GroupM brand safety boss calls out Facebook and Google over third party verification excuses

Arvind Hickman
By Arvind Hickman | 23 May 2017
GroupM global EVP brand safety, John Montgomery

GroupM's global head of global brand safety John Montgomery doesn't buy Facebook and Google's argument that user privacy will be compromised if they allow third party verification of their platforms.

He has warned the social media companies that they need to take greater responsibility over the content published on their user generated platforms by allowing third party scrutiny or risk further advertiser backlashes like the YouTube boycott this year.

In a wide-ranging interview with AdNews, Montgomery points out that the brand safety contextual crises that engulfed YouTube is partly driven by a proxy war between traditional publishers and the technology giants eating up their client's advertising budgets.

It also explose the scale and challenges social media companies face by trying to shirk their roles as the world's largest publisher platforms.

Another problem clouding digital media, Montgomery explains later in this piece, is a race to the bottom where clients increasingly commoditise digital inventory and view it as a cost to be managed rather than as an investment that adds value.

But to begin with, Montgomery says, it's important to explain the context around the brand safety contextual debate.

A 'stupefying' amount of content

The major reason user generated content platforms are having problems policing unsavoury content is the volume of uncurated content being uploaded on a daily basis, which he describes as at a “stupefying” level.

“If you think about the volume, 350 hours every minute, a billion hours of YouTube video viewed a day, 1.4 billion daily active users on Facebook. There is no way you can curate that so you can't police it,” Montgomery says.

“The only way we can manage that is through automation so we are doing trillions of ad bids to try and find the right impression, the right place at the right time. So those two things are at a higher volume than they have ever been before.”

Another factor that has pushed contextual brand safety up the CMO's agenda is some of the headlines in the mainstream news, which began with News Corp's serious UK paper, The Times.

“Headlines like 'Big Brands Front Terror' - I'm sure that it's made some journalist's career. Not only that, but I think the established or the legacy media are seeing a lot of their revenue moving to Google and to Facebook and they want to make a very strong point that Google and Facebook are unsafe environments," he explains.

“I would argue that before the Times article jihadist websites probably got 50 impressions or maybe 1,000 impressions, now they probably have had 500,000 impressions. So it's not just an advertising thing, there is a social issue around giving those guys more attention than they deserve.

“This kind of conflict between Twitter, Facebook and Google, and the established press would be interesting to watch if it wasn't for the fact that our client's brands are stuck in the middle.”

Montgomery describes some coverage in the mainstream press, citing articles in News Corp's serious newspapers “feels like entrapment”. AdNews also carried out its own investigation when the issue broke to discover how it affected local brands.


Social media responsibility

Montgomery believes that social media platforms need to take responsibility for advertiser's brand safety, particularly for any threats that exist upon their platforms.

GroupM has agreements in place with traditional media publishers whereby it refuses to pay for any client ads that appear against unsafe content. This doesn't provide for any recompense for any damage to a brand's reputation and is managed on a “relationships basis”.

Social media platforms are different in that they warn advertisers that they should “use this at your own risk”.

“We absolutely believe that they should take full responsibility for our client's brand safety on the web. It's just that you have to acknowledge that when you've got 350 hours of content being uploaded every minute, there is just no way that you can personally supervise that,” Montgomery says.

“It absolutely doesn't take away from the fact that they have a responsibility to keep our client's brand safe. They are making money from it so they are monetising that space. That's the key.”

Use at Your Own Risk

Aside from the technical challenge of curating large volumes of content in a timely manner, social media platforms do not want to take ownership of it due to the threat of litigation in the US.

Media publishers are bound by strict rules around what content they can or cannot publish and defamation laws that protect individuals from having their reputations unfairly tarnished.

Social media companies place the burden on individuals, but defamation suits rarely see the light of day because most individuals cannot afford or be bothered to litigate against abuse they deal with on social media.

For brands, it leaves them in a difficult spot. They need to balance the reach and targeting benefits of social media platforms against the threat of having ads placed near user generated content that is offensive or linked to undesirable groups, such as terrorist sympathisers, far right groups, paedophiles and other harmful groups lurking on these platforms.

Media agency protections

Media agencies use several tools and layers of protection to prevent client ads from appearing in an unsafe environment.
This involves a combination of white lists, black lists, and on-platform filters that block ads from appearing against articles about specific topics.

For YouTube inventory, GroupM has signed up with OpenSlate, which rates video content on its brand safety credentials, providing an extra layer of scrutiny. Other media agency groups have similar tools and strict procedures in place.

Agencies attach special brand safety tags on ads before placing them on a publishers platform. These tags help media agencies identify undesirable webpages by scanning URL data, inbound and outbound links and metadata attached to each article.

Importantly, it also reads every word in an article and blocks ads on webpages that raise red flags.

“I would argue this would give you a 99.9% surety that you have some idea of what is on the site,” Montgomery explains.

“The big 'if' connected to this is if we are allowed to place a third party tag on a website. Google and Facebook and Twitter don't allow it although Twitter is coming around.”

Transparency on social media

Transparency vs privacy

Industry leaders have long called on social media companies to allow third-party vendors access to place tags on ads served on their platforms.

They argue that they cannot allow this in order to protect user privacy, although this is not a view shared by media buyers.

“I contest that,” Montgomery says. “I don't think that brand safety tag would break their contract with their users.

“As you well know, [Google's] agreement with any Google user is they can look in your underwear drawer if they want to. So I think that this is something that those guys don't want us to see inside their platforms. I mean, why would you if you don't have to?”

