Collective Commercial Action as a Censorship / Deplatforming Lever

Author: Analyst, advertising-coordination & content-moderation-via-infrastructure Date: 2026-06-07 Method: Primary sources prioritized (court filings, House Judiciary report, company statements, Reuters/AP/FT/CBC/TechCrunch). Each mechanism graded for documented coordination (agreement / concerted action) vs. parallel / convergent conduct (independent firms reacting to the same pressure or risk signal). Skeptical posture: a boycott or a defunding outcome is not proof of unlawful collusion; intent and agreement must be shown.

Executive frame

Three distinct "collective commercial action" architectures recur:

1. Advertiser brand-safety coalitions (GARM/WFA, Stop Hate for Profit) — buyers of advertising coordinate or are urged to withhold ad dollars from disfavored platforms/content.
2. Ratings / "disinformation" intermediaries (NewsGuard, Global Disinformation Index) — third parties produce blocklists/scores that ad networks and exclusion lists consume, with partial government funding.
3. Financial-infrastructure chokepoints (Visa/Mastercard/PayPal; the bank-regulator precedent of Operation Choke Point) — payment networks and their rules act as the de-facto on/off switch for whether content can be monetized at all.

The recurring mechanism is the same: deny money flow (ad revenue or payment processing) rather than directly remove speech. This is attractive precisely because it routes around the First Amendment (which constrains the state, not private firms) and around platform liability — but it also creates the antitrust/coercion exposure that the litigation below tests.

1. GARM (Global Alliance for Responsible Media) under the WFA

What it was. GARM was launched in June 2019 by the World Federation of Advertisers (WFA) as a cross-industry initiative to create shared "brand safety" definitions — common taxonomies of harmful content categories and "brand safety floor / suitability framework" standards that advertisers, agencies (the big holding companies / GroupM etc.) and platforms could adopt for where ads should and should not run. Members included large advertisers (Unilever, Mars, CVS Health, Ørsted, etc.) and agencies whose clients account for a very large share of global ad spend.

The House Judiciary trigger. On 2024-07-10 the House Judiciary Committee (Chairman Jim Jordan, R-OH) released an interim staff report, "GARM's Harm: How the World's Biggest Brands Seek to Control Online Speech," alleging GARM's "Steer Team" coordinated to demonetize/boycott disfavored platforms and outlets (it cites internal discussion touching Fox News, The Daily Wire, Breitbart, Spotify/Joe Rogan, and post-acquisition Twitter/X) and that this "likely" violated antitrust law. The Committee held a hearing the same day and sought documents from 40+ companies.

• Report PDF: https://judiciary.house.gov/sites/evo-subsites/republicans-judiciary.house.gov/files/evo-media-document/2024-07-10%20GARMs%20Harm%20-%20How%20the%20Worlds%20Biggest%20Brands%20Seek%20to%20Control%20Online%20Speech.pdf
• Committee release: https://judiciary.house.gov/media/in-the-news/house-judiciary-committee-report-exposes-shadowy-corporate-coordination-silence

The X lawsuit & dissolution.

• 2024-08-06: X Corp filed an antitrust suit (Sherman Act §1) in the Northern District of Texas against the WFA, GARM, and members (Unilever, Mars, CVS Health, Ørsted) alleging an illegal advertiser boycott.
• 2024-08-08: WFA announced it was discontinuing GARM's operations, citing limited resources. WFA CEO Stephan Loerke: "This decision was not made lightly, but GARM is a not-for-profit organization, and its resources are limited." WFA said it believed it had operated within competition rules and would defend itself; GARM as an entity ceased activity.
• https://fortune.com/2024/08/08/garm-ad-group-shutting-down-elon-musk-x-lawsuit/
• https://deadline.com/2024/08/elon-musk-lawsuit-x-advertisers-1236035749/
• 2024 (Aug 6): Rumble filed a parallel suit against WFA, WPP, Diageo and others.
• 2025-02: X amended its complaint, expanding defendants to ~a dozen+ brands (Nestlé, Colgate-Palmolive, Lego, Shell, Tyson Foods, Abbott, Pinterest, etc.).
• 2025-08-15: Judge Jane Boyle dismissed Rumble's complaint for lack of Texas nexus — "None of the allegations that gave rise to this lawsuit occurred in Texas" — without prejudice (Rumble could refile elsewhere/amend). https://www.mediapost.com/publications/article/408193/judge-tosses-rumble-suit-against-wfa-wpp-diageo.html
• 2026-03-26: Judge Jane Boyle (N.D. Tex., Dallas) dismissed X's antitrust suit WITH PREJUDICE for lack of jurisdiction and failure to state an antitrust claim. Boyle: "The very nature of the alleged conspiracy does not state an antitrust claim" — reasoning that firms have no obligation to advertise on any platform and declining to do so is not by itself an antitrust conspiracy.
• https://www.engadget.com/social-media/judge-tosses-out-xs-advertiser-boycott-lawsuit-184832071.html
• https://www.mlex.com/mlex/antitrust/articles/2458188
• https://news.quantosei.com/2026/03/26/musks-x-loses-suit-accusing-advertisers-of-boycotting-platform/

