What is Brand Safety?
Also known as: Ad placement safety, Brand-suitable advertising
What is brand safety?
Brand safety is the practice of preventing ads from appearing next to content that could harm an advertiser's reputation. It combines blocklists, contextual analysis, and third-party verification to filter inventory in real time. The goal is simple. Keep the logo away from terrorism, hate speech, piracy, and misinformation.
The discipline emerged in the early 2010s when programmatic buying scaled faster than publisher vetting. By 2017, the YouTube advertiser boycott pushed the industry to formalize standards. Today every major DSP, SSP, and verification vendor maps to the same shared framework: the GARM Brand Safety Floor maintained by the World Federation of Advertisers.
Brand safety is distinct from ad fraud. Fraud is about non-human or invalid traffic. Brand safety is about the editorial environment around a real human impression. Both matter. Both get measured. Most advertisers buy them together.
GARM Brand Safety Floor and Suitability Framework
GARM publishes 11 sensitive categories that form the shared floor for the industry. Every category has a high, medium, and low risk tier. Advertisers pick the tier that fits their brand. The framework is voluntary but adopted by every major holdco and platform.
| Category | Examples |
|---|---|
| Adult & Explicit Sexual Content | Pornography, explicit nudity |
| Arms & Ammunition | Firearms sales, illegal weapons content |
| Crime & Harmful acts to individuals and Society | Human trafficking, exploitation |
| Death, Injury or Military Conflict | Graphic violence, war footage |
| Online piracy | Pirated streams, illegal downloads |
| Hate speech & acts of aggression | Targeted hate against protected groups |
| Obscenity and Profanity | Excessive profanity, vulgar language |
| Illegal Drugs, Tobacco, e-cigarettes, Vaping, or Alcohol | Underage targeting, illegal sales |
| Spam or Harmful Content | Malware, phishing, MFA |
| Terrorism | Terrorist propaganda, recruitment |
| Debated Sensitive Social Issues | Highly divisive political content |
The full definitions sit on the WFA GARM site. Verification vendors classify every page against these 11 categories so advertisers can pick exclusions per campaign.
Pre-bid vs post-bid brand safety
Pre-bid checks block unsafe inventory before money moves. Post-bid checks measure what actually ran. Most enterprise advertisers use both, because each catches what the other misses.
Pre-bid sits inside the DSP. Before the platform bids on an impression, it queries the verification vendor's API. The vendor returns a verdict: safe, unsafe, or unknown. The DSP filters in milliseconds. According to IAB Tech Lab, pre-bid integrations now cover the majority of open web programmatic inventory.
Post-bid happens after the impression. A measurement tag fires inside the ad, scans the page, and reports back. The advertiser learns where the ad ran, on what content, and whether it was viewable. Post-bid is slower but independent, which makes it the source of truth for monthly reporting and make-good negotiations.
The trade-off is simple. Pre-bid stops waste. Post-bid proves what happened. A serious program runs both layers from different vendors so the measurement is not marking its own homework.
Major brand safety vendors
Four vendors handle the bulk of enterprise verification. Each has a slightly different focus, and most advertisers run at least two for cross-checking.
| Vendor | Strength | Notable products |
|---|---|---|
| DoubleVerify | Pre-bid avoidance, attention metrics | Authentic Brand Suitability, DV Authentic Attention |
| Integral Ad Science (IAS) | Contextual categorization, CTV measurement | Total Media Quality, Context Control |
| Oracle MOAT | Viewability and invalid traffic measurement | MOAT Analytics, MOAT Pre-Bid (sunset path announced) |
| HUMAN Security | Bot detection, sophisticated invalid traffic | MediaGuard, BotGuard for Advertising |
All four are accredited by the Media Rating Council for viewability and invalid traffic. Oracle announced the wind-down of the Oracle Advertising business unit in 2024, so MOAT customers are migrating to DV or IAS through 2026. Coverage varies by channel. CTV and retail media still trail open web in vendor maturity.
Brand safety vs brand suitability
Brand safety is the floor. Brand suitability is the line each advertiser draws above it.
A car insurance company running on a news site is brand safe next to a story about a fatal crash. It is not brand suitable. The content is legal, factual, and editorially sound, but the adjacency damages the message. Suitability is contextual, not categorical.
