Fraud isn’t an easy topic for the ad industry to discuss — in fact, I was the only speaker at GDC 2016 to tackle it. Many people from the audience asked me why are there not more resources out there to educate developers about fraud. That thought resonated with me, and in light of recent online conversations from a few Chartboost suspended accounts, I want to address those concerns with as much transparency as possible.
Let’s face it, advertising fraud presents an enormous, costly, and often undetected challenge for game devs. A study from the Interactive Advertising Bureau estimates that advertising fraud costs advertisers over $8 billion per year, with nearly $1.3 billion of that on mobile devices. This means ads are served but not viewed by human users — allowing fraudsters to cash in on the impressions. Right now mobile advertising is the fastest growing and most lucrative of any advertising platform, meaning the challenges of fraud will only get bigger.
When we started building Chartboost 5 years ago, it was critical for us to have a clear developer-first mentality, growing deep partnerships with the developer community based on trust, transparency, and control. I strongly believe that’s why today Chartboost is the largest mobile games marketplace where over 300,000 apps acquire valuable players and monetize their audiences. But fraudulent traffic jeopardizes the pillars of our platform as it lowers the quality of our network and muddies our partnership with publishers. We take fraud very seriously.
We’re all responsible
While often the challenge falls to the ad networks, the entire ecosystem is responsible for fraud prevention. The only way to protect the industry is to work together.
Here are a few ways to start from wherever you are in the industry:
In a CPI/CPA world, the ad networks invoice advertisers based on converted installs. This data is sent from the attribution platforms to the ad networks. The advertisers often look to the ad network to ensure quality installs, but they should also expect attribution platforms to be vigilant in preventing fraudulent behaviors.
In 2016, every advertiser should request transparency and control to optimize their campaigns on the fly and filter out low quality publishing apps. This protects the advertiser and contributes to a healthier network for ad partners.
As mentioned above, many fraudulent sources can be detected by looking at the attributes of an install. Attribution platforms have unique access to broad data within the app ecosystem and should leverage this intel to better prevent fraud. Signals such as abnormal time between clicks and installs, blacklisted or illegal IPs, VPN traffic, and increased duplicate installs from a single device are data points that should be automatically detected and managed on the attribution side. The main attribution platforms are already investing in advanced preventative technology.
Protecting the health of your network should be the top priority as an ad platform. Every ad platform needs to be serious about devoting internal efforts to combat fraud. From analyzing traffic sources to quickly acting on fraud signals, these mechanisms improve accuracy the longer they are in motion. And the more consistent we are across the industry, the less false-positives and the more successful we become in winning the battle.
When a publisher is flagged by fraud detection methods and suspended, I can understand the frustration. We hear the feedback across our channels. Ultimately, though, the responsibility to prevent fraud falls to the publisher. When a game developer opens their inventory to ad networks, it’s their job to ensure the traffic they provide is legitimate. This includes staying within any network’s terms of service. For developers who are just getting started, here’s a list of best practices to keep your app healthy.
We have zero tolerance for ad fraud
At Chartboost, we strive to make it easy for all mobile game developers to join as a new publisher. We also invest heavily in fraud prevention because it’s our duty to create a healthy, trustworthy ecosystem for both advertisers and publishers. Our traffic sources are closely monitored as we leverage hundreds of signals powered by machine learning algorithms and our human-powered industry expertise to stop fraud. We don’t tolerate fraudulent behavior and shut down accounts that present clear evidence of fraudulent activity.
That said, with automated systems there can be false-positives and we have made mistakes in the past. When we’re wrong, we’ll own up to that (and I have personally apologized). We’re continually improving our fraud algorithm with new signals to better root out fraud activity. I want to reiterate our commitment to making fraud detection mechanisms faster, smarter and more reliable. In the end, it’s about protecting all of the legitimate developers in the network.
Let’s fight it together
I know that fraud is a delicate topic to talk about publicly. On one hand, you’re acknowledging the existence of fraudulent traffic in your product. On the other hand, we as an industry have to limit details we share about fraud management because of the risk with fraudster workarounds. It’s also why ad networks have to be vague about the reasons why they suspend accounts.
The reality is that we’re all affected. We need to all take ownership and work together toward a real solution. This week, we’re proud to join our partner Adjust in their new alliance with 12 ad tech companies to combat fraud. Now, what are the next steps we need to take to build a fraud-free industry in 2017?