The 11 Categories and the Four Risk Levels: How the GARM Framework Actually Dictates Creator Ad Revenue

On June 17, 2022, the World Federation of Advertisers published a document that would quietly rewrite the financial realities of the creator economy. It was not a new algorithm update or a sweeping piece of international legislation. It was a standardized PDF titled the 'GARM Brand Safety Floor + Suitability Framework.' Inside its pages, the internet was neatly partitioned into 11 categories of content, each graded across four distinct levels of risk. For a multinational brand, the framework served as a corporate shield, a standardized way to ensure their logos never appeared next to a PR disaster. But for an independent creator uploading a video to YouTube, Twitch, or TikTok, those 11 categories became the invisible ledger that dictates whether their work generates a living wage or earns nothing at all.[1]

The Global Alliance for Responsible Media (GARM) was established as a cross-industry initiative to address a specific, multi-billion-dollar problem: keeping corporate advertising dollars away from harmful online material. As reported by Marketing Edge in October 2020, the alliance brought together marketers, media agencies, and platforms to create a 'safer digital media environment.' The core issue was that platforms and advertisers lacked a shared vocabulary. Without a universal definition of what constituted 'harmful,' brands were routinely pulling their entire ad spends from platforms whenever a controversy erupted. GARM’s solution was to build a standardized taxonomy so that the entire advertising industry could categorize digital risk 'in the same way across the board.' The approach yielded measurable results for advertisers; the World Federation of Advertisers later reported that ads appearing beside harmful content across measured platforms declined from 6.1% in 2020 to just 1.7% in 2023.[1][3]

The resulting framework divides all digital media into 11 specific categories. These range from the universally prohibited—such as child exploitation and terrorism—to more nuanced areas like 'Adult & Explicit Sexual Content,' 'Arms & Ammunition,' and 'Debated Sensitive Social Issues.' 'The challenge of harmful online content cannot be addressed if one is unable to describe it using consistent and understandable language,' the WFA noted when outlining the framework's purpose. Within these 11 categories, GARM established a hard boundary called the 'Brand Safety Floor.' Content that falls below this floor is deemed universally inappropriate for advertising support and is effectively blacklisted across the programmatic supply chain. If a creator's video breaches the floor, the ad spigot is shut off entirely.[1]

Above that floor lies the 'Suitability Framework,' which introduces the concept of varying risk appetites. Here, content is graded into three tiers: Low Risk, Medium Risk, and High Risk. This is where the framework transitions from objective safety to subjective brand preference. As the Internet Policy Review highlighted in a 2022 analysis, the shift from 'brand safety' to 'brand suitability' allows advertisers to dictate platform governance by deciding what contexts align with their specific corporate values. A video game review featuring simulated violence might be perfectly acceptable for an energy drink brand operating at a Medium Risk tolerance, but entirely unsuitable for a children's toy manufacturer requiring a Low Risk environment.[1][2]

For creators, the theoretical nuance of GARM's four-tier system is often compressed by the platforms that enforce it. YouTube, a founding member of GARM, translates these industry standards into its own 'Advertiser-friendly content guidelines.' When a creator uploads a video, they must navigate a self-certification questionnaire that mirrors GARM's 11 categories, disclosing whether their content contains inappropriate language, violence, or sensitive events. The platform's automated systems then assign a monetization state. Instead of four nuanced risk levels, creators typically face a binary reality: a green icon for full monetization, or a yellow icon indicating limited or no ads.[5]

This compression means that 'Medium Risk' and 'High Risk' content often suffer the exact same financial fate. If a creator discusses a 'Debated Sensitive Social Issue'—one of GARM's 11 categories—they frequently trigger the yellow icon, severely restricting their ad revenue even if the content is educational or journalistic. The Internet Policy Review notes that this dynamic has 'major implications for online media freedom,' as the drive for brand suitability encourages platforms to demonetize coverage of sensitive current events, effectively punishing creators who tackle serious or controversial topics.[1][2][5]

The enforcement of these rules is largely outsourced to third-party verification vendors. Companies in the ad-tech space use the GARM framework to build the algorithms that scan, flag, and block content at scale. According to Coinis, these vendors provide the technological infrastructure that allows brands to apply the GARM definitions programmatically across millions of daily ad auctions. Because these automated systems must process massive volumes of video and text in milliseconds, they frequently lack the capacity to understand context. A documentary about the history of firearms is easily conflated with a video promoting the sale of weapons, resulting in immediate demonetization for the creator.[4]

To mitigate these automated errors, platforms have attempted to carve out exceptions for context. YouTube's guidelines explicitly state that artistic content, music videos, and educational material may contain elements like inappropriate language or non-explicit sexual themes and still remain suitable for advertising. However, the burden of proof falls entirely on the creator. If the algorithm misinterprets the context of a video and flags it as a violation of the brand safety floor, the creator must manually request a human review. During the days it takes to process that appeal, the video typically garners the majority of its lifetime views—views that remain unmonetized.[5]

Even though the World Federation of Advertisers officially discontinued the GARM initiative in August 2024 following legal and financial pressures, the framework it built remains deeply embedded in the digital ecosystem. The ultimate effect is a subtle but pervasive homogenization of creator content. When the financial incentives are strictly aligned with the 'Low Risk' tier, creators are economically pressured to produce universally inoffensive material. As the Internet Policy Review observed, while brand suitability tools promise advertisers greater customization, they are 'in fact highly standardised and likely to promote further homogenisation across major platforms.' The 11 categories and four risk levels were designed to protect the reputations of multinational brands, but in practice, they serve as the editorial guidelines for the modern internet.[1][2]