Micro-Targeted Pseudo-Local News: Computational Analysis of Metric Media’s Digital Propaganda Network

AJ Cordeiro

Toronto Metropolitan University

Abstract

Local journalism’s decline has fueled “pseudo-local” news networks that mimic community outlets while advancing partisan narratives. This paper examines Metric Media, a large U.S. network, from a master’s thesis case study in micro-targeted digital propaganda. Using computational methods (sentiment analysis, topic modelling, and social media metrics) on content from 2019 to 2024, the study identifies thematic patterns and strategic tactics. Findings indicate extensive algorithmically generated content, a partisan agenda, and targeted distribution, with limited engagement. The analysis demonstrates how data-driven transparency can expose covert propaganda and inform policies to strengthen local news and democratic discourse.

Introduction

Democratic societies rely on robust local journalism to inform communities and hold power accountable. In recent years, however, the collapse of many local news outlets has created “news deserts” and information voids. Exploiting this gap is a new phenomenon: pseudo-local news networks. These operations establish websites that superficially resemble independent small-town news sources, often adopting generic community names, but are centrally controlled and funded by political or corporate interests. By leveraging the higher public trust traditionally enjoyed by local media (Benjamin, 2022), pseudo-local networks disseminate targeted propaganda under the guise of hometown news.

A prominent example is Metric Media, a U.S.-based conglomerate of hundreds of local-looking news sites launched in the late 2010s. Investigations by media researchers found that Metric Media’s network spanned approximately 450 sites in 2019 (Bengani, 2019) and expanded to more than 1,200 sites across all 50 states by mid-2020 (Bengani, 2021). These sites publish enormous volumes of localized stories on topics such as school events, real estate trends, and municipal announcements, which are largely generated by algorithms or bots, aiming to create the appearance of bustling local journalism. Intermixed with this neutral local content is a small fraction of politically charged stories aligned with conservative talking points, often authored by affiliates rather than independent reporters. This strategy creates a “pink slime” journalism effect: a blend of formulaic filler and partisan messaging that is difficult for readers to distinguish from genuine community news (Bengani, 2019).

The rise of such pseudo-local outlets raises urgent concerns for democracy. Scholars note that disinformation and politicized falsehoods, when laundered through credible-seeming local channels, can erode civic discourse and public trust in media institutions (Bennett & Livingston, 2018). Unlike overt fake news on social media, pseudo-local news leverages the structural authority of journalism, using familiar news formats and benign local updates, to subtly prime readers on divisive issues. For example, a Metric Media site might repeatedly post neutral-sounding updates on voter registration deadlines alongside op-eds warning about voter fraud, thus feeding a partisan narrative under a local veneer (Bengani, 2021). These tactics allow propagandists to micro-target communities (by geography or demographics) with tailored messages that often escape the scrutiny given to national media campaigns. Indeed, by 2022, the proliferation of fake local outlets was on pace to outnumber legitimate local newspapers in the United States, highlighting the scale of this trend (Abernathy, 2022; Benjamin, 2022).

This paper examines Metric Media’s pseudo-local news network from a master’s thesis case study to understand the structural patterns and influence of micro-targeted digital propaganda. It employs computational journalism techniques, using large-scale content scraping, sentiment analysis, topic modelling, and social media data analysis, to systematically “reverse-engineer” Metric Media’s content strategy and its reception. The goal is twofold: (1) to illuminate how such networks function and influence public opinion at the local level, and (2) to derive insights for communication policy on combating covert propaganda while strengthening authentic local media. By marrying data-driven analysis with communication theory, this research aims to contribute to a growing body of policy-relevant scholarship on misinformation, media trust, and democracy (Cordeiro, 2025). The remainder of the paper is organized as follows: first, a review of relevant literature on local news, misinformation, and propaganda. Next, an outline of the methodology, followed by the empirical findings on Metric Media’s content and engagement patterns. A discussion follows, regarding the implications of these findings for democratic accountability and media policy. Finally, policy interventions are proposed, and the paper concludes with reflections on safeguarding the information commons in an era of “pink slime” journalism.

Literature Review

Locals News, Trust, and Democratic Vulnerabilities

The foundational role of local journalism in democratic societies has been well documented. Local news outlets not only inform citizens about community affairs but also act as watchdogs over municipal governments, schools, and businesses. When local journalism declines, studies have found increases in government inefficiency and corruption, lower voter turnout, and a citizenry less equipped to engage in civic matters (Pickard, 2020; Napoli, 2019). Meyer et al. (2019) observe that the loss of credible local news creates a populace “starved of community-specific information,” leaving a vacuum that alternative sources can fill (Meyer et al., 2019). Unfortunately, this vacuum is increasingly being occupied by actors with political agendas. As traditional media falter, partisan publishers have launched pseudo-local outlets designed to look like news but serve as vehicles for propaganda (Benjamin, 2022). The public’s enduring trust in local media is being weaponized: people tend to give local-branded news the benefit of the doubt, which malicious actors exploit by disseminating biased content through local facsimiles (Benjamin, 2022).