Litigation status 2025–26: X's federal claim is dead with prejudice as of 2026-03-26 (a defeat for the collusion theory in court). Rumble's was dismissed without prejudice (2025-08-15) on venue grounds and could be re-pleaded. GARM itself remains dissolved.

Coordination grade — MIXED / GENUINELY CONTESTED.

• For coordination: GARM by design produced shared standards and convened competitors; the Judiciary report quotes internal communications suggesting concerted discussion of specific outlets/platforms. A standard-setting body of buyers who together control most ad spend is a classic antitrust risk surface, and "shared definitions of what's unsafe" is one step from "agreement on whom to defund."
• Against coordination (collusion): The court found the alleged facts insufficient to state a §1 claim — i.e., the pleaded conduct looked like parallel exercise of each advertiser's independent right not to buy ads, plus a voluntary framework, not an enforced agreement to boycott. Brand-safety avoidance of "unsafe" adjacency is also independently rational for each advertiser. The House report is a partisan, one-sided staff document, not an adjudication, and selectively quotes.
• Net grade: WEAK-TO-MODERATE documented coordination on standards; NOT established as unlawful collusion (court rejected the antitrust theory). The frame "advertiser cartel" is unproven as a legal matter.

2. Collective Shout — payment processors vs. Steam & itch.io (2025)

Actor: Collective Shout, an Australian anti-pornography / anti-"sexploitation" NGO (director Melinda Tankard Reist).

The mechanism (payment-processor-as-censor):

• 2025-07 (~mid/late July): Collective Shout published an open letter to Visa, Mastercard, PayPal asserting Steam (Valve) and itch.io hosted games depicting "rape, incest and child sexual abuse," and ran an email/letter campaign (it claims ~1,000+ emails to processors; its director told CBC the group escalated to processors after ~3,000 emails to Valve went unanswered).
• ~2025-07-16–18: Valve/Steam updated its onboarding rules (the new "Rule 15" barring content that may violate the rules/standards of payment processors, card networks, banks, or ISPs) and removed hundreds of adult games.
• 2025-07-24: itch.io deindexed/delisted ALL adult NSFW content from browse/search "as we work to comply with the requirements of our payment processors," saying determinations from Stripe/PayPal were pending. (itch.io later began conditionally reinstating some titles.)
• https://www.thesixthaxis.com/2025/07/24/itch-io-delists-all-adult-nsfw-content-due-to-pressure-from-payment-processors-and-collective-shout/
• https://aftermath.site/steam-itch-porn-censorship-collective-shout-visa-mastercard-paypal/
• https://www.cbc.ca/radio/day6/steam-itch-takedowns-credit-cards-1.7597563

Processor responses (denials):

• 2025-08-03: Mastercard denied directing any restriction: "Mastercard has not evaluated any game or required restrictions of any activity on game creator sites and platforms, contrary to media reports," while affirming it requires merchants to have controls so cards "cannot be used for unlawful purchases, including illegal adult content."
• 2025-08-03: Valve told a different story — that Mastercard's rules reached it indirectly: "Mastercard communicated with payment processors and their acquiring banks. Payment processors communicated this with Valve." When Valve cited its 2018 policy of distributing legally-distributable games, processors rejected it citing "risk to the Mastercard brand."
• https://techcrunch.com/2025/08/03/mastercard-denies-pressuring-game-platforms-valve-tells-a-different-story/
• https://www.thegamer.com/mastercard-response-to-gamers-adult-games-steam-itch-collective-shout/

Earlier Collective Shout campaigns:

• 2014: pressured Target Australia / Kmart to pull Grand Theft Auto V (succeeded at Target Australia).
• 2025-04: change.org petition against the Steam game "No Mercy" (rape/incest themes); after a complaint from UK tech secretary Peter Kyle, the game was pulled in the UK, then AU/Canada, and on 2025-04-10 the developer pulled it globally.
• https://en.wikipedia.org/wiki/Collective_Shout
• https://www.collectiveshout.org/steam_campaign_history

Outcome: A small NGO, by appealing to card-network risk rules rather than to the platforms or any government, produced one of the largest single content purges in PC-gaming history (thousands of titles delisted/deindexed across two platforms within ~2 weeks). itch.io partially reinstated; Steam's Rule 15 persists. No government order was involved — which is exactly the point: the lever bypasses both the platform's stated tolerance and any need for a censorship law.

