How GroupM deals with brand safety

GroupM's John Montgomery, EVP, Global Brand Safety, told WARC why advertisers can never totally eliminate the risk of appearing alongside inappropriate content; while it is possible to introduce strong safety measures, he adds, these come at a cost.

1. Is the focus on brand safety a useful correction or an over-reaction?

This isn’t an easy question to answer. Of course, the measures that we are taking are necessary because of the complexity inherent in digital and we must ensure that our clients ads are viewed in a safe environment, but there is a need for us to bring balance into, particularly the contextual brand safety discussion, so that we can weigh risk against reward.

As much as we approach safe contextual placement with a zero-tolerance attitude, we can never completely eliminate the risk of an ad appearing adjacent to inappropriate content. We must remember that the nature of automated behavioral advertising is directed by an algorithm that selects a user (it’s our ambition to place the right ad in front of the right prospect at the right time) not a web-site. So if a user matches the target profile and happens to stray onto an inappropriate site – they may see an ad directed at them. No leading marketer wants their ad to be seen in an inappropriate environment, but I would argue that the reputational harm rests mostly in the publicity, should that brand’s ad be discovered by a reporter and subsequently broadcast to a larger audience, rather than the exposure to the target user who chose to frequent the “inappropriate” site in the first place. Normally a very small percentage of marketers’ ads appear adjacent to risky content; the real danger lies in the embarrassment when this is exposed in the press.

GroupM takes very strong measures to ensure that ads are safely placed, but also makes it clear that the more filters that are applied (like white lists, category avoidance, monetization floors etc), the more we limit reach and increase relative costs of reaching the audience. In this process, we have a risk assessment discussion with our clients to help them decide the implications of their brand safety strategy.

Thus, a call for perspective and balance.

2. What is your definition of brand safety?

Brand Safety is GroupM’s collective description of any risk that a brand may face in the digital supply chain. Currently we have nine areas identified:

Digital opens many opportunities to us through the use of data, but along with those opportunities come some risk and we have to use technology and vigilance to mitigate that risk.

GroupM have led the market in all the categories listed on this schematic (in some cases we developed the technology for the category and shared it with the industry – like privacy and anti-piracy) and have received wide recognition for doing so. Two such areas are:

Privacy: GroupM led the team that developed communications for online privacy self-regulation in the US (the blue icon that appears in the top left of all behaviorally targeted ads, was designed by our team) and we have represented the industry several times in Washington to discourage regulation. Our digital world would have looked very different if we had opt-in cookie regulation, which was being suggested by US congress at the time.

Anti-Piracy: Some of GroupM’s largest clients are content creators and they lose billions to IP infringement. One of the main ways that the illegal download sites make money is through advertising that is placed on their sites – our client’s ads find their way to the sites because our target audiences are often illegally downloading material. GroupM operationalized technology and systems to ensure that our clients ads are kept off these sites and last years shared this tech with the industry which has now become TAG ant-piracy initiative and is implemented across the industry and prevents hundreds of millions of $’s from going to pirate site.

3. What changes has the industry made since the start of the year? Are they going far enough, in your view?

The main areas of concern has been with contextual brand safety in the large social media platforms like YouTube and to a lesser degree Facebook, Twitter and Snapchat. Clients want to avoid being adjacent to inappropriate content and the press has been on high alert to monitor the situation and call out brands for "supporting" such content. Apart from the adjacency issue, brands have been worried about areas of fake news where untruthful and sensationalistic content is distributed for ideological and profit reasons and is supported by ad revenue.

Facebook and Google in particular have responded with urgency, but for some, not enough is being done, quickly enough.

Facebook has employed an additional 3000 people worldwide to investigate fake news and content quality and Google has invested heavily in machine learning and human controls to keep inappropriate channels off their platform. The environment is definitely safer than it was six months ago, but we are far from declaring it safe enough.

4. What has been the biggest impact on investment in digital? Has there been a flight to quality, or a shift of budget out of digital

Definitely the former. We have not noticed clients moving significant amounts of investment out of digital altogether (the channel continues to grow), but publishers of premium content who can assure clients of a high degree of quality control in all the areas of brand safety are favored.

5. Is there a better appreciation among clients of the context an ad is in (versus chasing the cheapest impression regardless of location)?

Yes, particularly amongst branded advertisers. Advertisers measuring on a cost per action (a sale or an online interaction) are less concerned as they pay only once the interaction has occurred.

6. What is the role of media agencies in cleaning this up?

Media agencies are ideally placed to measure digital inventory quality and apply the brand safety tools to protect clients from risks in the digital supply chain. We also help our clients understand the risks that their brands face; particularly if a client tries to drive down pricing - the risks increase.

7. What has been the impact on pricing?

We are encouraging our clients to include a quality dimension into their price measurement - like a cost per human, (i.e. fraud free) viewable impression. When non-viewable, fraudulent impressions are eliminated from the inventory measurement, prices are stable or even lower.

8. Is there a role for blockchain technology in cleaning up digital advertising?

Probably. In theory, blockchain technology should be helpful in identifying and eliminating risky inventory, but common wisdom has it that a useful solution is more than a year away. This is a development that we are watching carefully.