Shadow Politics: Commercial Digital Influencers,“Data,” and Disinformation in India

Abstract

This article builds on ethnographic research among an emergent group of self-styled political consultants and digital influencers in India to highlight the contours of what is defined here as “shadow politics” and its implications for disinformation research and policy. Shadow politics refers to the dual structure of “official” and “unofficial” streams of campaign organization that can integrate diverse influence actors with the party campaign system. The specificity of this form of political influence management emerges from the growing uptake for digital tools and how commercial consultants anchor “data” to the goal of “narrative building” to favor their political clients, thereby delinking data practices from the moor of political moralities. It sets the stage to draw extreme content as one data type among many to choose from. Through such data practices, a vast substratum of indirectly sanctioned influence operators is linked to the public campaign stream as a “shadow,” with incentive structures to “innovate” upon excitable and inflated content.

“We have 50 ways of microtargeting voters,” claimed an influential political consultant in India, as we were drawing closer to the end of a long ethnographic interview. By taking a captivating number to the multiple tools he uses for election campaigns, my interlocutor was rhetorically conveying the precision and depth of his election operations. Throughout the interview, he listened to my questions patiently, trying to help me with as much information as he could share about his activities, most of which were guarded from full public view. In fact, he avoids media interviews altogether and hadn’t met a single researcher until that day. He had heeded to a common friend’s request and agreed to meet me when I was in Bangalore in 2020 for another round of fieldwork, to offer some insights into digital influence operations that he crafts and steers for a major political party in India. He picked a plush five-star hotel in the city for the meeting, signaling his affluence. As he got out of a chauffeur-driven car in casual jeans and a light-colored cotton shirt, I could guess he was in his late 1930s. After leading digital campaigns in several elections, he had come to a politically quieter city like Bangalore for a respite from hectic engagements. His plan was to go back to New Delhi to “enter politics.” I was intrigued that he didn’t consider digital influence activities for political parties as being “inside politics.” Demarcating a tacit boundary between actual politics and his activities as a campaign strategist, he said his job till now had been to develop online and on-ground tools and strategies that can ramp up and carefully blur “organic traction” and “manufactured traction” among voters, balancing between costs and expectations of political leaders, between what could go public and what should remain concealed in the background.

In this article, I delve into the narratives and strategies of a new class of political consultants in India and digitally enabled campaign practices that have grown around them, to examine the implications of these developments for the production and dissemination of extreme speech and disinformation in the global South context. I approach the constellation of actors, practices, and networks of digitally mediated election campaigns and influence operations as “shadow politics.” Shadow politics highlights two interlinked phenomena. First, the centrality of discourses surrounding “data,” especially promises of “data-tested” strategies that commercial political consultants advance not only to enhance their professional standing but also to convince their political clients about the value they proffer in terms of tracking and calibrating voter sentiments and boosting the brand value of political sponsors. It refers to the ways in which different strands of data—social media traces, telephone numbers, neighborhood morphologies, religious-caste-ethnic-gender data, surveys, panel data, and so on—are assembled as patchworked and discontinuous “shadow texts” (Zuboff, 2019) in political party spaces, with purported effects upon election outcomes. Second, data practices have become the means for a shadow sector of dispersed digital influence operations, including disinformation-for-hire, to latch on to an outward-facing, self-styled “professional” technology enterprise. Such diffused and “bottom-up” networks are integrated into party campaign structures via indirect sanctions and incentives to create inflamed content and intrusive distribution. A key characteristic of these networks is the vastly variegated field of actors ranging from ambitious techies aspiring to make a transition to formal politics and party loyalists who have picked up technology to climb up the ladder of power and social status to people who seek livelihood out of social media work. Thus, shadow politics throws light on the distinctive dual structure of “official” information technology (IT) wing with in-house members of political parties and public-facing consultants they hire, and indirectly sanctioned “unofficial” and dispersed influence operations. I show how the ambitions of political consultants promising the potential of “data” and the diverse networks of influence management that are drawn into the main campaign unit as the “shadow” have mobilized an underregulated field of digitally mediated operations that has sought to delink data practices from the moor of political moralities, thereby deepening the conditions of disinformation and vitriol.