GARM splits this into two products. The Brand Safety Floor is the shared baseline, blocked by nearly every advertiser. The Brand Suitability Framework adds three risk tiers, high, medium, and low, that each advertiser configures. A QSR brand might accept medium-risk gaming content. A pharma brand will not.
This is where contextual classification has returned to the centre of the stack. Vendors like IAS and DV now offer page-level NLP that scores tone, sentiment, and topic, not just URL category. The shift matters as third-party cookies disappear and pure audience targeting weakens.
Real-world example: a brand safety incident
In February 2017, The Times of London reported that ads from Mercedes-Benz, Marks & Spencer, the BBC, and the UK government were running on YouTube videos promoting extremism. The story triggered a global advertiser boycott. Within weeks, more than 250 brands paused YouTube spend.
Google's response was structural. The platform tightened monetization eligibility, raised the subscriber threshold for the YouTube Partner Program, and rolled out preferred partner inventory tiers. Annual revenue impact was estimated in the hundreds of millions of dollars in 2017 alone, per Bloomberg coverage.
The lesson held. A single Sunday newspaper investigation rewrote ad operations across an entire platform. Today every major buyer requires GARM-aligned reporting, third-party verification, and inclusion or exclusion lists at campaign launch. The cost of an incident is measured in pulled budgets and CFO scrutiny, not just CPMs.
Brand safety in 2026
Three forces are reshaping the discipline this year.
AI-generated content at scale. NewsGuard's tracker logged growth from 49 to over 1,150 AI-generated unreliable news sites between May 2023 and 2024. Most are MFA, optimized to attract programmatic spend, with no human editorial. Static blocklists cannot keep pace. Real-time contextual classification has become the default defence.
Deepfakes and synthetic media. Audio and video generated by diffusion models now appear in news cycles within hours of real events. Verification vendors are adding provenance signals from the C2PA standard to score whether media is authentic, modified, or fully synthetic. Advertisers who run on news inventory are starting to require provenance metadata as a buying condition.
The contextual return. As cookies wind down and regulators tighten audience targeting, contextual is back. The difference from 2010 contextual is depth. Modern systems read full page semantics, not keyword bags. Campaigns now define suitability as a vector of topics and tones, not a list of URLs to block. Brand safety stops being a defensive add-on and starts being the primary targeting mechanism. Pair this with a tight blacklist and a curated whitelist of trusted publishers, and the supply path tightens without the audience reach collapsing.
For advertisers running programmatic campaigns, the operating model is shifting from "buy audiences, filter inventory" to "buy contexts, verify identity." Brand safety is no longer a checkbox at launch. It is the core of how serious media is bought.
Related terms
Frequently asked questions
What is the difference between brand safety and brand suitability?
Brand safety is universal. It blocks content nearly every advertiser agrees is unsafe, like terrorism or child exploitation. Brand suitability is advertiser-specific. A whisky brand can run next to bar reviews. A children's brand cannot. GARM defines a shared safety floor and a separate suitability framework with high, medium, and low risk tiers.
How does pre-bid brand safety work?
Pre-bid checks happen before the DSP places a bid. The verification vendor scans the URL, page content, and historical signals in real time. If the placement violates the campaign's safety rules, the bid is blocked. This stops the impression from being bought at all, which is cheaper than detecting the problem after spend.
Who are the major brand safety vendors?
DoubleVerify, Integral Ad Science (IAS), and Oracle MOAT dominate verification. HUMAN (formerly White Ops) leads on bot detection and ad fraud. Each is MRC-accredited for viewability and invalid traffic measurement. Most enterprise advertisers run at least one pre-bid layer plus one post-bid measurement vendor for independent reporting.
Is brand safety still a problem with AI-generated content?
Yes, and it is harder than before. Generative AI lets bad actors produce thousands of low-quality MFA pages, deepfake videos, and synthetic news sites at scale. According to NewsGuard's 2024 tracker, AI-generated unreliable news sites grew from 49 to over 1,150 in twelve months. Static blocklists cannot keep up, so contextual classification matters more.
How much does brand unsafe inventory cost advertisers?
The ANA's 2023 Programmatic Media Supply Chain Transparency Study found 23 percent of programmatic spend was wasted on MFA and unsafe inventory, equal to roughly $20 billion across the open web annually. Brand safety controls do not eliminate the loss. They redirect spend toward inventory that performs and reports cleanly.