Industrialized Disinformation: 2020 Global Inventory of Organized Social Media Manipulation

Contributions

• Global scope: Systematically documents cyber troop activity in 81 countries, the most comprehensive inventory to date as of 2021.
• Three-tiered actor taxonomy: Distinguishes government agencies, political parties, private contractors, and civil society groups, revealing the professionalization and privatization of disinformation operations.
• Capacity classification: Develops a three-point scale (high, medium, low capacity) characterizing 81 countries by cyber troop resource commitment, coordination, and permanence (permanent vs. temporary teams).
• Tactical granularity: Catalogs four communication strategies (disinformation, data-driven ads, trolling/harassment, mass-reporting) and four valence patterns (pro-government/pro-party, smear campaigns, suppression through harassment, division/polarization).
• Account type inventory: Identifies prevalence of three account types: automated (57 countries), human-curated (79 countries), hacked/stolen (14 countries).
• Platform removals data: Documents Facebook and Twitter removing 10,893 accounts, 12,588 pages, 603 groups, 1,556 Instagram accounts, and 294,096 Twitter accounts between January 2019 and November 2020; notes platforms remain reactive rather than proactive.
• Expenditure estimates: Quantifies approximately US$10 million spent on political advertisements by cyber troops and US$60 million spent by governments on contracts with private firms for propaganda services (2009–2020).

Method

The report employs a four-stage mixed-methods approach combining qualitative content analysis, secondary literature review, country case studies, and expert consultation:

1. Content analysis: Qualitative coding of news articles reporting cyber troop activity using a standardized spreadsheet of variables (account types, strategies, messaging valence, organizational form, country presence). Search terms included bot, disinformation, fake account, fake news, information warfare, military misinformation, propaganda, psychological operations, psyops, social media, sock puppet, troll. Purposive sampling across major, professional, and amateur news sources; three-point credibility scale ranked articles by publication quality; only 1-2 tier sources retained.
2. Secondary literature review: Desk research examining government reports, think tank papers, academic articles, and civil society research about cyber troop activity; compiled public archive available in Zotero database tagged by country.
3. Country case studies: Short-form case studies synthesizing findings from content analysis and secondary literature, providing context on media environment, organizational structure, tactics, and capacity.
4. Expert consultation: Peer review of case studies by subject matter experts to validate findings, fact-check data, provide citations to open-source material, and ensure accuracy.

Results

Geographic Scope

Evidence of cyber troop activity in 81 countries (increase from 70 in previous 2019 report). Organized by capacity:

• High capacity (16 countries): Permanent teams with large budgets and sophisticated coordination: Australia, China, Egypt, India, Iran, Iraq, Israel, Myanmar, Pakistan, Philippines, Russia, Saudi Arabia, Ukraine, United Arab Emirates, United Kingdom, United States, Venezuela, Vietnam.
• Medium capacity (40 countries): Temporary or seasonal teams, consistent strategies, year-round staff: Armenia, Austria, Azerbaijan, Bahrain, Belarus, Bolivia, Brazil, Cambodia, Cuba, Czech Republic, Eritrea, Ethiopia, Georgia, Guatemala, Hungary, Indonesia, Kenya, Kazakhstan, Kuwait, Lebanon, Libya, Malaysia, Malta, Mexico, Nigeria, North Korea, Poland, Qatar, Rwanda, South Korea, Sri Lanka, Syria, Taiwan, Tajikistan, Thailand, Turkey, Yemen.
• Low capacity (25 countries): Temporary teams, limited budgets, dormant between elections: Angola, Argentina, Bosnia & Herzegovina, Colombia, Costa Rica, Croatia, Ecuador, El Salvador, Germany, Ghana, Greece, Honduras, Italy, Kyrgyzstan, Macedonia, Moldova, Netherlands, Oman, Serbia, South Africa, Spain, Sudan, Sweden, Tunisia, Uzbekistan, Zimbabwe.

Account Types

• 57 countries use automated accounts (bots); often combined with human-curated accounts to amplify reach.
• 79 countries use human-curated accounts; often operated at low cost by youth groups, students, or unemployed individuals.
• 14 countries report hacked, stolen, or impersonation accounts (including fake Instagram profile of Ali Karimi, Iranian political leader).

Organizational Forms and Actors

Government agencies in 62 countries use social media for computational propaganda (communication ministries, military, police forces, state-funded media). Russia, China, and Iran-backed campaigns target both domestic and international audiences; examples include Philippine Police using Facebook to shape military narratives (2020) and Belarusian government using state-funded media infrastructure.

