Ad networks traffic, fraud, and the calculation behind it

Ad networks connect advertisers with inventory, but the buyer still needs to understand delivery, fees and the quality of reported events. Start with consistent definitions for impressions, clicks, installs and purchases. Investigate unusual patterns against raw reporting and product outcomes before deciding whether the issue is a measurement mismatch, invalid traffic or fraud.

How Do Ad Networks Serve Ads?

Let’s start with how ad networks serve ads. Ad networks work by partnering with publishers and advertisers. Advertisers launch ads on the network, which then serves them to the publisher’s website or app. Ad networks use various targeting methods to ensure the ads are shown to the most relevant audience. This can include demographic targeting, behavioral targeting, and contextual targeting.

How Do Ad Networks Make Money?

Ad networks make money by taking a cut of the ad revenue. Typically, the advertiser pays the ad network on a cost-per-click (CPC) or cost per thousand impressions (CPM) basis, and the ad network pays the publisher a portion of that revenue. The ad network takes a percentage of the revenue as their fee. Some ad networks also charge fees for services such as ad optimization and fraud prevention.

So, What is Ad Fraud?

Ad fraud involves ad networks exaggerating impressions, clicks, or conversion data to gain a financial advantage while squandering the advertiser’s money. The activity encompasses a wide variety of dishonest practices in internet advertising. This can result in advertisers paying for ads that were never actually seen by a human. Examples of ad fraud include click farms, bot traffic, and ad stacking. Ad fraud not only harms advertisers but also damages the credibility of the entire industryCybercriminals who work alone or in groups can perpetrate ad fraud. No matter who is responsible, the outcome is always the same: businesses squander money on ads that the target consumers do not see.

Do Ad Networks Hijack User Devices?

One concern around ad networks is the potential for device hijacking. This occurs when an ad network serves an ad that contains malicious code, which then takes over the user’s device. This can result in anything from annoying pop-ups to serious security breaches. While there have been instances of ad networks inadvertently serving malicious ads, the industry has taken steps to combat this issue. Many ad networks use advanced fraud detection tools to ensure that only safe and secure ads are served to users. Yet, ad fraud is still a significant issue in the industry.

Examples of Fraudulent Ad Network Activity

Ad fraud is also known as invalid traffic (IVT) because it generates activity that doesn’t originate from a real user. There are two kinds of IVT:

• General invalid traffic (GIVT) can be detected using standard filtration techniques like lists or other common tests.
• Sophisticated invalid traffic (SIVT) is only detectable with multi-point collaboration, advanced analytics, and/or human intervention.

Here are some of the SIVT tactics employed by ad fraudsters:

• Domain Spoofing
• Click Fraud
• Cookie Stuffing
• Click Injection
• Pixel Stuffing
• Ad Stacking
• Ad Injection
• Geo Masking
• Bots/Non Human Traffic

Most fraudulent ad networks use a combination of these tactics together to defraud advertisers. Let’s examine a couple of high-profile cases.

The Methbot Scandal

Uncovered in 2016, the fraud involved a group of cybercriminals who created a network of fake websites and bots to generate fraudulent views and clicks on video ads, resulting in advertisers paying for non-existent views. The Methbot fraud scheme was highly sophisticated, with the criminals creating thousands of fake web domains and using data centers to make it look like the traffic was coming from legitimate sources – a mixture of both geo masking and a sophisticated example of domain spoofing. The operators of Methbot used a network of proxies and fake IP addresses to hide the fact that the ad views and clicks were actually coming from servers located in data centers in Eastern Europe. The use of these proxies and fake IP addresses allowed the scheme to evade detection and make it appear as though the ad views and clicks were coming from legitimate U.S.-based users. At its peak, the scheme reportedly generated between $3 million and $5 million in daily fraudulent revenue.

