While ad fraud has become part of every marketer’s vocabulary, attribution fraud—the practice of gaming outdated attribution models to justify self-serving means—has been mostly ignored up until now.
As marketers and media owners shift dollars into digital channels, however, serving ads outside of an attribution model’s measurable capabilities in order to achieve financial benefit has quietly been ramping up in response. Attribution fraud is fast becoming the ad industry’s next big quality headache.
Attribution fraud includes everything from retargeting users about to convert to knowingly cookie-bombing users with non-viewable ads. And the costs of attribution fraud are not just the valuable advertising dollars wasted on ineffective media.
Attribution fraud leads marketers down a rabbit hole of fundamentally misunderstanding their customers’ behavior. It leads advertisers to misinterpret how various channels, devices and tactics lead to consumer actions.
But all hope is not lost. With advances in cross-device attribution and experimental design methodologies, marketers are able to create randomized controlled experiments that measure combinations of investments in channels, regions or media types.
These randomized controlled experiments are able to accurately attribute business outcomes to specific marketing activities—and leave scant opportunities for fraudsters in the process.
Don’t leave the door open for fraudsters
The multifaceted, multi-channel nature of consumer behavior and advertising programs today has made true attribution difficult to measure—but marketers tend to default to what they can easily measure. And what is easy to measure is often first or last touch.
Fraudulent ad impressions without value frequently clutter attribution methodologies such as last touch/view. Marketers are unwittingly leaving the door wide open for fraudsters by using attribution models which are neither accurate nor effective. And even sophisticated multi-touch or split-funnel attribution models still rely on impressions that are measurable—leaving these models vulnerable to fraud as well.
Take a look at advertising investments relative to desired outcomes
Marketers should consider taking a more holistic view of campaigns across all channels to better protect themselves from fraud.
One way to do this is by activating a scientific experimentation framework through a cross-channel technology measurement partner. Here’s what such a framework looks like:
1. Run many small, isolated experiments to understand the impact of each investment on desired business outcomes.
Marketers can run many parallel experiments to achieve a holistic understanding of their advertising investments and returns. This “design-of-experiments” methodology is superior to traditional media-mix models for three reasons:
- Runs in real-time. Media-mix models measure the relationship between historical investments and outcomes over multi-year periods. But multi-year models are often outdated by the time they are complete. Design-of-experiments methodology, in contrast, allows advertisers to run many short-duration, randomized controlled experiments in order to pulse different levels of media investments across regions, devices and channels.
- Normalizes reporting. Existing attribution models suffer from the lack of apples-to-apples comparisons of outcomes from digital channels. The design-of-experiment model uses true outcomes data (i.e., offline or online sales), however—which remains comparable no matter the channel—to calculate the optimal investment level and saturation level of each channel. This method can include purely offline channels such as print and traditional TV in addition to digital channels
- Demonstrates causation instead of correlation. Traditional models are only able to show correlation—not causation. The learnings from the sum of experiments, however, can be used to establish a causal link between marketing outcomes and media investments.
2. Consumers use multiple devices, so all measurements should, too.
Consumers use multiple devices to research, plan, compare, consider and complete their purchases. When looking at attribution models, it is important to consider both the cross-device nature of today’s modern consumer and the role of each device within the conversion path.
Graph technologies identify unique users across multiple touchpoints, including touchpoints outside of paid media, and are crucial to identify and curb impression-level attribution fraud. Transparent multi-touch attribution models built on this foundation can help advertisers recognize fraudulent impressions and discredit them, thereby assigning credit to the rightful touchpoint in the consumer journey.
3.Work with partners you trust.
As more money flows into an industry, the incentive for fraudsters to find new ways of gaming the system increases. The advertising industry is no exception to this trend. So just like in any other financial relationship, it is very important for advertisers to know and trust the partners and vendors with whom they are working.
Given the vast amounts of money spent on digital by global advertisers, there has never been a better time to transition to more complete attribution models that represent the reality as a minimum standard.
Last-touch modeling is outdated, impression-only models are incomplete, and walled gardens (intentionally) make it difficult for advertisers to uniformly measure advertising performance. As digital marketing becomes more complex and fragmented, it is critical for marketers to modernize their attribution modelling to ensure they fully understand the ROI of their media investments.
The good news is that over 50% of marketers in US companies already plan to use multichannel attribution models in 2017, according to eMarketer. These marketers have the right idea. Only a holistic look at advertising investments across all devices, channels and media types will lead to a true understanding of attribution. And only then will attribution fraud become a thing of the past.
Advertisers are more concerned than ever about brand safety, and one of the primary ways they're trying to keep their ads from appearing in unfriendly places is through whitelisting. But as more and more brands turn to whitelisting, some are talking about the impact this will have.
We all know that Facebook is a viable source of huge amounts of mobile traffic with relatively cheap CPCs). It’s too good an opportunity to ignore in today’s digital landscape - even if your mobile landing-page experience isn’t up to snuff.
For years, advertisers have tolerated a big elephant in the room: the fact that their digital ads aren't always appearing where they would want them to.
Deep learning tools are the next major area of AI-based research, and it will spark a wave of future innovation in every industry – bringing a new era of marketing which both advertisers and end-users will benefit from.