This trend intersects with the broader phenomenon of mis- and disinformation in the digital age. Wardle and Derakhshan’s (2017) framework on “information disorder” distinguishes between misinformation (unintentional falsehoods), disinformation (deliberate falsehoods), and malinformation (harmful truths) in today’s media environment. Tandoc, Lim, and Ling (2017) note that “fake news” comes in various forms, including fabricated stories, partisan propaganda, and clickbait (Tandoc et al, 2017). The pseudo-local model exemplified by Metric Media aligns with the propaganda category: it uses real information (e.g. factual local statistics or events). However, it packages it with deceptive intent and sponsorship. By hiding political advocacy behind the aesthetic of routine local reporting, these outlets muddy the waters between genuine journalism and partisan influence operations.

The term “pink slime journalism” has been coined by reporters to describe these low-cost, mass-produced local news lookalikes (a reference to processed filler in meat) (Bengani, 2019). Initial journalistic investigations into pink slime journalism were pivotal in exposing Metric Media and similar networks. In late 2019, the Columbia Journalism Review reported an “intricately linked network” of about 450 pseudo-local sites that were publishing algorithmically generated stories combined with conservative political content (Bengani, 2019). By mid-2020, this network had nearly tripled to more than 1,200 sites nationwide (Bengani, 2021). Such growth is not organic, but rather is fueled by strategic investment from advocacy groups and political donors (Bengani, 2024). These findings underscore that pseudo-local networks are not random or grassroots but rather a deliberate propaganda infrastructure designed to influence public opinion at the local level.

Propaganda in the Digital Age: Micro-Targeting and Algorithms

The activities of Metric Media and similar networks can be viewed through classic and contemporary theories of propaganda. Jowett and O’Donnell (2012) define propaganda as “the deliberate, systematic attempt to shape perceptions, manipulate cognitions, and direct behaviour to achieve a response that furthers the desired intent of the propagandist.” In the past, propaganda efforts often relied on centralized mass media or overt political advertising. Today, as Herman and Chomsky’s (1988) Manufacturing Consent and Bennett and Livingston’s (2018) disinformation order thesis suggest, propaganda has evolved into more insidious forms that exploit media fragmentation and the personalization of online content (Herman & Chomsky, 1988; Bennett & Livingston, 2018). Instead of broadcasting a single message to a national audience (the traditional mass-propaganda model), propagandists can now micro-target specific segments of the population with tailored messages delivered via niche channels. This transformation is enabled by the same digital platforms and data analytics that drive modern advertising.

Metric Media’s model exemplifies micro-targeted propaganda in action. By establishing separate local news sites for each state, county, or city (often with names such as Springfield Times or Mobile Courant), the network can serve customized content to each community. Much of this content is automatically generated (Cordeiro, 2025). This provides a facade of normal local journalism. This assembly-line news production enables the network to scale rapidly and cost-effectively across locales (Cordeiro, 2025). Crucially, it also provides cover for the injection of partisan narratives. The propaganda is camouflaged amidst voluminous benign coverage: only a small percentage of Metric Media’s articles carry an overt political bias, and those that do are often written in a measured, informational tone to avoid detection (Cordeiro, 2025).

Another hallmark of modern propaganda in this context is the use of platform algorithms and data analytics to target messages. Social media and search engines are key distribution channels for pseudo-local news. This strategy leverages what Zuboff (2019) terms surveillance capitalism: the vast collection of personal data by platforms, which can be repurposed to micro-target content or advertisements (Zuboff, 2019). The underlying logic is the same as in commercial micro-targeting, but the product being sold is a political ideology or candidate.

Research on social media’s role in misinformation and disinformation provides additional context. Studies have found that false or hyper-partisan news often spreads through networks of aligned communities and can be amplified by recommendation algorithms that favour engagement (Vosoughi, Roy, & Aral, 2018). However, pseudo-local outlets often have modest followings, and their content may not naturally trend organically (Cordeiro, 2025). Instead, these outlets rely on orchestrated dissemination: using multiple low-visibility pages, coordinated posting, and sometimes external boosting (by political organizations or automated bots) to ensure their narratives circulate.

Emerging Countermeasures and Gaps

Awareness of pseudo-local news networks has grown among journalists and scholars, prompting discussions on how to counter them. Traditional fact-checking is insufficient here, as many individual stories on these sites may be factually correct or only subtly slanted, thereby not triggering obvious flags. The issue is more about structural deception (who is behind the information and why) rather than clear falsehoods in the content. Consequently, countermeasures emphasize transparency and source verification. For example, researchers and watchdog groups have begun compiling lists of known “pink slime” sites and their ownership. The activist-curated MassMove dataset (2024) is one such effort that tracks deceptive local news networks and was used in this study (MassMove et al., 2024). These open-source intelligence (OSINT) approaches help map the scale of the networks and provide data for analysis.

Media literacy initiatives form another line of defence. Hobbs (2011) and other digital literacy experts argue that educating the public to critically evaluate news sources through checking masthead information, researching a site’s background, and being wary of one-sided local outlets is vital in an era of information disorder. The challenge, however, is that pseudo-local sites are specifically engineered to appear innocuous, bypassing the heuristics that readers might normally use to spot fake news. As Darr (2024) points out, many consumers cannot easily distinguish these partisan outlets from genuine local journalism, especially when the content is not blatantly false or sensational. The credibility by association with “local news” is precisely the point (Benjamin, 2022).

