Your AI does not need tickets to the big game
Only when we address all of the ways that fraud and inefficiency have plagued media buying will we end their threat to a healthy, thriving marketplace.
Over the past few years, marketing has been transformed by new technology. And the flood of data makes it easier and more efficient to identify, reach and convert potential customers and stakeholders. With relative ease, marketers today can find those individuals who might most benefit from their company or organization’s products and services. From the comfort of their desks, and with just a few keystrokes, marketers can open up new markets and reach a global audience.
As with every advance, however, this progress has attracted some bad actors. These individuals have sought to take advantage of our messy rush to expand the advertising ecosystem, creating some issues we may need to avoid.
The problem is complex, ranging from how many ads are actually seen by anybody and the definition of ad measurement to “hidden” rebates, where media agencies don’t disclose or return savings to the advertising clients. In addition, bad business practices of a bygone era may be lingering, artificially distorting our marketing efforts and confounding some of our efforts to optimize every advertising dollar.
The cost of fraud
Previously believed to cost advertisers $7.2 billion globally each year, the actual cost of advertising fraud was predicted to reach $16.4 billion in 2017, according to the Association of National Advertisers. To put this in context, nearly 20 percent of total digital ad spend is wasted on fraudulent advertising placements. This is one of the main reasons many US digital marketers put ad fraud and brand safety as top concerns.
For most of the latter half of the 20th century, the business of advertising was a business of insiders. In this world, media planners were often wined and dined, given premium tickets to popular events or taken to exclusive travel destinations for spurious media training events. Though these indulgences were sometimes wrapped in the facade of education, their ultimate purpose was to make sure a media company received its “fair share” of clients’ media budgets.
Unfortunately, some may believe that these freebies are part of the media buyer’s or agency’s compensation and are justified. However, this undue influence warps a brand’s advertising plan in three key areas: planning, buying and delivery.
Putting an end to outdated and harmful practices
It’s going to take a well-coordinated effort from many players to root out and put an end these practices that inflate impressions and artificially boost performance data. It’s important to understand that we are the ones who left this door open for fraudsters.
We created a tangle of scenarios, involving humans and machines, that gave unique opportunities for dishonest brokers to misrepresent the impact they’re delivering and take advantage of us. Only when we start to address all of these aspects will we start toward ending their threat to a healthy, thriving marketplace.
A place to start is the oft-used but generally misunderstood double-edged sword of transparency. A truly transparent marketplace, one in which there is no place to hide, should by its nature be trustworthy. However, the acknowledgment of a thing doesn’t always lead to the comprehension of the thing. We have to be smarter.
To eliminate fraud, we need to step up our game and ask ourselves some tough questions about:
- How we buy.
- From whom we buy.
- How we target.
- How we engage publishers.
- What metrics we use to buy and judge success.
- Which intermediaries (if any) we should work with.
- Which intermediaries our agencies work with.
We need to make sure we understand the blind spots we’ve created where fraud can occur. Unless we are critical of ourselves and our peers and understand and address all the factors, the door will always be ajar.
What role can artificial intelligence (AI) and machine learning play?
Leading media audit companies are exploring the potential of transparent algorithms to measure ROI and plan media. These algorithms, or artificial intelligence (AI), can generate the initial media plan based on independent data and prior performance measurements. The advertiser, agency or buyer is free to make edits, but the software automatically tracks which changes are made, when, why and by whom.
This input from senior advertising professionals is absolutely important and, over time, can help refine AI to be smarter. Eventually, this need to tweak plans created by AI should diminish. When less than 10 to 15 percent of a media plan is being adjusted, with good justification, advertisers can be more confident that the influence of outside forces has been eliminated.
With this ability to enhance transparency and accountability in the buying process, an algorithm or AI can generate very specific media orders — a virtual shopping list for the media placement grocery store. Instead of having a broad category like “Local TV,” it can generate specific requirements by city, network, daypart, programs, websites and content.
Ideally, the algorithm would be directly connected to the system that actually executes the buy. This isn’t always possible today, but working toward that goal will help create an environment where we’re sure that what is planned is what is bought.
In the delivery phase, we know that certain media have well-discussed problems with fraud. There is no excuse for accepting fraudulent machine-generated impressions that have done nothing to help us achieve our business goals. Independent, transparent and automated measurement of the delivery of inventory is important to ensure that what is purchased is actually placed and delivered to the correct human audience. This, too, should be connected to the algorithm.
Ideally, the marketer is closing the loop with ROI data from sales, because this is the ultimate arbiter of the value received for the inventory delivered. This simple change will allow us to better identify messages and channels that truly move the market.
It is also the case that when fighting fraud, we cannot show too much of our hand, or we risk jeopardizing future lines of detection and the protection we need. The best marketers need to start to build an understanding of what fraud is and where it will potentially occur and build a roadmap we can follow in our own media buying to reduce that exposure.
Small steps? Giant leaps? We are at the start of a long journey. Look on the bright side: this is surmountable, and it will take down a load of bad actors in the ecosystem on the way.