The Big Issue: Brand Safety

Every week this July, weve taken one of the four biggest issues in mobile advertising right now and examining them close up. With ad blocking, viewability and ad fraud already addressed, this final piece in the series turns its attentions to the topic of brand safety.

Ad misplacement brand safetyBrand safety can be used in a number of ways – most notably as an umbrella term covering ad fraud and viewability, the topics discussed in the last two pieces in this series. Most commonly, though, it refers to instances where an ad appears next to content that is inappropriate for the brand being advertised, also known as brand risk or ad misplacement.

In May, Ad Week reported that Nielsen had pulled advertising from Twitter after its sponsored tweets appeared on profile pages for Homemade Porn and Daily Dick Pictures. And just this month, Co-op and Kuoni demanded their pre-roll ads were removed from a YouTube clip of controversial Rihanna video Bitch Better Have My Money, which is packed with nudity and sexualised violence. “This is certainly not something we would ever wish to be associated with,” a Co-op spokesperson told The Daily Mail.

Integral Ad Science reports that, of the display impressions on networks and exchanges it processed in Q4 2014, 13.7 per cent were blocked because they represented a moderate or very high risk to brands. Those arent gigantic numbers – and its worth noting that these are impressions that didnt get past the filtering process – but this doesnt need to happen at huge scale to be a risk.

A single misplaced ad can have a major impact because, unlike every other issue weve examined in this series, this directly affects consumers.

By definition, individuals are unlikely to notice an ad that doesnt meet viewability standards. Theyll probably even fail to notice whether its the wrong ad, due to an ad injector in their browser. But if someone sees an ad next to inappropriate content, and is offended? Theyll definitely remember that. Worse, these stories often get passed around between friends – and they make for great headlines.

“Weve not just the trade press but the nationals carrying stories of big premium brands that have been seen to be running their ads on salacious content, or content which is inappropriate for that particular brand at that particular moment,” says Nick Welch, business development director at AdmantX. “That in itself is a news story, so it carries weight and it gathers momentum, and can be really damaging for a brand.”

For example, The Sun reported in April that ads for major brands including Asda, Marks & Spencer, British Gas and Nissan were appearing “on websites devoted to paedophilia, incest, bestiality and racism”.

So its no surprise that an IAB study in February identified brand safety as one of the top concerns for UK media agencies – but in the same survey, 22 per cent of agencies admitted to having no idea about the topic.

Not just porn and gambling
In one way, though, brand safety is simplest topic weve covered in this series. After all, its just a case of keeping inappropriate publisher and advertiser content apart. Theres no malicious intent on the side of either party, and inappropriate content can be managed with white lists of brand safe sites, and black lists of sites that contain inappropriate content. As Matt Keating, UK manager for BuzzCity, puts it: “If you have all your ducks in a row as a supplier, you should not be allowing the wrong kind of ad to appear on the wrong kind of the site.”

When you look at it this way, brand safety sounds like something we should have already consigned to the history books. But its also one of the most complex issues weve examined, because its so subjective.

“I suspect if you lined up five brand advertisers and asked what brand safety is, you wouldnt get the same answer,” says Phunware VP, advertising strategy, Jon Hook. “And the same is true with agencies.”

The definition of brand unsafe content is a little like pornography: youll know it when you see it. Its not as simple as good and bad sites and, according to AdmantXs Welch, the kind of obvious clashes mentioned so far are only the tip of the iceberg.

“Salacious content is something that I think everyone needs to get some perspective on,” he says. “Its what Id call the first level of brand safety – no porn, no gambling – and that stuff is easy, but its really only a small fraction of the story. Whats really important for a brand for is avoiding content that might be carrying a negative story about that brand or industry, or even a celebrity that endorses the brand.

“Lets say its a Nike campaign which is running in a sports news app which is carrying a story about the Mo Farrah doping scandal. The last thing Nike is going to want to do is run that ad against that content.”

Welch points to a couple of recent real life examples, including Pirelli Tyres. Advertising against a seemingly sensible set of keywords including car and tyre, Pirelli found its ads appearing on stories about fatal car accidents.

“Another example is Greece,” he says. “Greece is in the news for all the wrong reasons at the moment because the economys in freefall, you cant get any money out of the banks – and yet you have travel advertisers running ads against that coverage to travel to Greece. It might be the right time to reach that user, because theyre interested in buying a holiday, but the content is not going to have a positive impact on how that advertising message is received. Theyre not going to be interested in buying a holiday to Greece while theyre reading about the banks are running dry.”