The literature indicates that pseudo-local news networks like Metric Media sit at the intersection of several contemporary issues: the collapse of local journalism, the spread of online disinformation, and the sophisticated targeting capabilities enabled by digital platforms. They represent a structural and policy challenge rather than just a series of misleading articles. This study builds on this foundation by providing empirical analysis of Metric Media’s content and strategies, thereby informing potential policy responses to this emergent threat to democratic communication.

Methodology

Research Design and Data Collection

Adopting a case study research design (Yin, 2018), this work focused on Metric Media’s network of pseudo-local news sites. The approach was multi-method, integrating computational journalism techniques with content analysis and network analysis. The study period spanned July 2019 to January 2024, covering multiple election cycles and providing a longitudinal view of the network’s activity. The primary dataset consisted of text scraped from Metric Media websites, and the secondary dataset comprised social media data (Facebook pages) associated with those sites.

• Identifying Metric Media Sites: Using an open-source dataset compiled by an activist researcher collective (MassMove, 2024) to identify known Metric Media outlets, this research gained hundreds of domain names and site titles affiliated with Metric Media or its sister networks (sometimes branded as Local Government Information Services, Newsinator, etc.). Using this list as a starting point ensured comprehensive coverage and mitigated the risk of missing less prominent sites (Cordeiro, 2025). For each site, an RSS feed was identified, which is a publicly available web feed that syndicates recent articles. RSS (Really Simple Syndication) feeds provide article metadata (titles, summaries, timestamps, and sometimes author names) in a standardized XML format (O’Reilly, 2007). By using RSS feeds, this work systematically harvested content updates from each site without requiring direct web scraping of the human-facing pages. This approach is both efficient and respectful of site integrity.
• Web Content Scraping: A custom Python-based scraper was developed to iterate through the list of Metric Media RSS feeds. Over several iterations from late 2023 to early 2024, the scraper collected all available feed entries (going back to each feed’s inception or latest archive). In total, thousands of articles were gathered from over 1,000 Metric Media websites. The data for each article included the publication date, headline, a summary or the first paragraph (as provided by RSS), and the source site name. These entries were compiled into a master dataset. For consistency, duplicate entries and non-English content (if any) were filtered out, though Metric Media’s content is almost exclusively English. By the end of data collection, the researcher had compiled a database of approximately 7,500 article entries, representing a substantial sample of Metric Media’s output during the study period (Cordeiro, 2025). All data collected was publicly available, and no private data was accessed, consistent with ethical norms for OSINT research (Townsend & Wallace, 2017).
• Social Media Data: To assess how Metric Media disseminates content and engages audiences on broader platforms, this work also collected data from Facebook, where many of the local sites maintain pages. Using a combination of Facebook’s web interface and a scraping tool (Zúñiga, 2023), this study retrieved posts from pages associated with Metric Media outlets. This yielded a dataset of 7,510 Facebook posts from Metric Media pages between April 2017 and February 2024 (Cordeiro, 2025). Each record contains the page name, the post content (typically a headline or caption with a link to the article), the timestamp, and engagement metrics, specifically the numbers of likes, shares, and comments on the post. These metrics were collected by parsing the publicly visible counters on each post (without using any private API), ensuring that the research work only gathered data available to any Facebook user (Cordeiro, 2025). X/Twitter was explored but found to be minimally used by these outlets (many had no X/Twitter presence or had abandoned accounts), so the social analysis focused on Facebook, which is more popular for local news distribution (Cordeiro, 2025).

Analytical Techniques

The analysis proceeded in three main components: sentiment analysis of content, topic modelling/clustering of content, and social network and engagement analysis of the Facebook data. Each component was designed to address specific research questions about Metric Media’s operations:

• Sentiment Analysis: The case study assessed the emotional tone of Metric Media’s content to determine whether it skews positive, negative, or neutral. Given claims that these sites maintain a neutral façade (Cordeiro, 2025), this analysis provides quantitative validation. The research used a lexicon-based sentiment analysis tool (TextBlob in Python) to calculate sentiment polarity scores for each article summary. The overall distribution of sentiment provided insight into whether the network primarily disseminates emotionally charged content or maintains a straight-news tone.

• Topic Modelling and Cluster Analysis: To uncover the major themes in Metric Media’s content, this work employed unsupervised topic modelling. Topic identification was conducted using TF–IDF vectorization followed by unsupervised K-means clustering. These clusters were then interpreted qualitatively to identify recurring narrative structures across the corpus. The study then experimented with different numbers of topics; based on coherence measures and interpretability, the researcher settled on a model with five dominant topics (Cordeiro, 2025). These topics were characterized by sets of keywords, which the researcher interpreted and labelled (for example, one topic’s top terms included “county, population, census, age, increase”, which was labelled as a Demographics theme). The five thematic clusters identified were: (i) Demographic Change, (ii) Religion & Culture, (iii) Civic Governance & Crime, (iv) Economy & Business, and (v) Education Policy. The research then qualitatively analyzed representative articles from each cluster to understand the narratives being promoted.

• Social Media Engagement and Network Analysis: The Facebook dataset was analyzed to evaluate how Metric Media’s content is disseminated and received on social platforms. The study first mapped the network of Facebook pages: each page corresponded to a pseudo-local site, and regional patterns were noted (e.g., multiple pages managed by the same central account, although these details are not fully visible externally). The research then examined posting frequency over time. By aggregating monthly post counts, the study identified bursts of activity, notably spikes in the lead-up to the November 2020 U.S. election and the 2022 midterms. These spikes suggested a concerted effort to push content during election seasons, consistent with prior reports that these networks increase politically relevant output at election time (Bengani, 2019; 2021).