The article begins by outlining recent scholarship on political marketing, digital influence operations, and the ways they are shaping the conditions and consequences of disinformation globally. In the next section, an overview of the field of digital influence activities and their growing role in electoral politics in India highlights the importance of the discourses and imaginaries of “data.” Throughout, I consider data as the “ordering” of information geared toward deliberate goals (Bycher, 2018), which involves gathering, annotating, sorting, and rendering of strands of information, as well as a widely circulating emic term with symbolic significance in political practices that employ digital technologies. Here, I build on anthropological studies that depart from considering “data” as “universalizing epistemology” and to instead “disambiguate different ways” that “data” are imagined, promised, and made to work (Douglas-Jones et al., 2021, p. 10). I outline a range of vocabularies that have emerged in relation to the promised impacts of data-backed campaigns, and how, such practices not only court “data” tied to aspirations of verifiable, partible, and “scorable” reality (Citron & Pasquale, 2014) but also paradoxically to feed its epistemic adversary, namely building the “narrative,” including by means of disinformation and extreme speech. The next section opens up the other aspect of “shadow politics”—the vast, loosely structured networks of influencers that political parties have cultivated for influence enhancement. Finally, I ask if the actually existing conditions to push back against concerted disinformation campaigns within compromised regulatory environments rest increasingly on the extent of funding that parallel and competing digital influence narratives manage to obtain in a multiparty system and how well twisted they are in the opposite direction. I conclude by discussing regulatory approaches that can account for influence actors and their practices by moving beyond platform determinism.

The analysis draws from ethnographic fieldwork I have carried out in Bangalore, Mumbai, and Delhi, including adjoining cities such as Kalyan, Noida, and Bhiwani together with research assistants, and in particular, in-depth conversations and hanging out with political consultants, politicians, and digital campaigners between 2013 and 2020, and continued conversations with consultants in 2021–2023. ¹ Political consultants we interacted with ranged from very influential high-ranking executives in consultancy firms and strategists who headed “IT” cells within the parties at the national and regional levels to entrepreneurs who had just launched their startups as well as former employees of established political consultancy firms and informal workers. They worked for or were associated with the right-wing Bharatiya Janata Party (BJP), opposition parties Indian National Congress (INC) party and Aam Aadmi Party (AAP), and two regional parties. In the years of on-ground fieldwork and virtual discussions, the relative prominence of different social media platforms has shifted, and the specific mix of strategies has been in a state of flux. For instance, extensive use of Facebook and Twitter during the 2014 general elections has given way to a more diverse mix of channels, most prominently WhatsApp, the Namo App (an application supported by the ruling BJP), Telegram, and domestic platforms such as Koo App and ShareChat. The analysis that follows offers a view of the broader trend clustering around the discourses, imaginaries, and applications of “data” and the organization of actors that crisscross platforms and the specifics of campaign strategy in different elections.

Political Marketing and “Professional” Disinformation Actors

A major stream of scholarship in political communication studies, political science and social media studies has linked the growing presence of tech-savvy political marketers and digital influence services in campaign structures as the sign of the consolidation of “political marketing”—“the utilization of commercial marketing techniques and concepts in politics” (Ahmad, 2021; Henneberg et al., 2009; Lees-Marschment, 2010, p. 1; Lilleker & Negrine, 2002; Ormrod, 2007). Studies have observed that the techniques of political marketing have spread around the world with a comparable, if not uniform, set of tactics, tools, and templates (Gibson & Römmele, 2001; Mancini, 1999) and “product ideas and consultants” (Lees-Marschment et al., 2010, p. xv). Concomitantly, this has given rise to the “ascendency of a new form of political professional: the technical expert who sells his or her services on the open market—at the heart of decision-making within political parties” (Mancini, 1999, p. 231).

Some key features of political marketing include systematic gathering and sorting of voters’ opinion through polls, focus groups, and digital data analytics; “continuous exchanges . . . with selected target groups and stakeholders as well as competitors” (Strömbäck, 2010, p. 18) through a variety of communication technologies but most prominently online channels; and a “shift of power upward and outward to leaders and external media and public relations consultants” with strategies of competitive marketing and consumer appeal (Gibson & Römmele, 2001, pp. 31–32). While not all features are present in election campaigns in different parts of the world, broader trends toward the marketing orientation are palpable. For instance, in Ghana, Kobby Mensah notes how political parties have employed research teams on the ground to conduct focus group discussions and surveys, and experts to “identify the segmentation base—geographic, demographic, behavioral, or psychographic—upon which a political market could be categorized” (Mensah, 2010, p. 198). Lees-Marschment (2010) has proposed a theoretical model to delineate three ideal types: “market-oriented party”—that identifies voter needs and expectations and incorporates them through different stages of designing the party’s “product”; “sales-oriented party”—that “aims to sell what they decide is the best for the people”; and “product-oriented party”—that does not show “awareness and utilization of communication techniques and market intelligence” (p. 9). Disputing with normative positions that assess political marketing as essentially a negative development for democracies, scholars who have contributed to this model and others who take a more optimistic view of the development have suggested that market orientation (as opposed to a purely sales or product orientation) can help parties to “identify the wants and needs of selected targeted groups in the electorate and to design a product to meet the expressed as well as latent wants and needs of these groups” (Bannon, 2005; Henneberg & O’Shaughnessy, 2007; Strömbäck, 2010, p. 19). They see the potential for market orientation and its adoption of communication technologies in political marketing to contribute directly to the requirements of deliberative democracy.