Coordination grade — DOCUMENTED CAUSAL CHAIN; "coordination" is asymmetric, not a cartel.

• This is not competitor collusion. It is a one-NGO-to-infrastructure pressure mechanism plus processor risk-rule enforcement transmitted through intermediaries. Valve's account (processor→Valve transmission of card-brand rules) is first-party and specific; the card networks' denials are narrow ("we didn't evaluate specific games") and compatible with Valve's account (rules enforced via acquirers, not by Mastercard reviewing each title).
• Net grade: STRONG documented mechanism that infrastructure leverage caused the takedowns; the card networks' role is real but mediated and partly deniable (plausible-deniability-by-intermediary is itself a feature of the mechanism). It is not evidence of an unlawful agreement among competitors.

3. Ratings / "disinformation" intermediaries & the ad-defunding stack

NewsGuard & Global Disinformation Index (GDI) + State Dept / Global Engagement Center (GEC)

Mechanism: Third parties score/rate news sources for "trust" or "disinformation risk." Ad-tech exclusion lists and brand-safety vendors can ingest those ratings, steering ad dollars away from low-rated (disproportionately, critics allege, conservative) outlets — a defunding lever one layer removed from any platform decision.

Government-funding controversy:

• 2023-02: Washington Examiner series alleged GDI's "Dynamic Exclusion List" blacklisted conservative outlets and that GDI/NewsGuard received U.S. government support.
• The State Department's Global Engagement Center (GEC) funded GDI (~$100k) and NewsGuard, in part via intermediary Park Capital Investment Group — later acknowledged in litigation. State denied funding domestic "blacklisting."
• 2023: National Endowment for Democracy said it would stop funding GDI; Congress moved to bar Pentagon GDI funding.
• The Daily Wire & The Federalist (with Texas AG) sued the State Department alleging First Amendment violations (government-funded viewpoint suppression / "censorship by proxy").
• 2024-12 (GEC sunset): the Global Engagement Center's authorization lapsed / it was shut down amid the controversy.
• 2025-12: State Dept (Sec. Rubio) imposed a visa sanction on GDI co-founder Clare Melford, citing "organized efforts to coerce American platforms to censor, demonetize, and suppress American viewpoints."
• https://en.wikipedia.org/wiki/Global_Disinformation_Index
• https://www.cato.org/blog/another-lawsuit-alleges-more-government-censorship-proxy-state-departments-funds-blacklisting
• https://www.washingtonexaminer.com/news/2640919/state-department-faces-lawsuit-from-conservative-media-over-funding-censorship-scheme/

Coordination grade — this is the strongest "government-can't-mandate-so-fund-a-proxy" case, but causation to actual ad loss is under-proven. Government funding of rating bodies is documented; what is weaker is rigorous proof that those specific ratings caused measurable ad-revenue loss to named outlets (the chain rating→exclusion-list→advertiser→lost-dollars is plausible and alleged but rarely quantified in the record). Grade: government-proxy structure DOCUMENTED; downstream defunding effect PARTIALLY DOCUMENTED / contested.

Stop Hate for Profit (2020 Facebook ad boycott)

• 2020-06/07: Coalition (ADL, NAACP, Color of Change, Common Sense Media, Free Press, Sleeping Giants) organized 1,000+ companies to pause Facebook/Instagram ad spend for July 2020 with a list of policy demands.
• Outcome: Largely symbolic financially — top-100 advertisers are ~6% of Meta revenue; aggregate July ad spend reportedly rose. Meta made modest concessions (civil-rights hire, audit, expanded hate-speech/Holocaust-denial policy) but met none of the demands at "green" level.
• https://en.wikipedia.org/wiki/2020_Facebook_ad_boycotts
• https://www.npr.org/2020/07/01/885853634/big-brands-abandon-facebook-threatening-to-derail-a-70b-advertising-juggernaut

Coordination grade — OVERT, ORGANIZED, PUBLIC coordination (an admitted boycott) — but it was a transparent advocacy campaign, not covert cartel conduct, and it had limited effect. Legally a coordinated boycott; commercially weak. Grade: HIGH documented coordination, LOW outcome.