For engagement metrics, the study calculated the average number of likes, shares, and comments per post for each year to identify trends (Cordeiro, 2025). Overall engagement was relatively low: a typical post garnered only a handful of likes and perhaps a share or two (Cordeiro, 2025). This supports the notion that the network’s influence may derive less from broad grassroots engagement and more from strategic amplification (Cordeiro, 2025).

Throughout the analysis, the research employed a triangulation strategy: comparing our computational findings with external investigative reports and theoretical frameworks. For instance, when the topic model identified a cluster of religious-themed content, it was recalled that Bengani (2024) reported that Catholic organizations partnered with Metric Media to promote certain views (Cordeiro, 2025). This helped confirm that the cluster was not a random artifact but tied to a known influence campaign (Cordeiro, 2025). The research similarly cross-referenced any surprising patterns with media reports to avoid overinterpretation (Bengani, 2019, 2021; Bartholomew, 2022). By integrating computational data with journalistic evidence, the work aimed to ensure validity and guard against confirmation bias: if our data-driven results aligned with independent reports, our confidence in those results increased; if they diverged, we treated them cautiously (Bartholomew, 2022).

Limitations

Every methodology has limitations, and this study is no exception. One constraint was data completeness: despite the extensive scraping, the research may not have captured all Metric Media sites or articles (new sites may have been launched after our dataset was compiled, or some older content may have been excluded from RSS retention). The focus on RSS summaries means that the study analyzed the gist of articles rather than the full text. However, summaries usually contain the key information, some nuance may be lost. Additionally, sentiment analysis using a lexicon (e.g., TextBlob) can misclassify sarcasm or context-dependent tone (Cordeiro, 2025). However, given that Metric Media content is largely straight-faced and report-like, this risk is minimal (and the neutral-majority finding is so pronounced as to withstand minor misclassifications) (Cordeiro, 2025).

On social media, Facebook engagement counts are a proxy for reach but not a direct measure of influence. The study also lacked data on whether Metric Media used Facebook’s paid promotion, as the engagement analysis is based on organic interactions (Cordeiro, 2025). If the network boosted posts through ads, the reach could be larger than the public metrics suggest, a factor not fully captured here (Cordeiro, 2025). Another limitation is that causal effects on public opinion are beyond the scope of this study. The research identified content patterns and potential intent, but cannot conclusively determine the extent to which these pseudo-local narratives changed people’s beliefs or actions (Cordeiro, 2025). That would require survey or experimental research on audiences.

Finally, from a methodological standpoint, it is important to note that the research approach itself is part of an emerging paradigm of countering disinformation. By using the very tools of big data and analytics that propagandists employ (for micro-targeting), the study turned them toward analysis and accountability (Cordeiro, 2025). This approach needs to be continually refined. Nonetheless, within these bounds, the research’s methodology provided a robust examination of Metric Media’s network and offered insights that purely qualitative or purely quantitative approaches might miss when used in isolation.

Findings

The analysis of Metric Media’s pseudo-local news network yielded several key findings regarding its content characteristics, thematic agenda, and distribution strategy. This work presents the findings in three parts: (1) the nature of content and its sentiment/profile, (2) the major thematic patterns (topics) identified across the network’s output, and (3) the distribution and engagement patterns on social media. Together, these findings paint a picture of a highly coordinated propaganda operation that blends into the local information ecosystem while advancing a partisan narrative.

1. Content Characteristics and Sentiment Profile

• Automated Local News Fabrication: Metric Media sites overwhelmingly relied on automated content generation to populate their pages. The data confirmed that the vast majority of published items are boilerplate local “news” stories created with minimal to zero human journalism (Cordeiro, 2025). By churning out thousands of such hyper-local tidbits (often dozens per site per day), the network creates a dense fog of local information that serves as camouflage (Bengani, 2021).

• The sentiment analysis provided quantitative evidence for what journalists had qualitatively observed: most content on these sites is written in a neutral, just-the-facts style. Sentiment analysis revealed that approximately 62% of articles were classified as neutral, 27% as positive, and 11% as negative, suggesting a predominantly neutral tonal presentation (Cordeiro, 2025). The predominance of neutral-toned content is consistent with the strategy of appearing non-partisan and credible. It also indicates that much of the auto-generated content consists of mundane community reporting that naturally exhibits no strong sentiment. The relatively small proportion of negative articles suggests that overtly critical or alarming content is used sparingly, likely reserved for topics where a negative frame serves a particular agenda (e.g., crime reports that might stoke fear or articles criticizing a policy). This pattern of selective negativity is a hallmark of stealth propaganda: most of the time, the network wears a friendly local demeanour, but at key moments it injects negativity or fear to influence readers’ perceptions of certain issues (Benkler et al., 2018).

• Human-Written Partisan Content: Interspersed within the automated output, Metric Media sites do include a smaller subset of articles that bear human bylines or are clearly written with editorial intent. These are the pieces that advance political or ideological viewpoints. According to Bengani (2021), only about 10% of content on such sites is original journalism with a human byline, and the study’s observations are consistent with this estimate. This mirrors Bengani's (2019) findings: content choices often align with election cycles and partisan narratives, suggesting behind-the-scenes coordination.

