Plans & Pricing
Does My Company Subscribe?
Phill HaymanHead of Technical Services, Sales EngineeringIntegral Ad Science
Although it’s a seemingly cyclical newsworthy topic, the issue of brand safety has come to a head in the UK. eMarketer’s Bill Fisher spoke with Phill Hayman, head of technical services, sales engineering at Integral Ad Science, about some of the reasons why this has become such a hot topic of late, and how Integral is trying to minimize some of the pain points.
eMarketer: Why is brand safety such a big story now?
Phill Hayman: In the last six to nine months there’s been a lot of news about news—fake news. There’s a lot of news floating about that may not be permissible for brands to serve against, so brands are on the watch-out.
You also have a lot of big walled gardens—the likes of Google and YouTube—which have their own tools to figure out what is and isn’t appropriate content. They don’t necessarily want third-party ad verification scripts. They’re not allowing companies like Integral to sit on their pages and monitor brand safety.
But the onus is changing. In the programmatic landscape, most publishers are familiar with ad blocking functionality and brands wanting to prevent appearing against unsavory content. But you can’t easily do that in some of the walled gardens. The focus is now on how we can have holistic campaign measurement across all channels. That’s why it’s become an important topic to a lot of brands, because they understand there are gaps and they need methods to fill those gaps.
eMarketer: Do you think these walls are going to come down anytime soon?
Hayman: Whether they come down totally is probably a way away. Ultimately, the brands have the biggest leverage here. And if those [walled garden] platforms want to continue taking money yet they can’t ensure a brand-safe environment, then the status quo as it currently exists will need to change.
There are a few obstacles that would need to be overcome, but whether those businesses do it in-house in a more transparent fashion and go through the same audit process that an Integral or our competitors are going through, or whether they give us a green light to audit, I think those are conversations that Integral can now have with those platforms.
eMarketer: In terms of measuring brand safety, how difficult is that to do, and how easy will it be to sort out the issues?
Hayman: If you ask anyone to analyze a page in real time and then come back within a couple of milliseconds and tell you what the content of that page is based on an entire contextual analysis of that page, that’s difficult to do, if not impossible at scale.
The issue with YouTube and a lot of the platforms that have content uploaded in real time is how you rate it quickly enough. That’s probably why you’re seeing brands saying, “I’m not happy advertising on YouTube for a little while until we know exactly what’s happening.”
Videos present their own difficulties, where there’s not a lot of content on the page, where there’s no metadata, maybe there’s no URL. For example, for YouTube, it doesn’t say what the video is about in the URL, so it becomes harder to understand what the content is and that’s an area that does need development. There are not many businesses out there delivering accurate contextual analysis of videos.
eMarketer: Are you able to get down to a level where you’re able to say, based on the context of a brand that wants to put its particular ad somewhere, that this is a brand-safe environment?
Hayman: The key for any brand is that it has its own degree of tolerance to what is appropriate content. You need to let the brand set those thresholds.
The way that we do it at Integral is every page gets a variety of scores. We have seven brand-based categories, ranging from adult content to violence, and a machine learning model that has all the key words that might be applicable to that particular brand-safety topic.
If it’s “violence” you’d have anything that’s related to violence, and we’d build a model out that would look across that page so all the URLs that we see, we can stack rank them. We’ll build a score based on the severity and frequency of the key words, and we do that in real time.
eMarketer: How does the brand determine what they want or don’t want?
Hayman: Every page has those scores, and then a brand, based on their tolerance for different levels of risk in those different key areas, can set the thresholds when they’re setting up a campaign. They can say, I want to make sure that I’m not targeting anything that is in any way associated to violence, but I’m OK with slightly softer settings on alcohol, for example.
So the key is having a contextual platform that analyzes pages, that takes into account the frequency or the severity of the page and then builds a holistic score for that page, for that particular brand-safety segment, and then passes it onto the marketers to be able to actually decide on what settings are appropriate for them.
Not a PRO subscriber? Find out how to become one.
You've never experienced research like this.
Nearly all Fortune 500 companies rely on us.
Inquire about corporate subscriptions today.