At the IAB HowTo conference in Warsaw, Poland, Karina Klymenko, Head of MGID’s Creative and Compliance Departments, reflected on rising brand safety concerns and shared her insights on ad quality. In this article, we summarize her talk: read on to discover practical ways for platforms to maintain high ad standards and helpful tips for advertisers dealing with advertising challenges.
What is Ad Quality From an Ad Platform's Perspective?
There is perhaps no metric that is more misinterpreted than ad quality, and it arguably sparks a significant number of disputes. Some interpret it solely as glossy creatives, considering true-to-life visuals not attractive enough; others refer to specific product categories and content, deeming promises of results or timelines for achieving them, types of advertisers and so on as less than.
So, to ensure quality control and analyze cause-and-effect relationships, it's important to recognize that ad quality isn't just one thing but rather a multifaceted concept that should be considered from different perspectives.
- Publishers: This includes assessing the quality of traffic, the overall user experience on the publisher's website, adherence to the publisher's editorial policy (the kind of content they post) and the quality of content on the publisher's website.
- Ad experience: Central to ad quality is the user's interaction with the ad itself. Factors such as the promoted product or service, the design of the ad and the underlying concept all contribute to the overall ad experience.
- Customer safety: Ensuring the safety and transparency of the ad experience is paramount. Ads should be clearly labeled as sponsored content to provide users with transparency regarding their origin. Advertisers must offer an accurate representation of the content or products users will encounter upon clicking the ad. Additionally, ads should seamlessly integrate with the style and format of the surrounding page.
- Brand safety: The advertising environment plays a critical role in maintaining brand safety. This involves considering the context in which ads are displayed and assessing ad adjacency to ensure alignment with brand values and objectives.
How to Protect Users From Bad Ads?
Multi-factor moderation: Our initial moderation process incorporates multiple layers of scrutiny to assess the quality and safety of ads:
- Internal automated model: The automated model works with each individual component of the creative, leveraging either historical data and previous decisions or predictive characteristics about the decision-making process. The model is trained daily and supplemented with new rules through which creatives pass. For us, flexible logic is extremely important, encompassing context and details, with a spectrum of solutions much broader than the "black and white." The model not only makes decisions that significantly expedite moderation but also helps group creatives according to specific principles, provides tips for manual processing and prevents misunderstandings and errors.
- Manual moderation: Manual moderation remains a crucial step in the verification process, handling all sensitive categories, contentious cases and ambiguous messages that may sound offensive to certain user groups. Manual moderation applies both to clients with low trust levels aiming to avoid violations and brand clients to ensure that creatives precisely adhere to brand guidelines. The compliance support team operates at every level and handles the technical aspects of moderation, while local teams make decisions on policy formation and compliance for each active campaign within the system.
- Third-party tools: We leverage third-party solutions to enhance our security measures and predict ad component behavior.
Despite the advancements in AI technology, content assessment tools often struggle to detect subtle contextual cues and sentiments, particularly in mediums like video. To address this, platforms should have local teams of human moderators who understand the specific context of ads and their impact on the audience.
Moderators provide an extra layer of evaluation to identify cultural differences, political ideologies and subject matters that AI may overlook, especially at the local level.
Developing an effective compliance policy requires attention to detail and regular updates. If you're partnering with organizations whose ad quality standards haven't been reviewed in the last six months, asking questions can prompt necessary updates to outdated policies.
Here are some key aspects to consider:
- Local laws and regulations: Regularly monitoring changes in local laws is crucial as notifications about these changes are often not received via email. Our local compliance teams provide timely updates to ensure adherence to legal requirements.
- Brand safety: Preventing conflicting or competing brands from appearing together is essential. Continuous monitoring and updates are necessary to maintain brand safety standards.
- Diversity: Embracing diversity in creative policies is imperative. We actively support initiatives to promote diversity and help advertisers avoid stereotypes, especially in industries where they are prevalent.
Clickbait VS Click Factor
Another important component of ad quality is the concept of clickbait, which is indeed a vivid indicator of low-quality advertising. However, this is highly debatable, as clickbait practices are often used to describe any positive product characteristic or concept. Although, it does not typically indicate the likelihood of the product’s result. It is crucial at this point to perceive and convey the difference between promises of guaranteed results and the probability of achieving them.
Every advertisement operates on human emotions and needs, offering solutions to their pain points. For this, there is a term called the click factor. Avoiding cheap promises but addressing customers' problems and needs — servicing their emotions and interests — constitutes quality advertising.
What Difference Does AI Bring to Ad Quality?
Leveraging the power of AI, advertisers can now optimize ad targeting, personalize content and enhance overall campaign performance with unprecedented precision. One of the most significant advantages of AI is its ability to address key challenges in ad quality. Among other benefits, it helps to solve 4 main problems.
- Copyright infringement: There’s no need to worry about the risk of copyright infringement, since AI generation creates unique images every time.
- Diversity: Advertisers can generate original images, allowing for experimentation and building multiple creatives. This keeps campaigns fresh and also maximizes seasonal opportunities.
- Localization: MGID’s image-to-image model can create variations of one image for different markets, each optimized for specific regions, languages or target audiences.
- A/B testing: Simple and fast A/B testing process
Below you will find examples of the Nestlé Brazil campaign, which was entirely launched using AI-generated visuals of babies (and continues to employ similar concepts to this day). Despite the delicate nature of children-related topics, the brand managed to create a safe space with these baby models, thanks to the high-quality output of AI technology.
Conclusion
Maintaining high-quality ads requires teamwork between advertisers, publishers and platforms. When choosing an ad platform, prioritize those that follow industry standards and regularly update their policies. This ensures a positive user experience and builds trust across the advertising community.
