How to ensure brand safety with RTB inventory

Brand safe. These two words have been the root cause of more lost hours of sleep for me than I care to admit. On the surface, it seems so simple -- find and deliver online ads to the appropriate users on brand-safe content. Once you dig into it though, you quickly realize you've just peeled off the top layer of an onion -- a very large onion.

How to ensure brand safety with RTB inventory

The basics of brand safety

Most ad servers today have out-of-the-box functionality that allows ad operations teams to manage basic brand safety. For example, if advertisers don't want their ads appearing in the message board section of a site, it can be excluded in the targeting so the ads won't show up in the wrong place. On a direct buy with the publisher, this might be good enough for advertisers as long as they trust that the publisher has tagged its site properly.

With ad networks though, the "trust" part starts to get more complicated. Ad networks generally use a similar ad server tagging structure, and they can also block sites, channels, zones, and content categories, for example. So you need to trust (ethically and technically) the network with which you are working, and then the sites those networks work with. It's important to understand if and how a network vets each of its inventory providers before sending them tags, as well as if and how it continues to monitor where the tags are actually serving ads.

However, what about URL-level brand safety? A reputable news site seems like a pretty safe place to advertise, and you probably want your ads served on it. But we've all heard the story about the cruise line company serving its ads on the "contextually relevant" content of a news site in hopes of getting in front of people that are interested in going on cruises -- only to see its ads appear on pages with stories about cruise catastrophes. The news site domain itself was safe enough, but when the catastrophe news story hit, the URL was created dynamically so you couldn't avoid it. Catastrophe for the cruise company running the ad struck as well. To prevent situations like this from occurring again, technology providers have responded with negative contextual targets, enabling advertisers to specify content that should be avoided.

Brand safety in a real-time bidding (RTB) world

Today, campaigns can be run easily and efficiently using RTB inventory. But again, how do you know where your ads are actually served, and how will you prevent them from ending up on unsafe content -- say, subwaydouchery.com? The RTB inventory platforms all have built-in controls that allow you to manage where you're buying at a basic level -- whitelists, blacklists, content categories, sensitive attributes, and others.

On the surface it seems like you'll have no problems at all, but once you peel back a few more layers of the onion, it's more complicated. Due to the added complexities involved in getting an ad served in RTB, the true domain of where the ad gets served isn't always clear. Often the referring URL resolves to an ad server domain like doubleclick.net or openx.org. Your blacklist will catch the legit domains it recognizes, but there are quite a few that slip through.

So how do you deal with brand safety when it comes to RTB buying?

First, companies like Peer39 and Proximic offer brand safety data segments available to use as part of the bid evaluation. Once you have one or more of these offerings layered on top of your buy, you probably feel pretty good about your controls. But sometimes you can't help but wonder where your ads are actually served. If you're like me, you need a little more information and one more line of defense. This is where the ad verification services come in. Providers like comScore (AdXpose) and DoubleVerify offer post-delivery reporting at the URL level. This is great for maintaining blacklists. You can also set up email alerts for brand safety by content, keyword, and site list, for example. You can even go as far as blocking your ad from appearing if there's something about the page you just don't like.

The trust issue has opened the door to a seemingly endless stream of technology providers with business models built around buying brand-safe media. New companies appear every week with solutions to the brand-safe riddle. You should listen, ask questions, do a trial, and try to make sense of the solution's impact on what you're trying to achieve. I say "make sense of the impact" because the complexities of buying and reporting on impressions make it very difficult to understand what happened and how it would have been different without the technology. That means you really have to understand how the service works -- and how or if it works with all of the other services you use.

Lee Byrne is vice president of ad operations at Bizo.

On Twitter? Follow iMedia Connection at @iMediaTweet.

"People in helmets on a construction site" image via Shutterstock.

 

Comments

David Hahn
David Hahn May 25, 2012 at 3:13 PM

Lee, thanks for an interesting article โ€“ we agree that as advertisers and agencies shift their buying strategies and move budget into RTB they need to be very aware of brand safety but also contextual relevance and Ad Viewability (3MS). Blacklists and whitelists in our mind are a V1.0 solution that have been used to date but not the longer term answer. Marketers can't really sleep at night - or safely scale their campaigns unless their brand is pro-actively protected at the page/url level. You make a good point about the issue with domain level technology - particularly in the instance of news sites that would seem consistently safe places to advertise. We see this often at AdSafe. For example, due to our pro-active approach and the fact that we look at each page on a site - we were able to prevent a cruise brand client's ads from appearing alongside any breaking news about the Costa Concordia disaster in real time. We've been providing the same level or proactive solutions for 100's of clients in both traditional network and publisher buys as well as via our RTB pre-bid solutions โ€“ helping them ensure their brands are protected not only from inadvertent news stories but overall brand safety.

As we move beyond brand safety we can now provide similar predictive measures for RTB around viewability and contextual classification. We believe that successful advertising in RTB not only means leveraging cookies but also leveraging contextual (page level) information as well as what the level of viewability a user potentially has with the ad โ€“ to that end our technology allows buyers to customize their bidding decisions based on the probability of a page's safety level against various risk categories, the contextual relevance of that impression as well as the potential of the ad being seen based on the advertisers requirements. For more information please visit us at www.adsafemedia.com