4. Financial-infrastructure deplatforming precedents

Sources: Operation Choke Point — https://en.wikipedia.org/wiki/Operation_Choke_Point , https://www.consumerfinancialserviceslawmonitor.com/2017/08/operation-choke-point-terminated/ ; Pornhub — https://www.washingtonpost.com/business/mastercard-visa-investigate-pornhub-business-relationship/2020/12/06/43a12b6e-382f-11eb-aad9-8959227280c4_story.html , https://fortune.com/2020/12/10/visa-mastercard-pornhub-charges-blocked/ ; OnlyFans — https://time.com/6092947/onlyfans-sexual-content-ban/ , https://www.pymnts.com/news/social-commerce/2021/onlyfans-changes-course-wont-ban-explicit-content/

Coordination grade for the financial layer — the mechanism is REAL and REPEATEDLY EFFECTIVE, but it is mostly "rule enforcement / risk management," not competitor collusion. Visa and Mastercard moving on Pornhub within days of the same NYT column is parallel response to a common reputational/legal trigger, not proven agreement with each other. Operation Choke Point is the one case with a state actor behind it — the template the other cases are accused of mimicking without a government. Grade: mechanism HIGHLY DOCUMENTED; inter-firm collusion NOT established; state coercion established ONLY for Operation Choke Point.

5. Testing the frame

Hypothesis: Collective commercial action is used (a) to impose content controls governments cannot directly mandate, and (b) to shift/avoid legal liability.

(a) Imposing controls the state can't mandate — SUPPORTED as a structural description, with caveats.

• The First Amendment binds government, not Visa, an NGO, or an advertiser. Routing content control through private money flow genuinely accomplishes outcomes a U.S. censorship statute could not. The Collective Shout→processor→Steam/itch.io episode is the cleanest proof: thousands of legal titles removed with no law and no court order.
• Caveat / skepticism: "the state couldn't mandate it" cuts both ways — it also means much of this is lawful private choice (advertisers and merchants picking whom to do business with), which is itself a protected activity. The GEC/GDI strand is the only one where government money arguably crosses into "censorship by proxy," and even there the causal link to concrete defunding is under-quantified, and the litigation is unresolved.

(b) Liability-shifting — STRONGLY SUPPORTED as the operative incentive.

• Every node disclaims authorship: Mastercard says it "did not evaluate any game"; Valve says processors made it; OnlyFans blamed "banks"; advertisers say they merely exercised independent brand-safety choices; the rating bodies say they only "inform." Diffusion of responsibility across a chain of nominally-independent private actors is the load-bearing feature — each link is individually defensible and individually deniable, while the aggregate effect is content removal. This is liability-management by design.

OVERALL GRADE

"Collective action as a censorship / liability-management lever": the LEVER is real and well-documented; the "COLLUSION/CARTEL" characterization is largely UNPROVEN and, where tested in court, REJECTED.

• Mechanism exists and works: YES — repeatedly (Pornhub 2020, OnlyFans 2021, Steam/itch.io 2025, Choke Point 2013–17). Money-flow denial is a more durable content-control tool than takedown requests.
• Liability-shifting via deniable intermediaries: YES — consistently observed; this is the strongest, best-supported part of the thesis.
• Unlawful coordination among competitors (the "ad cartel" / collusion claim): NOT ESTABLISHED. The marquee legal test — X Corp v. WFA/GARM — was dismissed with prejudice (2026-03-26) for failure to state an antitrust claim; Rumble's was dismissed on venue (2025-08-25... 08-15). Most "coordination" is better explained as parallel/convergent conduct: independent firms responding to the same reputational trigger, the same card-brand rule, or the same advocacy pressure. The House Judiciary report alleges agreement but is a partisan staff document, not an adjudication.
• Government-as-hidden-hand: PROVEN only for Operation Choke Point; alleged and partially funded-but-unadjudicated for GEC/GDI/NewsGuard; absent in Collective Shout and the GARM advertiser cases (those are private).

Bottom line: Treat "collective commercial action = coordinated censorship cartel" as an unproven (and in the GARM case court-rejected) strong claim. Treat "private financial/advertising infrastructure functions as an effective, deniability-preserving content-control and liability-management lever, sometimes pushed by single advocacy groups and occasionally by government funding" as well-documented. The honest distinction is between a powerful structural lever (real) and a proven conspiracy (mostly not).

← Research index · structured data: influence-ad-censorship.json · influence-ad-censorship.md