Measurement
Brand safety
Brand safety is the practice of keeping ads away from content that is harmful or universally unacceptable, such as terrorism, child abuse, hate speech or illegal activity, to protect an advertiser's reputation.
Brand safety is about avoiding the worst: a family car ad next to a terrorist video, or a bank ad funding a hate site. Such placements can hurt the brand and unintentionally pay for harmful content.
It is usually managed with a combination of tools: keyword and URL blocklists, content classification using the IAB Tech Lab content taxonomy (a shared list of content categories), verification vendors that scan pages and videos pre-bid and post-bid, and platform policies. Industry frameworks, notably the one developed by the WFA's Global Alliance for Responsible Media (GARM), defined a "floor" of content no ad should appear against, and GARM itself was wound down in 2024.
The common confusion is with brand suitability. Safety is the universal floor; suitability is each brand's own tolerance above it. Over-aggressive safety, such as blocking every article containing the word "shooting", can defund quality news and waste reach, so modern approaches favour context-aware classification over crude keyword lists. Brand safety is also connected to malvertising and made-for-advertising sites, which pose different kinds of reputational risk.
Think of it like this
Brand safety is like a caterer refusing to serve at events that are plainly illegal or hateful, regardless of who is paying.
An example
A consumer electronics brand's verification vendor blocks 0.8% of impressions pre-bid for appearing on pages classified as extremist or illegal, and removes a further 0.2% post-bid after video content analysis.
Related terms
Brand suitability
Brand suitability is choosing which content is appropriate for a specific brand's ads, based on its own values and risk tolerance, above the universal floor set by brand safety.
Ad verification
Ad verification is independent third-party measurement that checks whether ads were served to real people, in viewable positions, in safe and suitable environments, and in the intended geographies.
Contextual targeting
Contextual targeting places ads based on the content being viewed, such as the topic, keywords or sentiment of a page, video or app, rather than on data about the individual viewer.
Malvertising
Malvertising is using online ads to spread malware, scams or forced redirects, by sneaking harmful code or deceptive landing pages into ad creatives served through legitimate ad networks.
Block rate
Block rate is the share of ad impressions or bid requests prevented from being served or bought because they failed brand-safety, suitability, fraud or targeting checks.
Sources: IAB Tech Lab Content Taxonomy, MRC standards and guidelines