In essence, the content strategy is one of volume and veneer. Metric Media floods local channels with neutral, automated news to build trust. Then it adds a drip of partisan messaging, enough to influence, but not so much as to immediately reveal the site’s true nature. This subtlety is arguably more dangerous than outright disinformation because it does not elicit skepticism among casual readers.

2. Issue Agendas and Thematic Patterns

The topic modelling and clustering analysis identified five major thematic clusters in Metric Media’s content. These represent the network’s de facto editorial agenda, the recurring subjects and narratives that it emphasizes across different locales. Significantly, each of these clusters corresponds to issues that have been politicized in U.S. national discourse, especially by conservative movements in recent years (Cordeiro, 2025). This indicates that, despite the hyper-local appearance, Metric Media’s content is guided by a centralized issue agenda (Cordeiro, 2025).

• Demographics & Population Change: One cluster of articles focused on demographic statistics, population change, migration patterns, age distributions, and related census data. These pieces often look innocuous. However, the prevalence of such content points to a narrative interest: discussions of demographic change can underlie political themes (e.g., immigration, urban-rural shifts) (Cordeiro, 2025). While on the surface, these are merely data points, the structural decision to highlight demographics is telling. It resonates with a known tactic in partisan media: framing demographic trends in ways that energize base concerns (Bakir & McStay, 2017). For policymakers, understanding this emphasis is vital, as it demonstrates that propaganda can operate through selective factual reporting, highlighting certain facts.

• Religion & “Traditional Values”: A second cluster revolves around religious and cultural content. These items appealed to religiously conservative readers and legitimized the sites through community-values content (Cordeiro, 2025). The analysis identified repeated references to churches, parishes, and religious education (Cordeiro, 2025). This aligns with external findings that Metric Media had partnerships or content-sharing arrangements with religious advocacy organizations (Bengani, 2021). By embedding moral and religious viewpoints in local reporting, the network effectively advances a conservative cultural agenda within local discourse.

• Civic Governance & Public Safety: Another prominent theme concerns local governance, law enforcement, and public safety. This cluster dovetails with national political messaging that stresses “law and order” and distrust of government. It is notable that these local sites extensively cover small crimes and local government controversies that might otherwise not receive continuous attention (Cordeiro, 2025). This underscores how the network can use ostensibly neutral civic topics as a vehicle for partisan points.

• Economy & Business (Local Angle on National Issues): The economic theme in Metric Media’s content includes local business news, employment statistics, gas prices, and cost-of-living stories. By localizing these issues, the network personalizes broader political debates: a reader is more likely to feel strongly about gas prices when shown that the price at their corner gas station has jumped, rather than reading a national statistic. Thus, Metric Media harnesses local business reporting to reinforce economic grievances and free-market ideological preferences (e.g., suspicion of government economic intervention) (Cordeiro, 2025).

• Education & Civil Society: The fifth cluster concerns education, schools, and other community institutions and has proven to be one of the most politicized content areas. This aligns perfectly with a conservative education reform narrative that favours privatization and portrays public schools as failing or ideologically “captured” (Cordeiro, 2025). The education theme is especially influential because it engages parents and community identity. By inserting partisan perspectives into local school news, the network taps into a deeply personal arena for many readers.

Across these themes, a clear pattern emerges: Metric Media’s ostensibly local journalism consistently foregrounds issues aligned with a conservative political agenda, albeit in a mild, report-like manner. The operation filters the information environment in a manner analogous to Herman and Chomsky’s propaganda model. However, here filtering is achieved through content selection and volume rather than through editorial suppression by corporate media owners (Herman & Chomsky, 1988). By using community-specific events as hooks, Metric Media systematically amplifies certain frames (safety, tradition, economic discontent) while omitting or minimizing others, thereby subtly steering public discourse (Cordeiro, 2025).

3. Distribution Strategies and Engagement Patterns

• Social Media: Metric Media’s online presence extends into social media, primarily via Facebook pages for each of its local “news” brands. The network analysis of these pages revealed a sprawling but shallow footprint (Cordeiro, 2025). There are Facebook pages corresponding to most sites, totalling in the high hundreds. However, most pages have a very small following. Many had only a few dozen to a few hundred followers, and some appeared effectively dormant after an initial burst of posts (Cordeiro, 2025).

  Engagement data confirms that the vast majority of these Facebook posts receive minimal attention. Average engagement per post was slightly higher (a few likes/shares), skewed by a handful of posts that got traction (Cordeiro, 2025). Metric Media’s content did not organically become viral or widely seen on Facebook.

  However, low organic engagement should not be equated with zero influence. There are a few important nuances: first, even modest engagement may involve the right people (e.g., local influencers, partisan group members) who further disseminate the content. Second, Metric Media has used an alternative distribution channel in addition to organic reach.

  Analyzing posting frequency over time revealed that Metric Media’s activity is not constant, but surged during politically significant periods. The study plotted the number of Facebook posts per month and saw clear spikes in October–November 2020 and again in fall 2022, with a minor rise in mid-2021 (perhaps around local elections or specific campaigns) (Cordeiro, 2025). During these spikes, not only did the volume of posts increase, but the content of posts also skewed more political. This aligns with Bengani’s finding that the network was leveraged to influence voters in the run-up to elections (Bengani, 2024).