In stark contrast to such ideal-type analysis of political marketing, empirical studies on disinformation have pinned down the consequences of digital data-centric and consumer-oriented political marketing in relation to the serious threats they have posed to democratic institutions and processes in liberal democracies in the West by facilitating cross-media manipulation and engineered trends online largely carried out by commercial political consultants (Howard & Kolayni, 2016; Jakesch et al., 2021). The landmark Cambridge Analytica scandal has exposed the dangers of digital data marketing techniques and underregulated political marketing not only for election integrity but also for national security. In her investigations, Emma Briant has highlighted the growth of “a new industry of surveillance capitalism . . . that exploits online source data to carry out foreign influence campaigns” (Briant, 2021, p. 125). Consultants who “market data collection and analysis services” mine data from platforms such as Facebook and Twitter “to learn behavioral patterns and predict and manipulate them” (126). Studies have also documented Cambridge Analytica’s campaign activities in the presidential elections in Nigeria and Kenya (Ekdale & Tully, 2019). Josh Goldstein and Shelby Grossman (2021) have observed that “political actors are increasingly outsourcing their disinformation work to third-party PR and marketing firms,” adding that Facebook and Twitter “attributed at least 15 [disinformation] operations to private firms, such as marketing and PR companies” in 2020. Aside from tactical resources, such firms have offered political parties and other public figures the benefit of “deniability, since platforms “will likely lack digital evidence” to link a disinformation operation to the actual (political) sponsor (Goldstein & Gorssmann, 2021). Jonathan Corpus Ong and Jason Cabanes (2018) have shown that the expansion of the logics and techniques of political marketing has encouraged existing advertising and public relations industry in the Philippines to extend their “mastery in corporate marketing” to “an unregulated and highly profitable industry of digital political campaigning” (p. 2). They discuss how “ad and public relations (PR) executives assemble their own teams of anonymous digital influencers and community-level fake account operators” to influence voter sentiments and manipulate election time discourses, even as these actors craft an aura of respectability as professionals (p. 2).

The Indian case I examine below contributes to the observation that disinformation production is intricately linked to the expanding logics of political marketing tied to commercial political consultancy and the growing uptake for digital tools and strategies. “Shadow politics”—explored in the next sections—offers a framework to analyze linkages between commercial political marketing, digital networks, and disinformation.

Political Consultancy and Digital Influence Activities in India

With an estimated 759 million “active” Internet users accessing the net at least once a month in 2023 (The Hindu, 2023) and the vast infrastructural push for digitalization, India has seen election campaigns becoming increasingly “digitized” (Tan, 2020). The ruling BJP’s pioneering use of social media and digital campaign strategies stands as a common reference for a trend that dates back to the late-1990s when social media channels facilitated anticorruption mobilizations in urban India, shaping the rise of the AAP, a new urban party, and an emergent group of volunteers who gathered on social media to advance the cause of Hindu-first India (Udupa, 2019b). Building on early efforts to mobilize overseas and domestic Hindu nationalist volunteers via mailing lists and websites in the 1980s, the BJP’s experiments have steadily kept pace with technological development. Concerted social media campaigning that started in the run-up to the 2014 general elections with the help of a closed team of advisors and an elite group of technologists (called Citizens for Accountable Governance or CAG) has today consolidated into an indispensable part of the party’s campaign structure. Investigative reports on Cambridge Analytica have also revealed that political parties including the opposition INC recruited its parent company (Strategic Communications Laboratories and its Indian subsidiary SCL India) as early as 2003 for campaign operations to profile and influence voters (Ram, 2018).

