Tapping AI Technology to Reduce Ad Fraud
Advertisers have come to expect some level of digital ad fraud, but it's getting out of hand. The Interactive Advertising Bureau (IAB) found that 36 percent of ad web traffic is fraudulent. That's a significant amount of an advertiser's budget going down the drain. And brands seem to have given up hope. Short of manually going through each impression to determine its impact, there's not much else they can do. However, one way brands can fight back is though programmatic media buying and artificial intelligence (AI) technologies, which can play a significant role in reducing the incidence of fraud.
The problem is that most marketers still don't have a good understanding of what programmatic buying is, as revealed in a recent Forrester survey showing that only 23 percent grasp the concept. For those marketers in the other 77 percent, programmatic media buying is the process through which machine-learning technologies like AI optimize advertising campaigns to ensure they meet key performance indicators. Rather than spend time identifying the right combination out of billions of campaign possibilities, marketers can plug their goals, such as audience, size, time of day and geography, among others, into the system and let AI take care of the rest.
Manually optimizing a campaign requires an advertiser to cast a large net and cut out the publishers that aren't delivering high-quality leads. AI, on the other hand, starts with a smaller pool of qualified targets and builds the campaign based on aspects that deliver the highest conversion rates.
What does this have to do with ad fraud?
First, it's important to understand how ad fraud works. The industry generally defines fraud as the deliberate practice of serving advertisements that have no potential to be viewed by a human user. Some of the more common types include the following:
- Bots: Shady publishers can purchase bot traffic to boost their page views and drive bidding wars. Of course, none of that traffic will actually result in any leads or conversions for the brand since the users aren't real.
- Retargeting schemes: Another use for bots involves retargeting, where an individual bot will mimic user behavior by clicking on several related web pages before finally selecting the ad.
- Altered viewability: Through iFrame "stuffing," an ad can be shrunk to a 1x1 pixel, making it all but invisible to consumers. Or multiple ads can be layered over one another in a practice known as ad stacking, rendering all but the top ad useless.
- False attribution: Despite not generating any real impressions, fraudulent ads can take credit for sales or conversions, stealing ad spend away from real, profitable channels.
There are mountains of data an advertiser can comb through to reduce fraud manually. This is not only time consuming, but also cost prohibitive. Furthermore, fraudsters are constantly finding new ways to get brands to fork over more of their campaign budgets, and keeping up with the latest tricks of the trade can be a full-time job in itself. There's also basic human error to consider; we all make mistakes, and we have a tendency to repeat them. This is where AI can really shine.
Because AI is a self-learning technology, it can more quickly identify and isolate incidences of fraud. It does this by analyzing traffic ratios and how those translate into impressions or conversions. In tracking IP addresses, AI can determine if a bot is being used. By constantly evolving and improving its rules and definitions in real time over the entire duration of the campaign, AI automates the decision-making process in weeding out fraudulent activity.
Of course, manual optimization is still important. AI isn't meant to replace advertisers, but support them in achieving their goals faster. Just like your email's spam filter won't catch every bad email without the user's input, AI relies on advertisers to flag certain websites they know engage in fraudulent activity.
Advertisers view ad fraud as a necessary industry evil, but it doesn't have to be this way. The marriage of AI technologies and manual optimization creates a fraud-fighting machine that keeps campaigns on track and puts those previously wasted dollars back into driving leads and conversions, allowing advertisers to keep hitting those key performance indicators.
Or Shani is the CEO and founder of Adgorithms, which combines the power of machine learning and artificial intelligence to maximize returns on ad campaigns.
Related story: 5 Social Media 3.0 Trends to Watch For