Political parties in 61 countries deploy social media propaganda, especially during elections. Examples: Tunisia's 2019 election used auto-amplified hashtags; US 2016 used fake accounts and micro-targeted ads; Bloomberg 2020 primary hired operatives for amplification; China's Uyghur-adjacent groups allegedly create fake pages amplifying pro-government content.

Private firms operating in 48 countries contract with state actors to run campaigns. ~US$60 million spent 2009–2020; growth accelerating (9 firms in 2017 → 48 in 2020). Examples: Archimedes Group (Israel-based, operated in Latin America & Africa), Eliminalia (Spain, supported Colombian elections), Estratera (Canada, US$1.38M on Ecuador Facebook ads).

Civil society and citizen influencers in 22 countries: cyber troops coordinate with civil society organizations, hacktivists, fringe movements, or individual influencers under implicit or explicit state sanction.

Strategies and Tactics

Four communication strategies:

1. Disinfo/manipulated media (76 countries): Creation of "fake news" websites, doctored memes/images/videos, doctored fact-check sites. Most common form. Deep-fake concerns noted but few documented 2019–2020 examples in political context (though growing).
2. Data-driven advertising (30 countries): Micro-targeted ads using voter data and behavioral profiling; examples include UK Conservative Party 2019 (90% of FB ads in early December promoted false claims).
3. Trolling, doxxing, harassment (59 countries): State-sponsored trolls suppress political dissent and journalists; example: Tajik Ministry of Education assigned teachers to troll opponents; Pakistan's troll networks reported trending hashtags against military.
4. Mass-reporting (7 countries): Coordinated reporting campaigns to remove legitimate content and suppress free speech.

Four messaging and valence strategies:

• Pro-government/pro-party propaganda (90% of study countries): Amplify supportive state/party narratives, counter-messaging, patriotic content.
• Attack opposition and mount smear campaigns (94% of study countries): Disinformation targeting opposition, conspiracy narratives, character assassination.
• Suppress participation through trolling/harassment (73% of study countries): Vocabulary of harassment silences political dissent (e.g., "terrorists," "foreign invaders"); suppress journalists and activists.
• Drive division and polarization (48% of study countries): Troll farms and conspiracy narratives to fragment social cohesion (examples: Nigeria troll farms allegedly linked to Russian IRA; Hong Kong protest disinformation; US election polarization).

Organizational Capacity and Budgets

Team size and permanence vary. High-capacity teams (China, Russia) operate military-style hierarchies: squad (10 people), company (50 people), battalion (100 people), brigade (500 people), potentially operating 11,500+ accounts. Medium-capacity teams employ smaller, year-round staff. Low-capacity teams emerge temporarily for elections or crises.

Expenditures documented:

• Between January 2019 and November 2020, cyber troops spent over US$10 million on Facebook political advertisements alone.
• Governments signed over US$60 million in contracts with private firms since 2009 (likely underestimate; reported contracts only).

Notes

Strengths: This is the most comprehensive global inventory of organized social media manipulation available as of 2021. The four-stage mixed-methods approach (content analysis + literature review + case studies + expert validation) mitigates individual researcher bias and language barriers (BBC monitoring, multiple-language team). The capacity classification (high/medium/low) and actor taxonomy (government/party/private/civil society) provide useful frameworks for future comparative analysis. The geographic coverage (81 countries) and specificity (account type prevalence, expenditure estimates, concrete examples) make it an essential reference.

Limitations: The inventory is built on publicly reported activity, almost certainly undercounting state-sponsored operations (most sophisticated campaigns remain covert). Language bias: the team used English-language news sources supplemented by translations, potentially missing campaigns in non-English media environments. The credibility scoring system (three-tier for news sources) is subjective, and the authors acknowledge secondary sources may perpetuate original reporting errors. The private-firm expenditure estimates rely on disclosed contracts and are explicitly noted as likely underestimates. The report does not quantify the impact of these campaigns on elections or public opinion — it catalogs organizational activity, not downstream outcomes.

On "cyber troops" terminology: The report uses "cyber troops" defined as "government or political party actors tasked with manipulating public opinion online." This is narrower than "disinformation" writ large (which includes organic, platform-driven, and criminalally motivated false information) and focuses on organized, intentional state or political action. This framing is analytically useful for distinguishing deliberate influence operations from grassroots misinformation or platform-algorithmic amplification.

Methodological note on scope: The report explicitly limits focus to state and political-party cyber troops, not private firms in isolation or organic online grassroots behavior. The inclusion of private firms is justified as those contracted to state actors; independent disinformation-for-profit (e.g., clickbait farms, scams) falls outside this boundary, though some private firms listed may operate independently. This boundary choice reflects the report's framing of the problem as industrialized state/party-sponsored disinformation, not all online falsehood.