• Effectiveness and Influence: A crucial question is the extent to which Metric Media’s propaganda network shapes opinions or discourse. While the study cannot measure opinion change directly, the findings allow some inferences. The strategy of high-volume, low-engagement content suggests that the propaganda value lies in presence and perception rather than in persuasion through individual articles. In other words, by existing and being visible (even minimally) in local information spaces, these pseudo-sites create an ambient narrative (Cordeiro, 2025). They may not persuade someone deeply overnight, but they contribute to an environment in which certain ideas seem more prevalent or validated.

  At the same time, the relatively low engagement metrics raise questions about reach. Many of these sites have not built genuine local audiences, which is logical given that the content is often generic and the operations are not truly embedded in communities (Cordeiro, 2025).

In summary, the findings indicate that Metric Media’s network functions as a wide but thin layer of propaganda across the country: wide in its geographic and topical coverage, thin in its immediate audience engagement. It relies on the aggregate effect of many small incursions into local discourse, rather than any single blockbuster fake news story. This makes it a challenging phenomenon to counter, since there is no single viral falsehood to debunk. Instead, one must recognize and address the structural presence of a propaganda apparatus masquerading as local journalism.

Discussion

The Anatomy of Pseudo-Local Propaganda

The analysis reveals a carefully calibrated system that aligns well with theoretical expectations of modern propaganda. Unlike traditional top-down propaganda disseminated via national broadcasters or newspapers, the pseudo-local model is distributed and granular, yet centrally orchestrated. It embodies what might be called “networked propaganda,” leveraging a network of micro-outlets to achieve macro effects.

One theoretical lens to consider is agenda-setting. Classic agenda-setting theory (McCombs & Shaw, 1972) posits that media do not tell people what to think, but significantly influence what people think about. Metric Media’s network demonstrates agenda-setting at the hyper-local level, but is guided by national partisan priorities (Cordeiro, 2025). By saturating local channels with particular issues (crime, schools, demographics), the network aims to make those issues salient to the public across many communities simultaneously. This is, in effect, a grassroots simulation of agenda-setting. This can have downstream political effects, for instance, by influencing which issues local candidates campaign on or which questions dominate town hall meetings.

From the perspective of propaganda studies, this strategy aligns with the concept of “hidden propaganda,” as described by Bernays (1947) and later scholars: propaganda is most effective when the audience is unaware of the persuader’s hand. Today, a slick website with a hometown name can pop up with no easy way for an average reader to know if it is produced by local reporters or by a remote political operative. This opacity is a core part of the pseudo-local playbook, enabling what the researcher earlier termed “message laundering” (Cordeiro, 2025). Metric Media’s network thus manages to fly under the radar of both readers and platform moderation systems by largely obeying the outward forms of legitimate news content.

Micro-targeting in this context is as much about where and when as it is about whom. Yes, the network targets particular demographics via social media, but equally important is targeting specific places and moments. By proliferating across many localities, Metric Media ensures that national narratives are reproduced in local contexts, potentially influencing local political discourse without appearing imposed. By timing content surges to coincide with election cycles, they maximize relevance when citizens are making decisions. This temporal targeting is akin to an election campaign strategy, except it masquerades as journalism rather than campaign ads.

Implications for Democracy and Public Discourse

The existence and operations of networks like Metric Media carry several implications for a democratic society:

• Erosion of Informed Citizenship: Democracy relies on voters having access to accurate and unbiased information about their communities. Pseudo-local sites muddy this information environment. Citizens in areas affected by these sites may develop a skewed perception of local reality. Over time, this can influence voting behaviour and civic engagement, fragmenting communities and stoking divisions by manipulating the local discourse.

• Undermining of Legitimate Local Media: Another casualty of pseudo-local networks is the remaining legitimate local press. Real local newspapers and radio are already struggling; now they face competition from an entity that pretends to do their job, often flooding social media and search results that the real outlets also rely on for traffic. In the long run, pseudo-local networks are parasitic: they mimic and exploit the credibility of local journalism, but in doing so, they can also undermine that credibility for everyone.

• Challenges to Accountability and Transparency: The findings also highlight that existing accountability mechanisms are ill-suited to this phenomenon. Democratic accountability is short-circuited, as those seeking to influence public opinion need not stand up and openly state their case; they can hide behind the mask of “news.” This avoidance of accountability not only deceives the public but also impedes legitimate debate, as one side of an issue employs covert tactics rather than engaging openly.

• Normalized Misinformation Tactics: There is a risk that the pseudo-local model, if left unaddressed, will become normalized as just another PR or political tactic. In such a scenario, the very notion of local journalism as an objective source could collapse, leading communities to retreat into partisan echo chambers for both local and national information. This trend would accelerate the fragmentation of the public sphere, as many scholars, including Sunstein (2018) and Vaidhyanathan (2018), warn.

The fact that engagement was modest and that these sites rely on illusion more than deep engagement suggests they are vulnerable to exposure. When people learn that a site is not what it claims, the spell can be broken: readers might abandon it, and platforms might adjust algorithms. The challenge is to perform that exposure at scale and promptly.