The concern that Google and Facebook doesn't allow third party verification on brand safety has been expressed privately to AdNews by third party verification firms that partner the platform.

They say they cannot block ads from appearing on unsafe sites if they identify a problem because Google and Facebook don't allow them to run brand safety companion tags.

When AdNews asked Montgomery if he believes social media platforms might be concerned that third party verification tags that block unsafe webpages could affect “quite a lot” of their inventory, he replies: “Yes, but that's what we want, right?”

“Every other medium allows third party measurement. Gosh, we started measuring from a third party perspective 70 or 80 years ago when the ABC was formed and we wanted some assurance that the copies [of newspapers] that were printed were actually distributed.

“So it's a fundamental practice in our business to have third party measurement and it's not just for contextual brand safety although that's the topic du jour. It's for viewability and fraud and audience behaviour and a number of other things.”

It's also a concern for third party verification bodies, who tell AdNews privately that there is a large “slice of utter garbage” that they cannot block ads from running against on the Google Display Network, for example.

Traditional vs digital

The bigger picture

The issue is far bigger than just social media platforms. It goes to the heart of the debate around how digital inventory is valued and is traded using programmatic trading desks and how premium inventory can sometimes be bundled with long tail low quality inventory on user-generated blogs, hobby websites and less savoury content.

As one former CMO recently told AdNews, "would a media owner want their Sydney Morning Herald front page getting the same kind of respect as granny’s crocheted website?

“Yet programmatic use them both as same thing because it's a big system that's got no ceiling and plenty of floor,” the experienced marketing executive, who spoke on the condition of anonymity, explained.

In a recent special report on transparency in AdNews, an overriding sentiment expressed by nearly everybody in the report is that you get what you pay for, and Montgomery agrees, but programmatic is just a means to this end.

“What most people would ask me is, “Isn't programmatic part of the problem?” I always say 'no' because programmatic is just an automatic way of buying. It's like blaming road accidents on motorcars, it's the operators,” he says.

“I think programmatic has been used to access low cost inventory in the long tail. The lower we drive cost - in a way, what has caused a lot of this is the drive for so called efficiency and how do we measure efficiency? We measure efficiency with CPMs.”

Short tail

'Cheaper fraudulent non-viewable shit'

When AdNews pressed Montgomery on what he meant by low cost long tail inventory, he bluntly responds: “cheaper fraudulent non-viewable shit.”

“The thrust of what I am trying to say to our clients is we have to try and change the way we measure,” Montgomery says.

“CPM has become a very convenient benchmark and, if you think about the people who are forcing lower costs, they are generally procurement who aren't experts in media. They will latch onto a measurement benchmark which we have been using for years called CPM (cost per mile/ i.e. thousand impressions).

Montgomery continues: “Then they will say, 'Right, efficiency means lower CPMs'. They will often say to agencies and often go into pitch because of a lower cost drive and say, 'You need to deliver these lower CPMs if you are going to win the business'.

“The result of that is when we win the business we go to our publisher partners and say, 'You know that $10 CPM we had last year? I'm sorry, it needs to be a $9 or an $8 CPM this year because it's part of our contract now.

"We have to deliver on those lower CPMs.' But what is a CPM? There's no quality element in a CPM. It's just a raw cost metric.”

Educating clients

Montgomery says that GroupM tackles this by trying to trying to educate clients to value inventory rather than treat it as a cost.

This education process includes trying to explain that lower cost inventory is less viewable that premium and far more susceptible to ad fraud.

“That means using less low quality inventory and rewarding publishers who give us better quality inventory. They get more money from us. It is an incentive for them to do that. So we have moved viewability up in the US over the last two years. We have moved it up dramatically - over 100%,” Montgomery explains.

In this scenario, CPMs may be more expensive, rising by around 20% in cost, but the viewability of inventory goes up much higher. Using the above example, inventory that is 50% viewable at a CPM of $10 is really a cost of $20 because you need twice as much inventory to reach anywhere near 100% viewability. 

In other words, it's a false economy that CMOs think they are getting a bargain when they buy cheaper CPMs.


Pay for quality, not quantity

GroupM would rather clients pay a little more upfront for more viewable, better results. And this often requires an investment in inventory that is offered by premium publishers. That said, GroupM also offers programmatic solutions that cater to more downmarket inventory.

While Montgomery admits social media platforms have valuable role to play in the media and marketing mix, their popularity with advertisers has been skewed and their lack of transparency is an ongoing concern.

“In media you never get fired for buying Facebook or Google because it's the established thing,” he explains.

“There is no denying that those are very, very valuable reach media and they are great media. Both companies are very smart, technological and media platform companies. All we are asking for is the ability to be able to exercise our own judgement with our own third party measurement in there.

“I've spoken to many hundreds of clients in the last ten weeks since I've been on the road and every one of them says, 'Why don't Google allow a third party measurement?'”

The brand safety issue, if nothing else, has shone a brighter light on YouTube and the advertiser pushback has forced Google to tighten controls and lift its game. Everyone, including Google, accept this is a good thing for the industry.

But there are may more areas of the social media and digital supply chain that lack transparency. It's a battle Montgomery and his peers will continue to champion and his advice to clients is simple:

“Whenever you talk to Facebook, Google, Twitter, Snapchat - any of the guys who aren't letting third party measurement in - insist that they do.”

Have something to say on this? Share your views in the comments section below. Or if you have a news story or tip-off, drop us a line at

Sign up to the AdNews newsletter, like us on Facebook or follow us on Twitter for breaking stories and campaigns throughout the day.

comments powered by Disqus