According to Forrester, ad fraud will reach $10.9 billion in wasted spend by 2021. This figure speaks to marketers’ constant struggle with digital advertising fraud and the havoc it wreaks on targeting, budgeting, and measurement – if you can’t even be sure that you’re reaching a human being, let alone the right ones, how can you justify marketing spend and set realistic metrics?
We spoke with Dr. Augustine Fou, digital ad fraud researcher at Marketing Science Consulting Group, to assess the current situation with ad fraud, discuss ways to prevent it, and bolster the business case for people-based marketing.
Q: By most projections, ad fraud is only projected to increase. To what can we attribute this growth?
A: A few factors are contributing to the continued increase in ad fraud. First, dollars continue to pour into digital. Even TV ads are starting to be bought and sold programmatically. So as the ad dollars grow, the ad fraud dollars also go up.
This has become such a large pool of dollars that every criminal element wants in. They have access to black hat hackers who make more and more advanced bots that can get by most, if not all, forms of fraud detection. So bad guys have both the motive and the means to make a ton of money from ad fraud.
Q: One way of spotting ad fraud that you’ve spoken with us about is simply looking at the data to scrutinize the peaks and valleys and getting line-item details from analytics platforms. As this was a year ago, do you think marketers’ analytics have matured past this point in detecting ad fraud?
A: Unfortunately, no. There remains an enormous education curve because historically, marketers have handed off the media buying to agencies. The reports and analytics they got back were simply not detailed enough for them to do anything with it themselves.
The good news is that this year, more and more marketers have taken matters into their own hands. Some have brought programmatic media buying in-house, giving them much more control, far lower “middle-men” costs, and far more detailed analytics to work with.
Others are running their own ROI experiments, like cutting digital programmatic spend and seeing if there is any drop in sales. If they see no reduction in business outcomes after the reduction of ad spend, then that ad spend may not have been worthwhile, whether it was due to fraud or not.
Long story short, the analytics still have a ways to go, but there are other experiments marketers should do now to double check their digital media for themselves.
Q: What other tactics and techniques should marketers practice and master to minimize ad fraud now and in the future?
A: Marketers need to optimize their campaigns to actual business outcomes. This is first and foremost. Any other quantity metric can be easily gamed by fraudsters. Even quality metrics like click-through rates can be fine-tuned by bots. If marketers optimize campaigns for actual business outcomes, even if the measurement is challenging, they will quickly see which media sources are not performing and which ones are. It will become painfully obvious that in some cases their ads are not being shown to human audiences and in other cases they are. They can then optimize their campaigns by shifting ad budgets to more productive media sources.
Note this does not even require any advanced technology to implement. After the first technique, marketers can look at their own analytics to see if there are telltale signs of suspicious activities and actively scrub them out.
Q: Can you share any success stories from companies who have stemmed the tide of ad fraud in their organizations?
A: Yes, there are success stories from both the advertiser side and the publisher side. P&G cut $140 million from their digital ad spend and saw no drop in volume. Chase reduced the number of websites carrying their ads from 400,000 to 5,000 — a 99% decrease — and saw no change in business outcomes.
On the publisher side, many organizations have reduced or eliminated traffic sourcing and audience extension, both of which increase their risk of fraud and bots. Good publishers have also put filters in place filters to weed out the obvious data center traffic and named bots in order to protect their advertisers from being exposed to these sources of general invalid traffic (GIVT).
And finally, some publishers have implemented ads.txt and/or stopped selling their unsold inventory through programmatic channels entirely. This greatly reduces their risk of being spoofed, which is when bad guys’ sites pretend to be their site in order to sell fake inventory into the exchanges.
Q: Does the reality of ad fraud prevent companies from investing as much as they could in people-based marketing?
The rampant amount of ad fraud should compel marketers to focus on people-based marketing.
A: The rampant amount of ad fraud should compel marketers to focus on people-based marketing. I think most marketers assumed that digital advertising was already that because they assume that their ads were shown to humans. Unfortunately the fraudulent activity of bots means that is not the case.
So we are basically getting back to good old fashioned marketing — to people! Marketers can ensure their ads are shown to people by buying media from good publishers or by using onboarded data that came from offline data sources where real humans could be verified.
Q: What else needs to happen across the ecosystem to make fraudulent activity far less damaging and/or lucrative?
A: The ball is really in the court of the marketers because ultimately it’s their ad dollars. As long as marketers continue to buy on a reach and frequency basis and continue to demand more and more “inventory,” the bad guys will continue laughing all the way to the bank because they can generate as much quantity of impressions as marketers want to buy.
Good publishers have already done their part. But human audiences are finite, and actually pretty small when you layer in targeting parameters. Human audiences are earned over time with good quality content.
If marketers bought high quality and small quantity, they would have already reduced their risk of ad fraud and increased the chances of good business outcomes.