However, beyond the party-specific effort of the ruling party and scattered efforts across different parties that would later have remarkable effects upon the way campaigns are designed and elections are fought in India, even as recently as the first decade of the new millennium, a small number of technologists who were trying to market their social media-based campaign strategies received blank stares or dismissive huffs when they approached political parties and individual politicians. In the narratives of our interlocutors, political leaders were skeptical about what social media strategists claimed they could accomplish, when they, and not these techies, knew how the constituencies behaved. By the second decade of 2000, the mood had shifted. Offering a marketer’s perspective, a data analytics professional in the field of political consultancy described this shift as the rise of “connected consumerism,” when consumers started connecting to a number of digital media devices and touchpoints—desktop computer, mobile phone, laptop, apps, and so on, thereby driving information and social media content across channels, and leaving behind traces of their activities in connected ways that could later be mapped, collated, and analyzed. This shift in consumer media technologies was the ideal time for the country’s domestic class of tech entrepreneurs to propose digital marketing and data-based solutions to the political class, set in a broader postreforms climate in India when high technology sector had become the torchbearer of India’s foray into the global marketplace (Upadhya, 2016). Together with citizen-driven mobilizations of social media channels for political cause and proactive use of digital avenues by the BJP and AAP, the technologized trends of “connected consumerism” pushed, as one of our interlocutors described, “under-the-tree discussions and cafeteria discussion . . . into online spaces on a large scale.” In their own narratives, political consultants have seen a positive shift after the 2014 general elections in how the political class has begun to appraise the potential of social media and digital influence strategies. As one interlocutor described, from an indifferent, “kuch kardo” [do something] attitude, the demands are growing sharper and regular. “They now specifically ask how many followers they have gained on Facebook or Twitter, and how much engagement a particular post received and so on.”

A range of entities populate the space of commercial political consultancy in India, and precise figures to demarcate the sector are hard to come by. Alongside international players like Cambridge Analytica that sparked a controversy in 2018 when both the BJP and INC traded allegations that their party leaders had used its services (Singh, 2019), there are also large IT companies offering data analytics solutions. In addition, there are established advertising firms such as Ogilvy and Mather which was recruited for the election campaign of Prime Minister Modi in 2014, startup companies with analytics and marketing services, older marketing firms extending their business into digital influence operations, micro-enterprises often run by just one or a handful of entrepreneurs, and dispersed groups of individuals who take up digital influence services either as third-party contractors or by directly attaching to a political party or a politician. One other trajectory of digital political consulting relates to organizations engaged in civic activism that offer a variety of services including “citizenship engagement,” “participatory democracy,” “civic training,” and training in social media use for politicians and their close aides.

The field is defined by constant flux and readjustments. New entities emerge especially during the election period, ramping up activities across disparate strands of digital influence operations. Several of them vanish after the elections or retreat to their “usual” marketing business, while stable companies stay on, and entrepreneurial consultants shift bases to regional states where elections are imminent. Consulting firms fill the interim periods between elections by offering to maintain social media handles and keep the “engagement cycle” active.

Compared with those who primarily engage in tasks such as creating memes or garnering likes online, self-presentations of professionalism are especially common among the class of workers with formal education degrees who are employed within large political consulting companies as well as entrepreneurs who offer more complex data analytics and campaign solutions like microtargeting. Such consultants frequently refer to digital election campaigns and their interactions with political consultants in the United States as ways to position themselves as professionals. Importantly, what characterizes them is the variously articulated promise of tangible impacts powered by specialized vocabularies and specific worldviews, which hinge, for the most part, on discourses of “data.”

Vocabularies of “Influence” and Discourses of “Data”

“We listen the data, we think about this data, after we think, we get into predictions,” averred a political consultant, as we listened to his grandiose invocations of “data” as the pivot of what he promises his clients. Yet another pitched it as “knowledge”: “We don’t think only about the signal, we bring knowledge.” Among campaign strategists, the buzzword is “data tested.” Often, the word “data” itself come with the symbolic weight of certainty. For instance, across poll strategists, a common reference on the national level is the case of the Indian Prime Minister Narendra Modi “taking jibes” at the opposition leader Rahul Gandhi for his privileges as a member of an elite political family whose members ruled India for close to three decades after independence. An election strategist explained that such jibes and mockery are a regular feature of campaign content because voters, especially in the younger generation, are eager to see antilegacy messages that courageously take on established hierarchies and legacy figures. A consultant in 2023 claimed that “decolonization” had emerged as the latest buzzword for right-wing campaigning to signal national pride, autonomy, and economic ascendancy. They exclaimed that decisions to infuse and improvize such tropes and rhetorical strategies “are all data tested.”

