A friend told me recently that I come across as “all doom and gloom.” I can accept that, as I have been researching and writing about digital ad fraud for the last 8 years. I describe the phenomena of digital ad fraud, the techniques used to commit it and cover it up, and cases that have finally been discovered and documented, all in hopes that more marketers would open their eyes to the fact that a part of their digital ad spend may be going into the pockets of cyber criminals. Time for a change in tact, at least for this article. I will focus on the good that I have seen and the success stories of how some have fought digital ad fraud and won.
Small Businesses Fought Ad Fraud and Won
I’m a small business owner and I always go “hell yeah!” when I hear of another small business owner beating digital ad fraud. I’m hearing more and more such examples, and this is very encouraging. A few classic examples include the following.
One small business looked at her Google Analytics in detail and noticed that when they turned on paid search campaigns, Android devices hitting her site surged by 100,000%. She isolated those campaign lines and figured out that the clicks were coming from search partner websites. These are sites that use Google ad tech to run search ads. They earn a portion of the ad revenue for every click; so they were using bots to click on ads and inflate their own revenue artificially. She turned off search partners in her campaigns and ensured her ads went on the main property — google.com — and solved her fraudulent click problem.
Another small business owner looked at the discrepancies between his Facebook advertising stats and the Google Analytics on his own site. Clicks on his Facebook ads should arrive on his site on a one-to-one basis. But he noticed up to a 90% discrepancy — Facebook reported 100 clicks but he saw only 10 arrivals on his site. Turns out many of the clicks came from ad impressions loaded on sites in the Facebook Audience Network (“FAN”). These are websites outside of Facebook that use its tech to show ads. These sites could inflate their own revenue by using bot traffic, that caused the ads to load and also clicked on them. Once he discovered where the fraud came from, he turned off FAN and solved it. The massive discrepancy went away and actual sales started to make a comeback.
Companies’ “Turn Off” Experiments Exposed the Fraud
Bigger companies also solved ad fraud by running experiments. A few of my favorites are the following. P&G turned off $200 million in digital ad spend, and saw no change in business outcomes. They went on to turn off more digital ad spend and reduce the complexity of their digital ad supply chain, to save more money and reduce more ad fraud. Chase reduced the number of sites showing its ads from 400,000 to 5,000, a 99% decrease. They saw no change in business outcomes. They turned off long tail sites that no one had ever heard of, were fake sites, or were suspicious in other ways, upon review. This saved them a ton of money that would have gone to these suspect sites.
Ebay also ran an experiment where they turned off their digital ad spend – paying affiliates to send them traffic and customers. What they found is that while the paid clicks to the site dropped, the non-paid clicks increased by the same amount. That meant that those users would have come to the site anyway, without having seen an ad and clicked on it. They realized much of their paid digital spend went to accidental or fraudulent clicks. Uber ran a similar experiment and paused their app-install marketing campaigns. They also found that when the spend was cut off, paid installs dropped, but organic installs rose by the same amount. Again, this meant that users were going to install the Uber app anyway, without seeing and clicking an ad. In Uber’s case, they were facing mobile app install fraud, where the mobile exchanges they were paying were falsifying records and in some cases fabricating them entirely.
Large Marketers Systematically Reduce Ad Fraud Themselves
A number of large marketers have also done well to see and reduce digital ad fraud themselves. The reason this is significant is because they don’t assume someone else took care of it and they are not relying on black-box ad verification technologies to detect ad fraud for them. Media agencies and other middlemen in the programmatic supply chain don’t have any incentive to clean up ad fraud because they rely on large volumes of dollars and ad impressions to make more money. Black-box detection tech also doesn’t cut it because there is no way to check if they detected correctly, or if they detected all the fraud anyway.
These large marketers took matters into their own hands, by getting more detailed reports and analyzing them more closely. By doing so, they could see things like the following. The number of ads served were far less than the number of bids won (which they paid for). Agency reports would only show them what they paid for, so they previously had no way to tell there was a problem. By analyzing and comparing the reports, this marketer found the fraudulent sites – the ones that had the greatest discrepancies — and turned them off. They also noticed mobile apps eating up sizable chunks of their impressions. For example, a flashlight app, emoji keyboard, and alarm clock app accounted for large numbers of impressions and appeared to be loading ad impressions continuously throughout the day. Do any humans leave a flashlight app on all day long? Again, by seeing this data for themselves, they could identify where the fraud was coming from and turn off those domains and apps that were causing the fraud.
Agency Teams and Individual Fraud Fighters Help Behind the Scenes
It is very rare that media agencies actually care about solving ad fraud, because it goes against their priority of making more money. But I will call out and praise two independent agencies that have been vigilant and truly have their clients’ best interest at heart – Camelot Strategic Marketing and Media, Heffler Media, and App Science, a division of Sabio, Inc. They have years of track record analyzing for ad fraud and reducing fraud for their clients’ campaigns. They do not rely on black box fraud detection; they rely on their own methodical analytics processes and experience to find fraudulent domains and apps and turn them off.
I will also give a shout out to individual fraud fighters within specific ad exchanges, and ad tech companies who do the hard work behind the scenes to systematically root out ad fraud. They are often met with resistance or hostility, even from their own sales teams because reducing fraud often means cutting off large chunks of ad impressions. They cannot speak out publicly, for fear of losing their jobs. But they continue to hunt fraud, heads down, to try to do the right thing.
Global Ecommerce Company Minds Their Own Stores, Closely
Further good news is that a number of World Federation of Advertisers (“WFA”) global advertisers are now monitoring their own digital ads and ecommerce sites with fraud analytics. They have “full funnel” measurement in place — 1) measuring the display ads that get sprayed out to thousands of sites via programmatic channels, 2) measuring the ecommerce sites to which the ads click, and 3) measuring the “thank you” and conversion pages, and finally 4) mapping their digital marketing activities to actual sales.
Again, they do not rely on black box fraud detection, although their media agency and the programmatic ad exchanges they buy from do. Their own routine analysis uncovers all the fraud that is not detected or reported by the fraud detection tech. They give their media agency short lists of domains and mobile apps to exclude every few weeks, and thus progressively clean their campaigns, even as new fraudulent domains and apps pop up. It boils down to something so simple and repeatable — turning off domains and apps — that it can be done with some common sense, if you have the right analytics to look at and know what to look for.
So, look, I took a break from the “doom and gloom.” There is good out there. There are individual people who are knowledgeable and doing the right thing, behind the scenes for fear of losing their jobs. There are specific small media agencies that are knowledgeable and tirelessly helping their clients fight ad fraud. The technology used in programmatic digital advertising is not bad, and offers many advantages simply not possible before. But like any technology it can be used for good, or it can be abused by fraudsters for their own gain.
As long as marketers and their true partners (not vendors and agencies that call themselves “partners”) are vigilant and have the right analytics to look at, much of the ad fraud can be solved. They won’t need to solve fraud for the whole industry, just for their own campaigns. And with that, I see light at the end of the tunnel, and a rainbow after the storm. Let’s together move out of the darkness of ad fraud, and into the light.