Get the most important digital marketing news each day.
How to protect yourself against ad fraud
Columnist Ratko Vidakovic walks through steps that marketers should take to protect themselves and their ad budgets against fraud.
We have come a long way in our examination of online ad fraud. So far, we have learned: what ad fraud looks like, why it exists and who should be responsible for stopping it. Today, we are going to learn about the practical matter of protecting yourself, as a marketer, against ad fraud.
In a perfect world, fraudulent inventory would never make its way into the marketplace. Each participant in the supply chain would do their part to act as a necessary check on fraud.
Unfortunately, we don’t live in a perfect world, and not all participants are carrying their respective load. In practice, this means that it’s only wise that marketers take matters into their own hands to protect themselves and their ad budgets.
The following steps will help you take evasive action against fraud.
Step 1: Use pre-bid contextual data segments
The easiest way for marketers to protect themselves during ad buys is to leverage contextual brand safety segments. These contextual targeting (or really anti-targeting) segments should be available through your demand-side platform (DSP). This data comes from providers like comScore, Integral Ad Science and others.
The great part about these segments is that they are “pre-bid.” This means that you avoid the fraud altogether, at the outset, instead of reacting and filtering it out after you’ve bought the impressions. The cost of these contextual segments can be anywhere from $0.02 to $0.20 CPM.
Using comScore Bid Ratings, as an example, you can select segments that let you target inventory with specific levels of invalid traffic. You also can select segments for targeting inventory at certain viewability tiers, ensuring that your campaigns not only avoid fraud, but also get seen.
The trade-off in being too strict with these targeting segments is that you constrain your potential pool of inventory. So if delivery is a top priority, you might run into scaling issues. In that case, you may need to relax your tolerance level.
Step 2: Use an ad verification vendor
Ad verification vendors provide campaign metrics for advertisers, including viewability, invalid traffic and more. Examples of such vendors include Integral Ad Science, DoubleVerify, MOAT and Pixalate. The cost of these tools is usually a flat rate in the range of $0.01 to $0.07 CPM for post-bid measurement.
As a media buyer, using an ad verification vendor requires that you at least have your own ad server, so you can include the vendor’s tracking code in your own ad tags. That’s how you put these vendors in place to measure your campaigns.
If you have the resources, I suggest working with multiple vendors. That will give you the ability to consistently evaluate, compare and validate their data against each other.
Step 3: Analyze your campaign data
Once you have data coming in from your verification vendors, it’s important to start auditing your campaigns for signs of fraud. They will provide you with metrics like viewability and invalid (or suspicious) traffic levels. Using this information, you can optimize your campaigns by shutting down low-quality publishers and placements. You can then reallocate your budget to higher-quality sites.
Auditing your campaigns without a verification vendor is less scientific, but still possible. When looking through your campaign data, look for suspicious publishers and investigate them (a process we can explore another day). Also, look for unusual spikes in proxy metrics like CTR (click-through rate). Oftentimes, a CTR above one percent is suspicious, especially in an industry where the average is below 0.1%.
Whether you audit your campaigns with a verification vendor or without, you must be vigilant about monitoring for fraud. Like I’ve said before, the fight against fraud is always a game of whack-a-mole. Despite the efforts of SSPs (supply-side platforms) and DSPs (demand-side platforms), as a marketer, it’s prudent not to rely on them. The practice of active monitoring is especially important in the open exchanges, where the barriers are lowest.
Step 4: Maintain a universal blacklist
A common practice by many media buyers is the use of domain whitelists and blacklists. Whitelists are lists of domains that you want to explicitly target. They ensure that your ads don’t get served elsewhere, which tends to happen with bundled sites. Whitelisting is not perfect, but it’s a best practice if you are trying to target specific publishers. It’s usually a core function within most DSPs, so consider it a free tool at your disposal.
Blacklists, as you might have guessed, have the opposite function. An actively managed blacklist is a good way to block low-quality sites from your campaigns. So keeping a universal blacklist updated is an important practice. The information you get from auditing your campaigns in Step 3 is how you know what domains to blacklist.
The only downside to managing whitelists and blacklists is that it’s a reactive and manual process. It also isn’t perfect, since some bad actors can spoof their impressions to evade such tactics. Regardless, you should still use them — especially since they provide a competitive advantage against other less-diligent advertisers.
Step 5: Consider private marketplaces
Private marketplaces are “privileged” auction environments. They take place at a higher ad-serving priority than the open public marketplace. You’ll sometimes hear these referred to as “Deal ID,” which is just the technical term for the mechanism that powers them.
Most fraudulent publishers don’t have private marketplaces, for obvious reasons. (For one, they don’t have sales or ad ops teams to support such deals.) This makes participating in private marketplaces generally a safer option for marketers.
One downside to private marketplaces, however, is that they limit scale, which is one of the main benefits of programmatic buying. If you run a retargeting campaign, for example, you want to find your audience wherever they happen to be online. But if you limit your buying to a handful of sites, you naturally limit your reach. So that’s the trade-off.
Bonus tips for dealing with ad fraud
Aside from the best practices we just covered, here are some more tips to help you deal with fraud.
1. Avoid easy-to-fake campaign goals like impressions and clicks. Those are merely proxy metrics for actual business outcomes, like signups, sales and subscriptions. So it’s important to remember that. Focus instead on actual business metrics. If you are using any sort of auto-optimization algorithm in your DSP, avoid click-based goals. Instead, go for more concrete goals like conversions. It’s too easy for algorithms to get tricked by placements affected by click bots.
2. Embrace a defensive buying mindset. Being defensive means not trusting open or unmonitored inventory. It essentially means being a little paranoid. For example, it can mean avoiding unmonitored open exchange inventory altogether, and instead, pursuing more reputable publishers, like those with private marketplaces. Another way to be defensive is to avoid publishers with non-transparent inventory or those that bundle multiple sites together.
3. Refuse to pay for fraudulent impressions by blocking payment for fraud. No marketer should have to pay for fraudulent impressions, period. By taking a zero-tolerance approach to ad fraud, you, as a marketer, send a message to vendors that it’s unacceptable. And if it continues, you have to eventually ask yourself if it makes sense to continue fighting. Without the support of your DSP, it might make more sense to explore other partners or channels.
That brings us to the end of our series on ad fraud. There are definitely more angles to cover in the future, but for now, we’ve covered a good amount of ground. I hope you found this information useful. And if you have any questions, please don’t hesitate to reach out.
Opinions expressed in this article are those of the guest author and not necessarily Marketing Land. Staff authors are listed here.