3ve

The 3ve ad fraud scheme was a complex operation that involved multiple types of fraud designed to deceive advertisers and steal their ad dollars, including ad stacking, pixel stuffing, domain spoofing, bot traffic, and faked clicks and installs. 3ve was involved in a botnet of over 1.5 million infected computers and mobile devices. The scheme worked by creating fake websites that appeared to be legitimate and then using the botnet to generate traffic to those sites. The traffic was then sold to advertisers, who believed that their ads were being viewed by real people.  However, the traffic was actually generated by bots that clicked on the ads and made it appear as though they were being viewed by real users. This made it challenging to detect the fraud as the bots seemed to behave like humans. These cases are two high-profile, very lucrative ad fraud schemes that were eventually spotted, and the perpetrators (both were actually run by the same group) were charged and ordered to pay out to the FTC. However, while these schemes included an eye watering amount of money, they represent just the tip of the iceberg regarding the disingenuous agents in the digital ad space.

The Cost of Ad Fraud 

A study by Forrester found 69% of brands spending $1 million per month reported that at least 20% of their budgets were being lost to digital ad fraud. In 2018, Adobe found that potentially 28% of web traffic came from bots or other non-human actors in a sample of their thousands of client websites. Based on this finding, one commentator estimated that the total cost of ad fraud might be as high as $66 billion. In 2022 and 2023, while many believe AI and blockchain technologies can help prevent ad fraud, the fact remains that estimates are still extremely high for the overall cost of ad fraud.

So, should you trust whatever your ad network is reporting?

Ad networks play an important role in the digital advertising ecosystem, but it is essential to approach the data they provide with a healthy dose of skepticism. This is especially true in cases where the reported numbers seem too good to be true or defy basic logic. Let’s take a look at a clear example. This is a real situation that happened with one of our clients: their target country (South Africa) counts 60M people. The network reported 90M clicks/month, which prompted a review of repeated clicks, reporting definitions and downstream outcomes. A click total can exceed the population, so the comparison alone does not establish fraud. The team also consulted DataPortal’s Digital 2023: South Africa report, which reported that more than a quarter of the country’s population lacked internet access. That historical population context does not convert a click count into unique people. The team cross-checked the reporting and paused advertising. For an investigation, request the impression, click and conversion definitions, reconcile independent app data and document unresolved discrepancies before deciding whether to continue buying the traffic.

How do publishers make money?

Publisher payment depends on the commercial arrangement and cannot be inferred from the advertiser’s CPI alone. In a hypothetical example, 1,000,000 clicks at a 1% CTR imply 100,000,000 impressions. If half those clicks produce installs at a $1 CPI, the advertiser spends $500,000 for 500,000 installs. That implies an advertiser-side CPM of $5 before separating fees; it does not tell you the publisher’s net revenue. Check that the same reporting period, event definitions and cost scope are used in every calculation. Then investigate the supply path and event quality rather than treating a low CPI as proof of fraud.

1. Validate and cross-check the data: Always validate the numbers provided by your ad network by comparing them with industry benchmarks and your own internal data. Look for any discrepancies or unusual patterns that could suggest fraudulent activity.
2. Investigate the ad network’s reputation: Research the ad network’s history and reputation in the industry. Check for any red flags or past instances of fraudulent behavior. Look for reviews and feedback from other advertisers who have worked with the network.
3. Use third-party tools and services: Partner with third-party ad verification and fraud prevention services to analyze and validate the data provided by your ad network. These services can help you identify and mitigate the risks of ad fraud.
4. Request transparency and detailed reporting: Ask your ad network to provide more detailed reporting, including data on impressions, clicks, conversions, and other relevant metrics. This will enable you to better understand the performance of your campaigns and identify any potential issues.
5. Maintain open communication with your ad network: Establish open lines of communication with your ad network and discuss any concerns or issues that you may have. If the ad network is unwilling or unable to provide satisfactory explanations for discrepancies in their reporting, consider pausing your advertising activities with them.

Conclusion

In conclusion, while ad networks are an essential component of the digital advertising ecosystem, it is crucial for advertisers to be vigilant and proactive in verifying the data they receive. By taking the necessary steps to ensure the integrity of your advertising campaigns, you can protect your advertising investment and maintain the trust of your target audience.