Mobile and social
In these cases, the ads are running on legitimate publishers, and the content itself may be totally fine for another advertiser. What makes it inappropriate is the particular combination of subject matter and brand messaging.

This somewhat undermines the view, common in the industry, that in-app inventory is automatically brand safe because of Apple and Googles restrictions on their app stores. There are considerably fewer salacious apps (to borrow Welchs term) than there are websites, but news apps are just as likely as news sites to post stories that could be damaging to a brand – and for technical reasons, the content can be harder to analyse.

“Anyone telling you brand safety isnt a problem in-app is lying,” says Phunwares Hook. “Ultimately, if this wasnt a problem, then we wouldnt be having this conversation.”

And its not just traditional news content thats a problem, of course. Social apps, in particular Facebook, are among the biggest and most desirable sources of inventory on mobile, but they represent a flood of user-generated content over which the publisher has – or at least exercises – very little control. However, speaking at the 614 Group Brand Safety Event in May, WPP CEO Sir Martin Sorrell claimed that this doesnt absolve social companies of their responsibilities.

“Its the sellers responsibility, not the buyers, to make sure its safe,” Sorrell said. “You hear Google saying we are not responsible for the content that flows through our pipes, and I think thats nonsense. I see Google and Facebook as media owners, not technology companies, and they have to be responsible for the content. The fact that theres a lot of content generated is irrelevant.”

Automatic for the people
This flood of content creates an online environment thats difficult for advertisers to navigate, which is one of the factors has contributed to the growing popularity of programmatic buying. But automating the trading process means theres less visibility of where ads appear, which can lead to breaches of brand safety.

“Huge swathes of impressions are traded with very little information,” says AdmantXs Welch. “The majority of the information that the advertiser might have is related to the user, and not the content on which theyre finding the user.”

“Programmatic is a fantastic tool but, like anything else, it depends on how you use it,” says BuzzCitys Keating. “It depends on the quality of the data youre inputting, and theres factors around the transparency – how many middle men are in this process, what controls do these guys have along the way?

“Or even the person whos selling directly to you – does he have visibility on exactly where the ad is running? The longer the chain, and the less visibility in that chain, then obviously the greater the chance that somethings going to go wrong.”

The first step towards improving brand safety is to make this process more transparent, and to this end, the DSTG and JICWEBS have published their Good Practice Principles for digital trading in the UK.

But for advertisers hoping to tackle brand safety head on, the best solution is working with specialist vendors. AdmantX offers semantic analysis technology that can dig into individual pages to establish whether they include any potentially damaging content. “Were able to read and process written content using the same processes a human beings would, at scale and speed,” says Welch. This natural language processing is then used to target

Theyre not the only company offering this kind of technology. Kontera, a real-time content analysis platform, was acquired by Amobee for $150m (£96m) last June. Netbase and Sprinklr offer similar natural language processing for social media.

Does it actually matter?
But before you start tackling brand safety, its worth mentioning one final thing that makes this topic different from the others weve covered in this series. While viewability and ad fraud are black-and-white issues – whether its down to placement or a non-human viewer, an ad that goes unseen is inarguably worthless – brand safety is in the eye of the beholder.

As mentioned earlier, identifying a case of brand risk is a know it when you see it situation. What might offend – or even stand out – a person is likely to be different from individual to individual. And if youre seeing an M&S ad running against pornographic content, its probably because youve chosen to visit that porn site.

“Some brands do have the mentality that they dont really care where a person downloads their app from, as long as they use it regularly,” says Phunwares Hook. “Its not about the brand, its about finding a consumer where they hang out and where they want to engage with your content. I once worked with a big insurance brand who couldnt get their head around advertising in anything other than a broadsheet because, its just not where my brand should be. Well, thats great for you, but its where your audience are.”

So do brands even need to worry about this issue? It probably depends on the brand in question.

“For some advertisers, their brand is just a tightly guarded critical asset,” says Fiksu chief strategy officer Craig Palli. “If youre a brand that really tailors to a family orientation, for example, you are going to want to be as far removed from anything controversial as is possible. Brands of this nature may seek to never have an advertisement shown in an app that they havent personally approved.”

Other advertisers, Palli says, are seeking “a balance of reach and safety” and are happy as long as their ads arent running “on any sort of nefarious content”. And theres another group, beyond that, which are purely interested in maximising their reach.

“Ive certainly met some marketers that suggest that whats important is that my brand be pervasive everywhere,” he says. “For those folks, its much more about lets get my ad out there in as many places as possible. Ultimately, its up to the brand to decide where the line that theyll draw is – but every brand should have one.”