As our interlocutors explained, within large political consultancy operations, “data tested” refers to a multiunit structure of “political intelligence” gathering, data analytics, content creation, and dissemination. Profiling online users and identifying “organic influencers” are key steps in the process. One interlocutor revealed that they use as many as 150 parameters to determine influencers. The actual number of parameters depends on the social media platform, the kind of content and desired outcomes in question but they would include basic parameters of gender and age to more complex data such as spending patterns and political “sentiments” expressed on different social media channels. The aim is to trace what a strategist defined as the “multidimensional context of the person”—as a friend, a family member, a professional, and a voter. A consultant who had more than 10 years of experience in the field described how profiling voters in India is a humongous task:

For legally obtained data that I can use to profile you—if I can use that word—there are enough data points today. What is different from Cambridge Analytica is that, in a way it [the West] is a homogenous society. You know there are Republicans, there are Democrats, and issues are also mapped like that—if you are pro-life, you are here, if you are anti-abortion, you are there. Gay rights here, if not, you are there, so you are able to segregate easily. In a heterogenous society like India, and so many regional parties around, you really do not understand where it [political orientation] is. Facebook and Twitter are not the only forums to go to. There are many other forums. It could be just from buying patterns. I can do a lot of analytics on purchase patterns. I can develop a psychographic profile based on purchase patterns. These are methods that are available. If you understood x,y,z stuff I would be able to know what is going to happen.

In this rendition, the consultant followed the playbook of data-based political operations—profiling the users, data gathering, data analysis and with “x, y, z” to finally predict voter behaviors, highlighting all the while how difficult this aspired progression is when analysts deal with a “heterogenous” society like India, thereby necessitating data like purchase patterns alongside social media content.

Hitting the Bull’s Eye

Beyond online data mapping and analysis, commercial consultants organize focus group discussions. An influential political consultant described what they organize as “heavy focus group discussions” with “all age groups,” from rural and urban areas, men and women. Focus groups are combined with what he described as “ethnographic studies” which are carried out to “understand” behavior. Following this, the strategy team assembles at a “whiteboard” to identify common issues and develop an “issue board” to map the issues that are “plaguing” the constituencies. Based on these themes, they would then conduct a “large quantitative study” to confirm whether they are “hearing the right things from the ground” or if it is an exaggerated or distorted version found in an intense focus group. Subsequently, “data-tested” model that builds on these activities and insights suggests phrases for the party leader’s speeches and rallies, and tests them on social media. Social media then “becomes a tool to measure the success of what they [politicians] are trying to say,” examining if people on social media “are echoing that sentiment or not.”

“There are times when it will go flat,” our interlocutor admits,

Nobody will have noticed that we said something of that sort because it was not an issue [for people]. But eight out of ten times, what we are trying to say hits the bull’s eye . . . where you start seeing the reactions on social media exactly on sentences that you wanted them to make an attempt [focus] on.

The desired outcome is an “upsurge of response.”

Panels

Social media is just one way to test whether there is a reverberation. It is combined with what campaign strategists describe as “mirroring panels” with actual people who represent a “huge spectrum from core BJP voters to core Congress voters and regional party voters.” To maintain “neutrality,” panel members are not given any clue that the data would feed a political campaign, and the questions would be developed within the framework of commercial marketing surveys designed for consumer products. Except, this time, the questions are about Delhi riots, as one strategist described in detail:

I put a panel of 3000 people together, and I start doing various things, for example, today something happened on Delhi riots. I go back to the panel and ask something about the Delhi riots. You set [organize] the panel to ask various questions to kind of understand where they are on social issues, who they vote for, what they think about various issues. You have a basic understanding of the panel in any case. Then you go back and start working on these issues. After this, someone physically goes out . . . the word is not “recruit” but “screens” them and says, Listen, we are doing academic research, we are trying to do a panel, we are just going to give you a call once in a week and ask a few questions and if you are fine with it, we are happy to gift you something.

Commercial marketing agencies described as “professional agencies” are employed to “recruit” the panel, gather responses at regular intervals, and reward participants with discount codes and petty cash—a practice common in commercial marketing surveys.

Grid

In large political parties and consultancy firms, sample groups for the panels and focus group discussions are drawn from “internal research” teams. These teams have produced an elaborate “grid” to design the samples that suit different campaign objectives in different regions. “We have developed a grid,” said a consultant, “so if we have to put together a sample which will get us a fair understanding of India . . . We have tried to sort of include every possible profile [in the grid].”

Across these instruments, “data” are offered as quantifiable scores churned from surveys and software scripts that would yield a precise toolkit for campaigns. Data emerge, to follow Antonia Waldorf, “a form of potential—it is value because . . . it can be transformed into something else” (Walford, 2021, p. 127). Promises of knowledge derived from “data” and similar grandstanding evince an ambition, if not the realization, of what Shoshana Zuboff (2019) describes as “total certainty” powered by “data” as a means to shape voter loyalties by eliciting, mining, slotting, and calibrating their opinions both via offline and online means of engagement.