The Role of Policy and Regulation

These findings reinforce the argument that purely market-driven or laissez-faire approaches are insufficient to deal with coordinated disinformation campaigns. There is a clear public interest in ensuring media transparency, particularly with respect to political influence. One reason pseudo-local sites find fertile ground is the dearth of competing local coverage. Rebuilding local news, through subsidies, public funding, or innovative business models, is a critical part of inoculating communities against these fake outlets. Pickard (2020) advocates treating local journalism as infrastructure that requires public investment. The cost of not doing so can be measured in susceptibility to manipulation, as this study illustrates.

Finally, our research underscores a meta-implication: the very tools of big data and AI that enable micro-propaganda can be repurposed for the defence of the truth. Just as Metric Media used data-driven targeting, this research used data-driven analysis to unmask them. This kind of computational watchdog journalism and research will become increasingly important. In a sense, we must fight fire with fire: algorithmic deception should be met with algorithmic detection and transparency.

Policy Implications

Confronting the challenges posed by micro-targeted pseudo-local news networks will require a multifaceted policy approach. The goal is to restore transparency, accountability, and trust in the local information sphere without unduly infringing on press freedom or open discourse. Based on the findings and the broader literature on media policy and disinformation, the researcher outlines several key policy implications and recommendations:

• Strengthen Transparency in Online News Media: One immediate step is to require greater transparency about the ownership, funding, and editorial control of news outlets, especially those operating primarily online. In the European Union, the new Digital Services Act (DSA) 2022 moves toward mandating transparency for online platforms and could inspire similar provisions on content provenance. While the DSA focuses on platform responsibilities, complementary regulations could address content providers directly.

• Expand Election and Advertising Regulations: Currently, a loophole allows political propaganda to slip through when packaged as “news.” Electoral oversight bodies (such as the Federal Election Commission in the U.S. or Elections Canada) should consider rules that treat coordinated pseudo-news campaigns as in-kind political contributions or as election advertising.

• Bolster Local Journalism as a Public Good: A long-term and fundamental solution is to rebuild the capacity of genuine local news, thereby depriving pseudo-local impostors of oxygen. Pickard (2020) argues for bold interventions to save journalism in the age of misinformation, and the rise of networks like Metric Media makes that argument even more urgent and concrete.

• Platform Accountability and Algorithmic Audits: Social media platforms play a facilitating role in the spread of pseudo-local news; therefore, policy should also address their responsibilities. Building on frameworks like Sun’s (2023) notion of regulating algorithmic disinformation, regulators could require platforms to identify and down-rank coordinated disinformation outlets. While heavy-handed moderation (such as banning these sites) can raise freedom-of-expression concerns, algorithmic transparency and tweaks that reduce the reach of deceptive content are a softer but effective tool.

• Media Literacy and Public Awareness Campaigns: Empowering citizens to discern pseudo-local propaganda is crucial and relatively uncontroversial from a policy standpoint. Governments, educational institutions, and civil society can collaborate on media literacy initiatives focused on local news. Essentially, sunlight is the best disinfectant: once exposed, these networks lose credibility. Therefore, an informed public is the first line of defence.

• Collaborative Monitoring and Information Sharing: Just as cybersecurity threats are addressed through information sharing between government and industry, a similar collaborative infrastructure could be developed to monitor disinformation networks. Communication regulators (such as the CRTC in Canada or the FCC in the U.S.), in partnership with academic researchers and journalism organizations, could establish an early-warning system for the proliferation of pseudo-local sites.

• Legal Accountability for Deceptive Campaign Practices: In some jurisdictions, laws prohibit impersonating news outlets or disseminating false information to influence voting behaviour (for example, some states prohibit deliberately spreading false information about voting procedures). Such laws could potentially be invoked or modified to cover the activities of pseudo-local networks. The challenge lies in enforcement, especially when network operators shield themselves behind layers of corporate entities. But making the law clear that this behaviour is unacceptable sets a norm and can empower watchdogs to pursue investigations. It also signals to would-be imitators that there could be legal repercussions.

• Balancing Acts: It is important to acknowledge that any policy response must balance efforts to combat disinformation with protections for freedom of expression and a free press. Policies should target deceptive behaviours and structural aspects (such as nondisclosure, coordination, and automation) rather than censor specific viewpoints. The intent is not to silence conservative (or any) perspectives, but to ensure that, if those perspectives are promoted, it is done transparently and through genuine community voices rather than fabricated personas. By requiring clear labels on the masquerade and by uplifting authentic local voices, policy can tilt the playing field back toward honest communication.

Conclusion

Local news has long been considered the lifeblood of democracy at the community level. This study uncovers a concerted effort to exploit that lifeblood by injecting it with a partisan toxin: stealth propaganda disguised as neighbourhood journalism. The Metric Media case study demonstrates how digital technologies and strategic coordination can resurrect the form of local news while subverting its function. Under innocuous banners and bland news items lurks a sophisticated influence operation, one that challenges our assumptions about where political information comes from and how it spreads.

By applying computational analysis to Metric Media’s sprawling network, the research has shed light on its operations. It found a network that behaves less like a collection of independent community outlets and more like a single propaganda machine with thousands of appendages (Bengani, 2024). Its content strategy (overwhelming neutrality punctuated by targeted bias) maximizes credibility and minimizes detection. Its thematic focus aligns strikingly with a national political agenda, proving that the local appearance is a deliberate illusion. Moreover, its use of social media and micro-targeting shows an adaptation of propaganda to the age of Big Data: the personalization of persuasion on a mass scale.

