Our CEO and Co-Founder, Jacek Chrusciany, has authored an article titled “Why Aren’t Brands Using the Tools They’ve Said They Need?” This piece continues his “Pathways to Transparent Media” series, featured in the Association of National Advertisers (ANA) Industry Insights.

Link to the original article

A reprint of this article can be found below.

The digital advertising industry and its infrastructure have come a long way since P&G's Marc Pritchard launched his famous and highly effective crusade to bring about a "cleaner and more transparent advertising ecosystem." And yet, many advertising campaigns today are still struggling with quality and transparency issues when it comes to the placement of their media buys.

As time goes on, less of the fault lies with advertising platforms, and more of it is shifting to the advertisers themselves. That's a tough pill for many to swallow, but it's true. Brands and their agencies have more safety mechanisms available to them than ever before. So, why aren't they eagerly embracing these tools?

In recent years, advertisers have put pressure on the top media platforms to improve brand safety, and the platforms have largely delivered: For example, both Meta and Google have built AI capabilities for categorizing content, and they've given advertisers corresponding controls for these capabilities. However, a recent study of advertiser campaign setups found that these controls are underutilized. Despite the threat of brand unsafe content and transparency issues, campaigns today still often run without any sensitive content exclusions, and on the broadest inventory tier that includes risky content. Consider:

Meta Settings

• In the campaign setups evaluated in the recent study, 37 percent of eligible campaigns used the "Expanded" feed content tier. Meta's "Expanded" content includes: "Physical or emotional distress," "Social issues that provoke debate," "Substance abuse or crime," "Mature sexual or suggestive topics," "Profanity, derogatory words, slurs, or vulgar sexual language," and "Injury, gore, or bodily functions/conditions."

• 15 percent of campaigns included Audience Network, out of which 24 percent did not use any Audience Network content filters. Not all Audience Network inventory is risky, but Meta has less control over the Audience Network content compared to its own inventory, so many advertisers require these placements to be turned off.

Google DV360 Settings

• 44 percent of campaigns did not exclude any sensitive categories. Examples of sensitive categories include: Sexual, Violence, Profanity, Drugs, Politics, Religion, Tragedy, Transportation accidents, Shocking, and Sensitive social issues.

• 36 percent of open programmatic campaigns had no inclusion list and no exclusion list. Such campaigns are at a particular risk of falling victim to "made for advertising" (MFA) websites and ad fraud.

• 26 percent of YouTube campaigns did not exclude Google Video Partners. (A recent Adalytics study provided an in-depth look at the challenges of GVP inventory quality.)

Google Search Ads Settings

• 26 percent of accounts don't include any negative keywords. Negative keywords are critical in ensuring that the ads don't show up for search queries a brand doesn't want to be associated with.

• 12 percent of campaigns did not exclude Google Search Partners, despite their advertisers' policies requiring Search Partners inventory to be turned off. Advertisers often choose to have GSP turned off to avoid showing their ads in contexts they have no control over.

It's easy to disparage advertisers for their failure to implement available campaign safety switches, but the problem isn't as simple as it might seem. The fact is that today's advertisers face a scale challenge. Many brands today, within a single initiative, are placing hundreds of thousands of ads across an ever-growing number of platforms. They're working across multiple regions and countries, often coordinating with several partners. Their ability to manually implement and review campaign controls on a platform-by-platform basis is limited, if not impossible.