Quantified and objective as it may be claimed, this exercise is also laced with interpretations, gut feelings, conjectures, and approximations, as some former employees of consultancy firms reveal. Interpretative gymnastics is especially evident in how consultants understand political leaders’ actions, voters’ ideological affiliations, and the public mood in general. Through various steps and instruments, “data” are made meaningful with interpretation, and paradoxically, aligned to the goal of “narrative building,” its epistemic opposite. In this nexus of narrative building, data become not only partible and scorable, but also selectable. Disinformation and extreme speech become important data types to select from. They are calibrated for campaign content based on the assessments of the campaign management teams and politicians around how far the content should tilt toward “negative” or “positive” narratives in a given electoral contest. From riots and jokes about politicians to policy issues such as taxation and mighty concepts such as decolonization, the panel structure, and associated instruments of grids, polls, and focus groups as well as interpretations and augmented debates on social media feed the effort to influence “narratives and perception” in a manner similar to any other consumer product—a sort of data relation that is unmoored from political moralities.

Within the professional practice of political consultancy, there are active efforts to purge morality from the day-to-day operations of data-driven commercial consultancy. “Our interest is business,” stated a consultant, “We’ll have conflict of interest if we stick to one party.” The list of political clients for one of our interlocutors, for instance, ranged from BJP, INC, and regional parties such as Telangana Rashtra Samithi (TRS), Sikkim Democratic Front (SDF), and Bahujan Samaj Party (BSP)—representing different and often opposing political positions—alongside a motley mix of individual politicians.

Some strategists who call themselves “politically agnostic” have reconciled to commercial consulting by “suspending” their opinion, as one interlocutor reflected:

The minute I started working at a political consultancy, the first thing I did was to train myself not to have an opinion because it started interfering with my day-to-day life . . . in the sense that, “Oh my god, how can I work for a person like this or a party like this”? That is when I had to choose—if I want to be in this field, I have to leave my personal opinions. Some level of maturity crept in and now I don’t have opinions at all . . . Since the time I have understood this, I have never had an issue. I have worked with people across party lines and this neutrality of my own has actually helped my mental health a lot.

A young law graduate who had just quit a political consultancy firm told me how there is little room for anything other than meeting the deliverables. “Deliverables come from your reporting manager,” he said, “. . . you know it’s a corporate setup, you have to show you have met the deliverables at the end of the week, at the end of the next day.” Extreme speech emerges and folds into this self-styled professional space, as the singular goal of campaign success and project deliverables obscures other planes of judgment to assess the implications of one’s activities (see also Ong & Cabanes, 2018).

Star strategists like Prashant Kishore became the face of this type of business-first digital influence operation in India (Biswas, 2021), but a handful of similar influential figures represent an elite group of strategists who combine online and offline campaigns, amassing media visibility, and subsequently, also developing a political voice in favor of the party or the leaders they serve. Below this creamy elite lies a vast stratum of mid-sized and micro-commercial players offering influence enhancement services to interested political parties and individual politicians. This highlights the importance of the second aspect of shadow politics—the vastly dispersed networks of informal influencers.

Shadow Politics

Although different political parties have chased the ambition to achieve “total certainty” about voter opinions with efforts to influence their voting decisions, digital strategy of major political parties, including the ruling party, stands out for its vast and complex organization. The official stream of digital campaigns—known as the “IT cell”—is headquartered in New Delhi but with state (subnational) and local units around the country. The official network channel is centrally controlled by the IT cell in New Delhi for content related to national-level issues. As one interlocutor described,

Politicians now realize that whether they are rural or urban MLA [elected representative in the state legislative house], they need to have an account [on social media] and they need to propagate the same four lines that the central leadership parrots out to everybody.

State and local units are allowed to compose messages for state-level issues but are strictly instructed to keep them within the ideological ambit of the party, while devoting significant energy to amplify content that comes via the central IT unit. Top-down content is largely iterated through a combination of surveys, panels, and selective content testing on social media, as described in the previous section.

