In programmatic advertising, brands relinquish direct control over where their ads appear. Automated bidding systems can place ads on millions of websites and apps within milliseconds, creating enormous reach but also significant risk. Brand safety and ad verification technologies exist to manage this risk, ensuring that advertising budgets are spent on legitimate impressions in appropriate environments.
Understanding Brand Safety
Brand safety refers to the practices and technologies that protect a brand's reputation by preventing its advertisements from appearing alongside content that could be harmful, offensive, or misaligned with the brand's values. A luxury fashion brand appearing next to extremist content, or a children's toy company advertising on a page with adult material, represents clear brand safety failures.
The Global Alliance for Responsible Media (GARM) established a standardized framework for categorizing content risks. The GARM Brand Safety Floor and Suitability Framework defines categories of content that are universally considered unsafe, as well as nuanced suitability levels that brands can customize based on their individual risk tolerance.
The GARM Content Categories
GARM identifies several high-risk content categories that most brands want to avoid:
- Adult and explicit content: Pornography, graphic sexual content, and nudity outside of artistic or educational contexts.
- Arms and ammunition: Content promoting weapons sales or glorifying violence involving weapons.
- Crime and harmful acts: Content depicting or promoting criminal activity, including terrorism, human trafficking, and illicit drug use.
- Death, injury, and military conflict: Graphic depictions of death, injury, or active military conflict zones.
- Misinformation: Demonstrably false information presented as factual, particularly around public health, elections, and science.
- Hate speech and discrimination: Content targeting individuals or groups based on protected characteristics.
- Spam and malware: Deceptive sites designed to distribute malware or harvest personal information.
How Ad Verification Works
Ad verification is a broader discipline that encompasses brand safety along with two other critical measurement areas: viewability and fraud detection. Verification vendors deploy technology at multiple points in the ad delivery chain to monitor and measure these dimensions.
Pre-Bid Verification
Pre-bid verification evaluates the quality and safety of an ad placement before a bid is submitted. When a bid request arrives at a DSP, pre-bid segments or real-time classification data from verification vendors can flag whether the page content is safe, whether the ad placement is viewable, and whether there are signals of fraudulent activity.
Pre-bid is the most cost-effective form of verification because it prevents wasted spend entirely. If a page is classified as unsafe, the DSP simply does not bid, and no money is spent. Major DSPs integrate pre-bid data from vendors like Integral Ad Science (IAS), DoubleVerify (DV), and Oracle Advertising (formerly MOAT).
Post-Bid Verification
Post-bid verification measures what actually happened after an ad was served. Verification tags, typically implemented as JavaScript pixels alongside the ad creative, collect data about the environment where the ad appeared, whether it was viewable, and whether it was seen by a human.
Post-bid verification provides more detailed and accurate measurement than pre-bid, since it analyzes the actual page content at the time of ad delivery rather than relying on pre-cached classifications. However, by the time post-bid verification flags an issue, the impression has already been purchased.
Viewability Measurement
Viewability measures whether an ad had the opportunity to be seen by a user. The Media Rating Council (MRC) established standard viewability definitions that have been widely adopted:
- Display ads: At least 50 percent of the ad's pixels must be in the viewable area of the browser window for at least one continuous second.
- Video ads: At least 50 percent of the ad's pixels must be in the viewable area for at least two continuous seconds with the video playing.
- Large display ads (242,500+ pixels): At least 30 percent of pixels must be viewable for one continuous second.
Viewability rates vary significantly by ad format and placement. Industry benchmarks typically show display viewability rates between 50 and 70 percent, meaning a substantial portion of purchased impressions are never actually seen by users. Ads placed below the fold, in background tabs, or in areas users scroll past quickly contribute to low viewability.
Ad Fraud Detection
Ad fraud is the deliberate manipulation of ad serving, measurement, or billing to generate revenue illegitimately. The Association of National Advertisers (ANA) has estimated that ad fraud costs the industry billions of dollars annually. Verification vendors use multiple detection methodologies to identify and filter fraudulent activity.
Common Types of Ad Fraud
- Bot traffic: Automated programs that simulate human browsing behavior to generate fake impressions and clicks. Sophisticated bots can mimic mouse movements, scroll patterns, and even simulate viewability.
- Domain spoofing: Fraudulent sellers misrepresent low-quality or non-existent websites as premium publishers to attract higher bids. The ads.txt standard was specifically designed to combat this by letting publishers declare their authorized sellers.
- Ad stacking: Multiple ads are layered on top of each other in a single ad slot. Only the top ad is visible, but impressions are counted for all of them.
- Pixel stuffing: Ads are served in tiny, invisible pixels on a page. Technically rendered but impossible for users to see or interact with.
- Click farms: Human operators in low-cost labor markets manually click on ads to generate fraudulent click revenue.
Detection Techniques
Verification vendors employ a layered approach to fraud detection. Device and browser analysis examines characteristics like screen resolution, installed plugins, and JavaScript execution patterns to distinguish real browsers from emulated environments. Behavioral analysis looks at interaction patterns, including mouse movements, scroll behavior, and session duration, to identify non-human activity. Network analysis evaluates IP addresses, data center traffic, and proxy usage to flag suspicious traffic sources. Pattern recognition uses machine learning to identify coordinated fraud activity across networks of sites and devices.
Implementing a Brand Safety Strategy
An effective brand safety strategy requires coordination across multiple components:
- Define brand suitability standards: Work with your brand team to define which content categories are absolute exclusions and which represent nuanced suitability decisions based on context and brand positioning.
- Select verification partners: Choose verification vendors whose technology and classification accuracy align with your needs. Run parallel tests to evaluate accuracy and coverage.
- Implement pre-bid and post-bid: Use pre-bid segments to prevent unsafe placements and post-bid tags to measure and verify what was delivered. The two approaches complement each other.
- Maintain inclusion and exclusion lists: Curate lists of approved and blocked domains based on your brand standards and verification data. Regularly update these lists as new sites emerge and content changes.
- Review and adjust regularly: Brand safety is not a set-it-and-forget-it discipline. Regular review of blocked content, verification reports, and emerging threats is essential for maintaining effective protection.
The Cost of Getting It Wrong
Brand safety failures carry real business consequences. Research has consistently shown that consumers who see a brand's ad alongside unsafe content develop negative associations with that brand. The damage extends beyond consumer perception: brand safety incidents can attract media coverage, trigger advertiser boycotts of platforms, and force costly crisis management responses.
Conversely, overly aggressive brand safety settings can significantly restrict campaign reach and performance. Blocking entire categories of news content, for example, can exclude high-quality journalism alongside genuinely unsafe material. The challenge is finding the right balance between protection and reach, which requires ongoing refinement rather than a one-time setup.
As digital advertising becomes more complex and spans more channels, from the open web to CTV to in-game advertising, brand safety and verification will only grow in importance. Advertisers who invest in robust verification programs protect not just their brand reputation but also the effectiveness of their entire digital advertising investment.
