Data & Analytics

How Ad Fraud Impacts Your Measurement, with Dr. Augustine Fou

September 6, 2016  |   LiveRamp

This is part two of our interview with Dr. Augustine Fou, digital ad fraud researcher at Marketing Science Consulting Group. For the first part, click here.

Ad fraudsters not only steal your ad dollars—they also mess up your marketing analytics and prevent you from truly understanding the impact your digital has on sales.

Thankfully, the tools they disrupt are the same ones you can use to spot those nefarious bots. In this second part of our interview, Dr. Fou will tell you how.

How sophisticated do my measurement capabilities need to be to spot bots?

Everyone should be trained and on the lookout for fraud. When you see something strange in the data and analytics, it probably should raise eyebrows and lead to further investigation.

I usually like to say, “anything that is too high, like very high click-through rates, or anything that is too low, like very low bounce rates,” should be treated as suspicious and studied further.

These are things that can be done by the marketer’s own analytics teams, and do not require paying for expensive new technologies.

Should I build in ad fraud avoidance tactics into my measurement from the start, or should I wait until I’m a little more advanced?

Every marketer should start with good, clean practices right away, knowing that fraud is everywhere. These practices can be obvious and simple to do.

Here are some examples of what you can do:

  • Lower your ad spend in the overnight hours—like 1am–5am—when humans are sleeping.
  • Reduce your ad spend on the weekends, when fewer humans are going to websites.
  • Focus on using metrics that are directly correlated to business outcomes, like sales and other conversion metrics, rather than quantity metrics that can be easily faked by bots—like the number of visitors, impressions, or clicks.

Every marketer can do this right away, even if they are just starting out.

How can I insert spotting tactics for ad fraud into my existing measurement?

Many marketers can start with their own analytics platforms. They should focus on the detail—for example, insist on getting line-item details for things like ad placements.

In the webinar, there was one slide where we saw 100% click-through rates:

Screen Shot 2016-07-25 at 3.34.36 PM

Those are greedy bad guys who programmed their bots to click on every link to earn the CPC ad revenue.

If you have line item details, you will see the obviously fraudulent CTRs of 100% or more. When you get rolled up average numbers on a monthly basis, all those details are lost and the fraud continues.

By insisting on line-item details, marketers can stay vigilant and spot obvious fraud in their own existing measurement platforms without spending money on advanced bot detection tech.


For more of Dr. Fou on ad fraud, its impact on your analytics, and how to stop it, read this RampUp blog.