Ad fraud, the practice of generating profits by fraudulently converting, clicking, viewing, or generating online interactions, is a challenge not only for the advertisers affected, but also for the entire online marketing industry. This is because ad fraud undermines the trust in online marketing campaigns, thus incentivizing advertisers to focus on other marketing channels.
Fraudulent traffic amounts to about 25 percent of all online advertising traffic. Ad fraud puts companies in legal risk as, if they contact fraudulent leads, they may violate the Telephone Consumer Protection Act (TCPA). Furthermore, since fraudsters often use stolen credit cards to complete credit card authorization forms, companies may be subject to chargebacks.
To avoid legal issues and mitigate the financial impact of ad fraud, advertisers need to take measures to identify it as soon as possible. Every day of delay may prove costly. User metrics, machine learning, and ad fraud solutions are three terms that need to be understood by anyone involved with reducing ad fraud. Each of those terms will be examined in more detail below.
3 Ways To Identify Ad Fraud
1. User Metrics
Companies can uncover ad fraud by using various user metrics, including click-to-convert rates, bounce rates, and traffic data.
High number of clicks with little conversations may indicate possible ad fraud. The reason is that crooks usually are not interested in becoming customers. All they need is to generate impressions, clicks, or transactions that look real enough so they will be paid by the affected companies.
Simply put, a bounce is a situation where a person visits the homepage of a website, without visiting any other pages on that website. High bounce rates may indicate ad fraud as fraudsters often just click on advertisements, without browsing through the websites to which those advertisements lead.
However, more advanced ad fraud will browse the site and even submit fake transactions such as fill in lead forms or make a purchase with stolen information to help hide their tracks.
High bounce rates may not always be caused by ad fraud. For example, if a website does not catch the attention of its visitors or lacks the information needed by those users, most visitors may prefer to leave it immediately after seeing its homepage. However, when high bounce rates are unusually high compared to other traffic sources, it is likely that there is ad fraud involved.
Traffic Data Signaling Ad Fraud
- Traffic coming from less known devices and outdated browsers
- Traffic coming from random countries that are not targeted by the advertiser concerned
- Multiple clicks coming from identical IP addresses.
To identify ad fraud, advertisers are advised to use the three metrics above, click-to-convert rates, bounce rates, and traffic data, and disregard vanity metrics like view-ability or NHT (None Human Traffic), as these metrics can be easily faked.
2. Machine Learning
Machine learning allows companies to identify ad fraud quickly and cost-efficiently. Machine learning applications enable companies to study large volumes of data by using the user metrics and the rates mentioned earlier and present the findings to ad fraud specialists.
Like any ad fraud specialist, they can then use the data to determine the presence of ad fraud and take the necessary measures. But with machine learning, those charged with defining and eliminating ad fraud for large enterprises with can leverage the scale of the technology to make it more effective across more streams of traffic.
Machine learning improves with more data. The more data you have the more effective it can be.
3. Ad Fraud Solutions
Ad fraud solutions like Anura use machine learning combined with industry experts to allow companies to easily distinguish, with a high level of accuracy and in real time, between real website visitors and fraudsters.
To make such a distinction, ad fraud solutions validate lead conversations by using user metrics, without relying on ineffective vanity metrics like view-ability and NHT. Anura collects hundreds of points of data on users to distinguish between imposters and legitimate users.
The information provided by ad fraud solutions enables marketers to take measures against ad fraud and make business decisions based on the actual behavior of their potential customers, rather than the behavior of malicious software or fraudsters.
The measures may include, for example, stopping the use of advertising channels that are most affected by ad fraud and focusing on channels that produce real results. Such measures will reduce chargebacks due to transactions with stolen credit cards and improve campaign performance.