Why Brand Safety is a Holistic Strategy, Not a Single Tool

No brand wants to neglect its online reputation. Creating, maintaining, and protecting that reputation in digital advertising includes implementing brand safety controls.

A typical definition of brand safety is “a set of measures that aim to protect the image and reputation of brands from the negative or damaging influence of questionable or inappropriate content when advertising online."

Yet a single brand safety solution cannot cover all the risks a brand needs to mitigate. Or a particular solution may not be a good fit for the brand’s safety strategy (if they have one).

A recent case in point — Integral Ad Science pulled a demo of its Context Control product after reviewers found that controversial publishers like 4chan and Infowars were categorized as “neutral”. Few brands would consider those sites to have neutral points-of-view.

But there’s more to this story than a tool “not working”.

The context behind IAS’ recent product demo

Context Control helps brands determine if a page is brand safe by analyzing the language used. The tool qualifies the sentiment of the content and its “underlying emotional classification”.

Many brands would want to avoid pages and publishers with negative emotional content, like violent and obscene language. That’s a valid brand safety control, and perhaps a critical one for “upbeat”, “sunny” brands. However, sentiment by itself doesn’t tell the whole brand safety story.

For instance, Context Control classified a white supremacist publisher as “netural” because its language usage had no clear positive or negative sentiment. While virtually no brand wants to appear next to overt hate speech, the tool worked as expected.

Natural language processing tools like Context Control seek to remove the limitations of earlier generations of ad technology. The first contextual targeting tools performed a simple keyword match on a page’s content, which could lead to disastrous outcomes. The need for brand safety initiatives became obvious.

In the opposite extreme, keyword blocking tools also fail to distinguish between contexts and can become overzealous. In the classic example, “shot” has different implications for brand safety if the content involves basketball or gun violence.

As products like Context Control introduce contextual awareness into the ad ecosystem, it’s tempting to see them as a panacea for brand safety. The context of content is everything, isn’t it?

Sort of. As the demo proved, tools that qualify emotional sentiment and other specific metrics are just one aspect of a brand safety strategy.

What makes a brand safety strategy?

Brand safety strategies start with the brand manager. They define the brand’s image and draw the lines between acceptable and unacceptable places for the brand to appear.

In doing so, the brand manager develops a clear understanding of brand safety risks for their brand. For example, it’s unlikely BP’s brand manager liked this placement with a news story on climate change protests.

One brand might find the possibility of appearing next to bad economic news a major risk — if a screenshot of that ad placement spread on social media, the fallout could inflict serious damage. Another brand considers appearing next to bad news to be a major opportunity.

Even “common sense” brand safety risks don’t apply to every brand. An adult products brand may consider appearing next to safe-for-work content to be a risk.

Once the risks are identified, the brand manager determines how best to mitigate them. A mitigation can be a human intervention, a programmatic setting, or some combination of both.

Mitigations can take forms like blocklists (for known incompatibilities in content) and allowlists (for the most sensitive brands). It’s worth noting that a mitigation isn’t limited to a particular tool, or a particular setting. For instance, individual sites and a range of site categories can both be blocked.

Tools like Context Control come into play when they provide information about unknown content. A brand may already block a white supremacist publisher via a hate speech blocklist. But what if tone is a brand safety risk for them? A tool could provide the information they need to make a decision.

The bottom line on brand safety

Brand safety risks extend beyond just content— low quality ad placements, poor site or app user experience, and poor ad experience in general are as risky as appearing on a harmful site. A user shouldn’t remember your brand as the one that broke the app they were using! Like publishers, brands shouldn’t neglect testing their ads in mitigating these risks.

The bottom line is that there’s no one-size-fits-all solution to brand safety. While some risks are fairly constant, different brands will have different risks, which require different types and levels of mitigation.

Instead of relying on tools to make decisions for them, a brand manager must identify their most important risks and implement mitigations that cover them.