How we can consign most ad fraud to the history books
We need a system where the company who pays for bad actors that proliferate with fraud must take responsibility.
A recent report from the ANA has exposed the challenge of ad fraud in programmatic marketing: the $88bn open programmatic market is rife with waste and that advertisers could save $20bn.
This report came out during Cannes Lions and was a big topic of conversation. In a time when the need for efficiencies in advertising is acute, this is surely something worth pursuing. Furthermore, in a time when sustainability is high on the agenda, removing waste (which contributes to carbon but delivers no value) is surely a problem worth solving. Not to mention the damage it does to programmatic reputation and budgets, with fraud being used to fund various nefarious activities.
That’s why I and a number of industry veterans have joined the UK Stop Ad Fraud Coalition to try to end this practice, for if we can remove this blight, we can leave a much better industry.
So how do we solve it? Is it just about the use of technology? Well, yes and no.
Before we look at solutions, it’s worth understanding my view of why fraud is such a problem for our industry. I want to show that largely it’s because of the perverse incentives the industry has (more on this later) and my favourite economic theory, “the tragedy of the commons.”
The tragedy of the commons is a concept that explains the challenges faced in situations where resources are collectively owned but lack proper regulation or management. Individuals may exploit and deplete these shared resources, such as polluting a park or beauty spot, for their own short-term benefit, often leading to their ultimate degradation or depletion.
For many actors in programmatic advertising space, there is a perverse incentive to allow fraud to proliferate, as well as the rise of ‘made for advertising’ sites, poor content, and a range of other ills that “pollute” the marketplace. For many actors (supply-side platforms, demand-side platforms, publishers, and even advertisers), allowing a little pollution to leak through can make them additional income in the short term at little risk of being caught.
Surely in the real world, these perverse incentives would not result in allowing the fraud that we see? Surely advertisers would not allow it and vote with their feet? This requires a little examination.
The motivation for fraudsters themselves is obvious: greed. It is also well known within advertising circles that many actors on the supply side turned a blind eye to fraud rates because, in a market where they were being heavily squeezed, this worked in their favour. Crucially, the supply side did not believe that, at least in the short run and for their individual company, that efforts to remove fraud would be rewarded by higher prices on the remaining quality inventory.
The most famous (but not the only) example of this was back in 2015 when AppNexus announced that it had removed around 40% of its inventory due to fraud concerns, the implication being that previously this amount was of questionable quality. Since 2015, a new generation of fraudsters has emerged, and as the ANA report shows, volume is rife.
Can’t advertisers vote with their feet? Well, to a degree, they have. The so-called walled gardens take at least 80% market share in digital marketing. The buy-side, too, is subject to some of these perverse incentives, the desire for ever-lower costs-per-thousand (CPMs) putting pressure on a proliferation of lower-quality inventory. The poor state of campaign measurement and incrementality also meant that advertisers were not aware of how much of what they were buying was garbage.
The solution: better vetting and access to data
So what is the solution? Economic theory can help with the established remedy to a problem of this type being transparency and regulation.
If potential polluters are visible and regulation enforced, problems habitats can be saved. For programmatic marketing, regulation will come, but for me, the trick is around transparency and responsibility aided by technology. This is a great example where technology needs to be the supporting act, not the main event because this challenge, as we have discussed, is more about human nature than technical necessity.
But how do we bring transparency to programmatic? Consider this from an advertiser’s perspective: most advertisers would never want to be seen paying bad actors in the market, and that is the instruction given to their media buyers. The challenge is that when you are buying from a programmatic platform, identifying those bad actors is not as easy as it should be.
Advertisers can see a report of which website addresses (URLs) they purchased and can use verification to attempt to remove issues; however, these systems are not always used properly and far from perfect anyway. Advertisers have no ability to see which companies are being paid by the platforms on the back of their media dollars in an open programmatic environment. This is crucial as the fraudsters love to change their sites and reload their platforms to exchanges without ever changing their billing identity — far too easy an option.
What we need is a setup where the company who pays (or decides to pay) for bad actors that proliferate with fraud must take responsibility. This means that either the supply side needs to take much much more responsibility for online fraud or provide the transparency on whom is being paid to the buy-side so these suppliers can go through procurement efforts.
Imagine if only publishers that had been vetted at a billing level as legitimate were paid, and you would clean up the majority of the issues in an instant. New publishers could submit themselves for an automated audit with their incomes placed in escrow until they had been shown to be of sufficient standards. There is more to transparency than this, with better access to data being crucial to clean up the industry.
UKSAFC and bodies like Isba have proposed standards on transparency that need to be more universally adopted. With transparency and responsibility in place, techniques like Ads.txt and sellers.json alongside the verification vendors (with their level of responsibility) can assist the industry in automating good behaviour.
Cleaning up fraud in programmatic is much more about fixing the human part, the perverse incentives than purely pointing to technology. As is so often the case, the human element is what is important, and technology’s role should be that of an enabler and automater (something to remember as AI continues to grow).
If we all engage to solve the human part, we can permanently consign the vast majority of fraud to history. I hope you will join me and all those at UKSAFC in believing this is something worth doing.
Robert Webster is global vice-president — Strategy at marketing transformation consultancy ControlvExposed