Ad fraud (and how to mitigate it.)

This year, digital media has really had to stay on their toes — poised and ready to jump at any second. With all the consumer privacy regulation changes, and now, the real slime of the earth deciding to take ad fraud to a whole new level, there’s a lot to keep an eye on.

A recent profile of ad fraud in North America by TrafficGuard and Juniper Research estimates that within the next five years, advertisers will be losing $100 million a day due to ad fraud. That represents an increase of 125 percent since 2018. They’re also predicting that by 2023, 1 in 5 ad transactions will be derived from fraud. That sounds bad. But until everyone is clear on exactly what ad fraud means, we can’t move on to the solutions. So, here we go.

The simplest way to explain it is this: Ad fraud is when you pay for ads to appear somewhere online, and you think they did — because the reports say so. But no one saw your ad. The data was skewed from fake clicks from bots or click farms.

Even worse, from the vantage point of the brand or agency paying to run the ads, ad fraud is difficult to detect. Before we explain how we’re working to mitigate the risks of ad fraud for our clients (and ourselves), we want to make sure you’re up to speed on all the basic ways ad fraud can happen.

Common Types of Ad Fraud

Unfortunately, the slime balls behind ad fraud are no dummies, and they’ve managed to create multiple pathways to stealing your marketing dollars. The list of digital mediums where ad fraud rears its ugly head continues to grow, but there are several common pathways you’ll want to keep tabs on to start.

Click fraud

This is the most common, simplest type of advertising fraud. Fake traffic is generated through either an automated clicking platform or through click farms. A sign of this would be a higher than usual click-throughs that don’t lead to a sale.

Domain spoofing

Basically, this is impersonating a website. By building a domain and design that appears to be the intended site where your ad would run, the people running these sites can receive payment for your ads, but those ads will never be seen. This is also a common identity theft tactic.

Ad stacking

This happens a lot with programmatic ad placements. It’s when a publisher sells multiple ads on a website, but they’re all stacked on top of each other blocking those behind the top ad. The ads gain fake impressions, even though they’re not really being seen by users.

Pixel stuffing

Similar to ad stacking, only the ads are placed within pixels on the webpage, and they are not visible to potential customers. You end up paying for impressions when there’s zero chance of the ad actually being seen.

Location-based ad fraud

Location is super important to industries such as retail, quick-serve restaurants, the automotive space, and many others. These businesses rely on potential customers being close in proximity to their business. So, in order to take advantage of that, fraudsters fake the geolocations on mobile devices to place them around a specific location, causing brands to deliver ads to people in their target locations who aren’t actually there. With fraudsters taking advantage of all of these weak spots — and this is just a few — it can start to feel like the situation is getting more and more dire, but is it? We believe there are ways around the fake data.

The Ad Fraud Situation Is Bad, But It’s Also Manageable

The amount of digital ad fraud has grown to represent billions of dollars, and naturally, many marketers are now asking, “Is this still an efficient and effective channel for my brand?”

Yes, yes it is. As long as you know what you’re doing. While the numbers are concerning on the surface, the amount by which a specific brand experiences ad fraud with their campaigns will vary drastically, and ultimately lies in the hands of whoever manages their media. All programmatic buys are not created equal, and anyone that's running and optimizing a campaign strictly based on their Demand Side Platform’s (DSP) algorithm without strategic oversight is contributing to these crazy high fraud stats. In layman’s terms, allowing a DSP to completely automate the way you’re running ads, taking a set-it and forget-it approach is a risky move.

Back when programmatic media was popularized... Years ago, when programmatic buying was the new and shiny object, many of us were attracted to the idea that we could hit KPIs with the platform algorithm doing all of the heavy lifting. We would push long-tail impression volume to get the largest sample to then prospect from for conversions. Never mind that many of those conversions weren’t actions taken by real humans. And by the way, we paid for those fake events.

Smart media teams today should avoid that approach alone. While algorithms are still a close friend, we can’t put all of our trust in their judgment. In order to avoid fraud and maintain a level of brand safety, we also tap into premium publisher white lists, private marketplace deals (PMPs), walled garden inventory (i.e. Google FLoC products) and first-party data from site tags and customer databases. Our cost per thousand impressions (CPMs) aren't going to be dirt cheap, but the value of what we're delivering is higher and more qualified.

The fact is that fraud exists, and it is undoubtedly present with the campaigns we run. However, we’ve found an approach to buying and reporting programmatic media that involves multiple checks to minimize bot traffic and wasted ad dollars. Here it is:

How To Minimize and Prevent Ad Fraud AND Have Successful Digital Ad Buys:

1. Use analytics to examine the traffic our campaigns are driving. Check that the geography of the session matches the targeting, and look for suspicious patterns as related to site traffic, bounce rates, time and user path once on the site.

2. Run every programmatic buy through a third-party ad server to get unbiased confirmation that the impressions reported by your DSP match with what you believe was purchased. If they don't, negotiate a makegood.

3. Pay an additional CPM to layer filters from Double Verify, Integral Ad Science, MOAT, etc on top of programmatic buys.

4. Apply proprietary exclusion lists and exclusion lists from DSP to avoid known domains for fraud.

5. Review site transparency reports from campaigns on a regular basis. This is done in order to confirm that ads are running in brand-safe environments, and it will allow you to proactively identify any suspicious domains so that you can add those to your already existing exclusion lists.

6. Always Push to optimize digital campaigns to the most specific KPIs (not just CTR) where possible. While bots can still trick conversion tracking, if you're an agency or marketing partner, when you’re layering actual data from the brand over your reports, you will be able to validate what you’re doing as legitimate.

While the full picture of ad fraud will continue to evolve as the fraudsters learn from their mistakes and find ways around our fixes, the dangers still lie mostly in unmonitored programmatic media buys. So keep your eyes open, and keep reading articles like these to stay ahead of what’s coming. No one hates losing advertising dollars more than we do, so let us know if we can help shed more light on any fraudulent media concerns you may have.