This controlled centricism is, however, strategically positioned in parallel to a more dispersed, flexible structure that can draw and retain volunteers, and grow at various inflection points synchronized with elections and other critical events. To distribute campaign content, such networks, especially on WhatsApp, connect “official” workers with other official workers, and connect official workers with “general sympathizers” and “well-wishers,” allowing sufficient space for official workers to draw “general sympathizers” to become more committed, and general sympathizers to draw other sympathizers and fence-sitters. The flexible parts of this network are not edges of a single core but constitute connected nodes of content building and influence enhancement. They include ideologically motivated technologists who utilize their business units and in-house employees to take up some “party work” as a declaration of their commitment to the party or to gain and establish political “connections” that would help them to stay in or grow business. One interlocutor in Mumbai, for instance, owned a small firm that designed marketing campaigning for Bollywood movies, but owing to his family’s close connections with the ruling party, he also offered “sentiment analysis” and social media mapping for the party’s regional unit.

Some consultants describe such networks as “the unofficial wing” of the party’s campaign structure. It is the party’s internal system that nurtures the so-called social media voluntary support and a vast number of “third-party pages,” “unofficial pages,” proxy workers, and clickbait amplifiers through an elaborate network of incentives and layered recruitment. Actors in the “unofficial” wing rarely have a clear view of the party’s “official” social media structure. One such digital influence worker remarked that the party’s social media structure was still a puzzle, even though he had been contributing to a local “IT Cell” for several years. “This is regulated at a very internal level,” he said,

We are not allowed to discuss these things, but broadly, the central cell sends instructions to local offices for recruitment. Through these local channels, people come in contact with the party and then they are provided with small monetary incentives to the tune of two to three thousand rupees.

He continued that once websites and Facebook pages are ready, the party does not give any more cash rewards. But they help us to realize some revenue through these [social media] platforms by making websites or portals. As we get to know people at the local level, it helps us in getting jobs for our friends and family. As you rise in position and influence, the sources of revenue increase and you get many more opportunities, both offline and online. Both financial and power-based opportunities, you know how it works in India, he intoned with a smile, hinting at local networks of patronage that he was able to activate by directly managing 10 Facebook pages for the party, aside from overseeing several more pages across Northern India.

To this “unofficial” stream, vast gray networks that operate at the local levels supply information ranging from telephone numbers of voters to microdetails about the locality including, in some cases, electricity bill amounts paid by the households that can help profile the voters. Crucially, the unofficial wing is allowed to “innovate” upon content, ramp up provocative language, and twisted accounts. In the thick rush of Mumbai traffic, late in the night in 2019 near the Lokhandwala circle, a political party worker told us inside a taxi, as we were heading toward the official residence of the politician he worked for: “These factories [for inflamed campaign rhetoric and influence augmentation] are underground . . . I mean sometimes literally underground, they have offices in the basements and rarely people know about them.”

The official and unofficial units are kept separate in the organizational structure of digital campaign work. They cooperate but in the organizational structure, they do not report to one another. Throwing more light on how this structure operates, one election strategist revealed that the two units work “by synchronizing . . . a kind of sync is achieved.” In other words, the two wings coordinate but behind opaque lines of communication and not through a formal organizational structure. The unofficial wing could then be defined as a “shadow”: a shadow that follows but does not have the status of the body.

Describing digital influence practices, including disinformation-for-hire, which have evolved through such a dual structure as “shadow politics” highlights the significance of “informal” or “gray” zones where extreme content and actions are experimented, sponsored, and disseminated but defining them as a “shadow” clarifies that such units follow the official unit and are inseparable from it. In other words, gray practices and networks—and the extreme content produced through them—are unofficially sanctioned, elaborately financed through diverse incentives, and are critical to digital influence operations. The specificities of digital influence operations that have evolved through such a dual structure cannot be fully captured by concepts such as “informal politics” to designate precarious brokerage and diversified political actions outside the realm of party politics (Desai, 2002; Piper & von Lieres, 2015) because of the digitally networked integration of unofficial units into party’s “main” campaign structure. The networked and incentivized nature of shadow streams also marks their distinction from “mass” grassroots support for political parties. “Mass support” cohering around ideologies, charismatic leaders, or the populist axis of elite and people (Brubaker, 2017) is not organizationally integrated with regular party campaign infrastructure through varied incentives. However, although analytically distinct, “mass,” “shadow,” and “informal” support merge in practice, and even on a phenomenological level, when—in the life cycle of their party engagement—supporters receiving incentives begin to adore their leaders and become ready to forego the former.

