Fraud, FloCs and practicality of Cohorts

by Alex Andreyev

Feb. 2021

The promise of digital advertising over the past decade proved the benefits of personalized marketing, direct results, and improved effectiveness over old-school media. We saw advertisers pull budget from expensive SuperBowl ads, and tech-driven agencies double down on people-based targeting. Today DTC and eComm companies’ growth is through the roof. Or is it?

Let’s start with Fraud. Nandini Jammi, Co-Founder of Check My Ads, does a great job breaking down how Uber uncovered $100M worth of fraud in their marketing investment. Wow, that’s 2/3 of their estimated marketing budget. Uber is a company that is considered a savvy technology developer with extensive analytics and measurement teams. So where does that leave the rest of the marketers? How do you catch fraud when you don’t have a process to track an impression, to a click, to a sale all within the same path? What can consumer brands do that have the majority of their sales at brick and mortar? P&G stock didn’t take a significant hit after cutting $200M in advertising budget in 2018, they just made their money work harder and then expanded into what works. And that’s exactly the point, if measured properly, sales will drive out fraud. Bots don’t buy things, cookies and IDFAs don’t buy things, people do. Fraud will always exist as long as there’s money to be made and people exist, but if marketers go beyond reach, impressions, clicks, and installs, and measure the impact media has on sales, dollars will move to real media and content.

Measuring fraud sounds great, but measurement has its own challenges. As the dominant force in advertising, Google continues to shake things up by pulling 3rd party measurement from YouTube and most recently doubled down on their sandbox environment. Either they’re building higher walls for their garden, cutting out competition and walling in advertisers, or they’re responding to political pressures for privacy (arguably a problem they helped create.) In either case or maybe both, they might be onto something. While there are great discussions on this topic, we agree with Myles Younger of Mighty Hive,

“The idea that 1:1 was the holy grail of digital advertising was always a fallacy. It’s great to see Google blowing up that myth.”

From a practical standpoint, no marketer is evaluating individual purchase behaviors. With 1000s of variables against each individual user, the only practical way is to create cohorts anyway. So, while maybe not as intriguing as the fantasy of people-based targeting, cohort-based targeting, analysis, and measurement is an effective middle ground and leaves the creepy effects of hyper-targeting out. Creating cohorts based on traffic sources or suspected user behaviors connected to sales or other business driving activity can quickly identify the cohorts of fraudulent inventory providers.

At Evidnt, we look at purchase behaviors across thousands of stores and millions of transactions across the US, and realized that individual performance was directly correlated to their neighborhoods. In 2020, by removing all user-based IDs and building a platform on hyperlocal behaviors of neighborhood cohorts, we were able to drive more sales and growth for our clients within the CPG and alcohol categories. The benefits of such targeting align Evidnt with Google’s approach to measurement.  This also expands brand capabilities to measure all media since Evidnt doesn’t need user IDs to plan, activate or measure purchase behaviors. Our anticipation is that the writing is on the wall for targeting using 1:1 measurement, and neighborhood cohorts future proofs our client capabilities for targeting and optimization, with or without cookies. 

So we’re ready to fly with the FLoC.

Learn More

For more information about Evidnt:

Contact Us

Help Us Support Ukraine!

X