terms 'new media' and Web 2.0 are also used in the literature.Similarly, we do not include the term 'social network'.However, marketing studies almost exclusively rely on the term 'social media', and we include the names of all prominent social media platforms (including SNSs) used in political marketing.Therefore, these issues have limited bearing on our review.The identification and categorization of variables were undertaken by a single coder, which is another limitation of our study.Further, the number of studies might not be sufficient for a comprehensive and exhaustive framework.The number of databases also limits our selection.Similarly, the exclusion of conference papers and book chapters limits the findings.Finally, the review is restricted by its marketing-dominant view, which is also reflected in the themes identified.

Fig. 2
Fig. 2 Geographic context of the studies

Table 1
Inclusion and exclusion criteria

Table 2
Search results

Table 3
Source of the articles

Table 4
Studies by social media platforms

Table 5
The focus of the study NoteThe number of studies does not add up to sixty-six as a few studies studied both candidate and party

Table 6
The levels of politics studiedNote The number of studies does not add up to sixty-six as a few studies studied multiple contexts

Table 7
Research methods employedNote The number of studies does not add up to sixty-six as a few studies relied on multiple methods1The systematic literature review we included in our selection covers the phenomenon of permanent campaigning, with a partial focus on social media indicators of permanent campaigning

Table 8
Data collection

Table 10
Antecedents and consequences of PSMM * Variables from qualitative studies are specified.The remaining variables are extracted from quantitative studies