Here, I build on Madhava Prasad’s (2009) theorization of “shadow sovereignty” that refers to extra-parliamentary authority that cinema celebrities in southern Indian politics wield as legitimate leaders of “the people” and signal a similarly “extra-legal” parallel structure of influence that “unofficial” wings of digital campaign networks wield. Shadow politics refers specifically to the manner in which dispersed and “bottom-up” networks are integrated into the party campaign structure via indirect sanctions and incentives. In the postcolonial context, shadow politics of digital influence operations is not a radical rupture with existing practices and long-standing conventions of campaign organizing. It does not dismantle but complement patronage tactics of “distributing money, distributing liquor, distributing salaries,” where they occur, as part of instrumental exchange and moral obligatory attachments between voters and leaders (Piliavksy, 2014). Digital influence strategies and shadow politics have instead emerged as an effort to cut competition above what is taken as the reality of favor distribution. ²

Conclusion: Implications for Disinformation Regulation

This article has defined the constellation of actors, networks, and practices of digitally mediated influence operations as “shadow politics.” On one hand, shadow politics highlights the ways in which different strands of data are assembled as “shadow texts”; but unlike corporate reality mining in Zuboff’s (2019) study, they are patchworked and discontinuous and they occur in political party spaces. Data here emerge not as a “universalizing epistemology,” but as anthropologists Rachel Douglas-Jones, Antonia Walford, and Nick Seaver (2020) grippingly state, “data’s multiplicity and particularity in diverse local worlds,” and the “different ways that data are talked about and done” (p 10). In the most recent election phase (national elections planned in April-May 2024), “data” narratives, alongside multimodal manipulations in the form of deep fakes, are also articulated in relation to artificial intelligence (AI) assisted services.

Such ways of talking about and doing with data among commercial political consultants have led to strategies that delink data practices from normative considerations about the implications of their activities for democratic processes and institutions. Extreme speech becomes available as one of the many choices in the campaign design for commercial influence operations. Precisely through such data practices, commercial operations overlap with and draw from vastly dispersed gray practices of amplifying party-favoring content through the dual structure of official and unofficial streams that develop as the body and its inseparable shadow. Shadow politics inflects existing electoral practices in the postcolonial context by offering data-backed strategies of “narrative building” and novel ways of integrating variegated influencers into the party campaign stream. As both data practices and indirectly sanctioned networks have become common features of (dis)information operations (Mare et al., 2019; Ong & Cabanes, 2018; Wasserman & Madrid-Morales, 2022), the analytical purchase of “shadow politics” might resonate beyond the postcolonial context with rapid digitalization examined here. For instance, Erkan Saka’s (2021) study on “AKTrolls” who favor President Erdogan in Turkey has demonstrated that troll networks are decentralized. In Brazil, Albuquerque and colleagues describe how, after the ascent of Bolsonaro to presidency, the “boundaries between institutional and anti-institutional agents” have blurred in terms of networks that produce and disseminate disinformation (Albuquerque et al., 2022, p. 3).

In terms of regulation, as political parties compete to gather data and coordinate narratives for digital influence, the actually existing conditions for mitigating disinformation come down to how much traction parallel and competing digital influence narratives manage to gain in a multiparty system and how well twisted (and distorted) they are to inject inverse effects. Echoing an intention that was common among several consultants and social media managers we had met, a party worker in Mumbai declared determinedly, “They [rival political party] run a campaign, we run a reverse campaign.”

The grim prospect of competing campaigns utilizing extreme content as a necessary ingredient in the data-narrative matrix is accentuated by the absence of coordinated, multiprong regulatory system for “online electoral governance” (Tan, 2020, p. 214). Regulatory controls over digital influence operations in India include the model code of conduct during the election time that bans campaign activities, caps on election expenditure, legal provisions against cybercrime and general criminal acts, and restrictions on interactive voice response or bulk short messaging services governed by the Telecom Regulatory Authority (Madan, 2022). In addition, content that is produced or amplified for digital influence also hits the filters of content moderation practices of social media platforms. These measures, however, are shown to have little effect in containing disinformation and extreme speech in the current regulatory climate (Tyagi, 2022; Udupa, 2019a).

Importantly, a vastly dispersed array of actors mobilized through official and unofficial networks, alongside equally diverse field of commercial political consultants stress the need for regulatory mechanisms in the global South contexts to go beyond platform determinism. The analysis around shadow politics challenges some of the regulatory energies that heavily assign platform accountability while not directing equal attention to industry/actor accountability in the disinformation space (Caplan, 2018; Ong, 2021). Transparency in election expenditure, regulations for campaign finance, professional code of conduct and co-regulatory models for commercial political consultancy, training and awareness raising among micro-entities, and resource allocation for timely factchecks are some aspects of industry/actor accountability that require added focus in the coming years.