Dr. Augustine Fou shares tips for identifying good publishers in a sea of bad actors.
We recently sat down with Dr. Augustine Fou, a recognized thought leader in digital strategy, integrated marketing and ad fraud research. In part two of our two-part interview, Dr. Fou shares his experiences with ad fraud, the bots behind it and how bad actors take money away from good publishers and advertisers.
In part one, Dr. Fou talked about his experiences with ad fraud, the bots behind it and how bad actors take money away from good publishers and advertisers.
What are some ways to identify legitimate sites from fake ones?
Publishers, especially those that are part of AAM, are not getting their fair share of credit as fraudulent sites are getting lumped under the label of “publisher” just because they sell ad inventory. By giving all websites the “publisher” label, we are doing a disservice to the good publishers. You have to look at a publisher’s business practices to know if they are doing the right thing. Good publishers don’t cheat, they don’t buy traffic and they don’t artificially inflate their ad inventory by reloading the page every 30 seconds. Bad publishers do all of that to cheat buyers.
The publishers that started out in print have the infrastructure to create websites and content that humans want to read. A lot of other publishers either don’t carry content or steal content for their sites. The pirated sites steal content from movie and music studios and they use that to attract traffic and make money on ads. There’s only a finite set of good publishers that have real content that real humans want to read. A lot of them are already AAM clients.
Even if you take the top 100,000 on Alexa, that’s going to account for 99.999 percent of humans visiting websites. Publishers with recognizable names account for most of the humans. But when it comes to programmatic buying, it’s hard to clean up the thousands of long-tail sites because there are so many people who have vested interest in it and don’t want it cleaned up.
What are some business practices to look for when evaluating the legitimacy of a site?
Legitimate sites naturally have clean traffic as a lot of humans go there because they publish real content that real humans want to read. A lot of fraudulent sites do not show signs of human activity. A lot of questionable long-tail—and at times outright fraudulent—sites have no content whatsoever, so there’s no reason for any humans to visit.
These publishers typically buy traffic to be sent to their site. This purchased traffic is generated by bots that can produce as much traffic as the bad publisher wants. So, if we measure bot activity on sites across the digital ecosystem, we can start to differentiate between good publishers and the fraudulent ones.
There are publishers in the middle that produce content and have some human audience but started buying traffic to grow their ad revenue. But once you show your boss that your site has more traffic and more ad revenue, what happens if you stop buying it? This is an unscrupulous business practice that some publishers have resorted to as a way to grow their ad revenue.
Should an advertiser see bots in their campaign results if they’re working with a publisher who does the right things?
Advertisers need to be wary of publishers that have some human traffic but are not adding filters to prevent ads from being shown to bots. Typically, these publishers don’t want to lop off their own revenue, so they keep doing it until they get caught. In those cases, the advertisers have to detect it themselves and that’s why there’s always been this mistrust. Even though the publishers have tracking tags on their website, the advertisers don’t trust the numbers that the publishers tell them because not all publishers are doing the right thing.
Why does ad fraud keep happening?
These are hackers who will always find a workaround to whatever solution we create to block them or whatever technology we use to detect the fraud. A tech arms race is not a long-term solution to the fraud problem. While we need the technology to detect fraudulent activity, we also need to vet the publisher’s business practices. At the end of the day, a good publisher does the right thing and filters out a bot, and a bad publisher doesn’t. I believe that the long-term solution is a fusion of using ad detection technology and looking at publishers’ business practices.
What’s one piece of advice you would give to advertisers and publishers to address ad fraud?
Keep being vigilant and keep pushing for details. When your media partners and vendors provide information in one rolled-up number, ask questions. Fraud hides easily in the averages, and if you accept them, you’re helping ad fraud continue to exist. There’s a lot of common sense here that doesn’t require high-tech solutions. When you see the same amount of traffic come to your site every single day in a perfectly horizontal line that shows the same quantity over time, that’s not real.
Ad fraud is still prevalent. It has not gone down despite the industry setting standards or the proliferation of fraud detection companies. There is something else going on here, and throwing more technology at it is not going to help. There needs to be another solution. The key thing for both advertisers and publishers to do is to directly measure things and insist on full transparency with line-item details on the impression-by-impression level or visit-by-visit level. Ad fraud will continue if we don’t pay attention to it.