These findings carry an overarching lesson: defending democracy in the digital age requires updating our notion of “the press” and our mechanisms for safeguarding it. This research is not just about one company’s tactics: it is emblematic of an emerging threat vector in the information landscape. As such, it compels responses from multiple stakeholders: regulators, tech platforms, journalists, educators, and citizens.

It is also worth highlighting a hopeful aspect: the very act of conducting this study and similar investigations is part of the solution. Each time researchers and journalists expose “pink slime” in the media, they diminish its power. In a sense, this data-driven methodology exemplifies a new form of accountability journalism, suited to an era in which adversaries wear digital disguises. Moving forward, collaborations between data scientists, journalists, and policymakers can amplify this impact, turning the tide against disinformation by using evidence and analysis as our tools.

In conclusion, Micro-Targeted Pseudo-Local News networks exploit cracks in our modern communications framework, but they are not invincible. They rely on secrecy, scale, and public naivety to succeed. By learning how they operate and by implementing smart policies to close those gaps, we can curtail their influence. More broadly, this case underscores the importance of renewing our commitment to local media as a pillar of democracy. Ensuring that citizens have access to trustworthy local news, whether by revitalizing existing outlets or by fostering new models, is not merely a cultural or economic issue but a democratic imperative. The health of our local democracies will ultimately depend on our ability to distinguish genuine community voices from ventriloquized ones, and to promote the former over the latter.

Metric Media’s digital propaganda network, once unveiled, serves as both a cautionary tale and a catalyst for action. It challenges us to adapt our policies and civic strategies to the realities of the digital information age. In doing so, we reaffirm the values of transparency, truth, and trust, values that must underpin the media systems of any healthy democracy. The findings and recommendations presented here aim to contribute to that urgent task, ensuring that the promise of the digital age is not undermined by those who would abuse its possibilities in the dark.

Works Cited

Abernathy, P. M. (2022). The expanding news desert: America’s continuing local news crisis—University of North Carolina Press.

Bakir, V., & McStay, A. (2017). Fake News and The Economy of Emotions: Problems, causes, solutions. Digital Journalism, 6, 1–22.

Bartholomew, R. E. (2022). Pink slime journalism and the future of local news. Journal of Media Ethics, 37(4), 214–228.

Benkler, Y., Faris, R., & Roberts, H. (2018). Network Propaganda: Manipulation, Disinformation, and Radicalization in American Politics. Oxford University Press.

Bengani, P. (2019). Hundreds of “pink slime” local news outlets are distributing algorithmic stories and conservative talking points. Columbia Journalism Review.

Bengani, P. (2021). As election nears, more “pink slime” news outlets surface. Columbia Journalism Review.

Bengani, P. (2024). Tracking the growth of partisan local news networks. Columbia Journalism Review.

Benjamin, D. (2022). The weaponization of local trust: Pseudo-local news and public perception. Journalism Practice, 16(7), 1332–1348.

Bennett, W. L., & Livingston, S. (2018). The disinformation order: Disruptive communication and the decline of democratic institutions. European Journal of Communication, 33(2), 122–139.

Bernays, E. L. (1947). The Engineering of Consent. The ANNALS of the American Academy of Political and Social Science, 250(1), 113–120.

Cordeiro, A.-J. (2025). Digital deceptions: Unveiling the impact of pseudo-local news on democracy and crafting countermeasures (Metric Media case study) (Master’s thesis, Concordia University).

Darr, J. P. (2024). The collapse of local news and democratic consequences. Political Communication, 41(1), 1–21.

Herman, E. S., & Chomsky, N. (1988). Manufacturing consent: The political economy of the mass media. Pantheon.

Hobbs, R. (2011). Digital and media literacy: Connecting culture and classroom. Corwin.

Jowett, G. S., & O’Donnell, V. (2012). Propaganda & persuasion (5th ed.). SAGE.

MassMove et al. (2024). MassMove dataset on pseudo-local news networks. Open-source investigative archive.

McCombs, M., & Shaw, D. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36(2), 176–187.

Meyer, P., et al. (2019). The information vacuum and local news decline. Journalism & Mass Communication Quarterly, 96(3), 789–807.

Napoli, P. M. (2019). Social media and the public interest. Columbia University Press.

O’Reilly, T. (2007). What is Web 2.0: Design patterns and business models for the next generation of software. Communications & Strategies, 65, 17–37.

Pickard, V. (2020). Democracy without journalism? Confronting the misinformation society. Oxford University Press.

Sun, H. (2023). Regulating Algorithmic Disinformation. The Columbia Journal of Law & the Arts, 46(4), 367–417.

Sunstein, C. R. (2018). Republic: Divided democracy in the age of social media.

Tandoc, E., Lim, Z. W., & Ling, R. (2017). Defining “fake news.” Digital Journalism, 6(2), 137–153.

Townsend, L., & Wallace, C. (2017). Social media research ethics—University of Aberdeen Research Report.

Vaidhyanathan, S. (2018). Antisocial media: How Facebook disconnects us and undermines democracy. Oxford University Press.

Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146–1151.

Wardle, C., & Derakhshan, H. (2017). Information disorder: Toward an interdisciplinary framework for research and policymaking. Council of Europe.

Yin, R. K. (2018). Case study research and applications (6th ed.). SAGE.

Zuboff, S. (2019). The age of surveillance capitalism. PublicAffairs.

Date modified:

2026-06-03