Combatting Invalid Traffic: How to Safeguard Your Digital Marketing Efforts?

Invalid Traffic, or IVT, refers to any activity that doesn’t come from a real user with a genuine interest in a website’s content, products, or services. IVT can distort analytics, waste advertising budgets, and negatively impact the digital advertising ecosystem. Bots, automated scripts, or human fraudsters typically generate it.

According to IAB’s Europe Guide to Ad Fraud guide, there are two most common categories of fraudulent IVT:

• General Invalid Traffic (GIVT) includes simpler forms of invalid traffic that can often be detected using standard industry tools and filters. This category encompasses obvious and easily identifiable types of invalid traffic.
• Sophisticated Invalid Traffic (SIVT) represents more complex and harder-to-detect forms of invalid traffic. It involves techniques designed to mimic legitimate user behavior and evade detection by standard filters.

Let’s take a look at some statistics:

• According to a leading scan traffic provider, “Masked IP” was Apple’s second most common type of Invalid Traffic (IVT), accounting for 25% of invalid Apple app traffic in Q1 2024. Similarly, “Defased” was Google’s second most common type of IVT, making up 17% of Google’s app IVT in the same period.
• Research shows that the “Masked IP” invalid traffic type impacted the most apps in the Google Play Store, with 72,000 instances in Q1 2024. The “Masked IP” invalid traffic type in the Apple App Store affected 15,000 apps, making it the most prevalent IVT in the APAC region during the same quarter.

Invalid Traffic Types

DecenterAds platform collaborates closely with prominent traffic scanner vendors, utilizing their standardized classification systems to identify various types of invalid traffic. These classifications help us effectively detect and manage fraudulent activities. For instance, app spoofing is one type of sophisticated invalid traffic we monitor. This occurs when the app identifier, such as the bundle ID reported to the exchange, doesn’t align with the app’s characteristics identified by a fraud detection provider.

Another SIVT type is click farm activity, where impressions come from users flagged as part of human click farms. This fraud involves groups of people hired to click on ads to inflate engagement metrics artificially. Similarly, cookie stuffing represents another SIVT category, characterized by activity from cookies that connect to the internet via an unusually high number of different IP addresses.

We also track display impression fraud, a sophisticated method where impressions are generated at a statistically significant and inflated rate from the same browser or device. This type of fraud skews metrics by creating an artificial volume of ad impressions.

General invalid traffic (GIVT) includes several types of non-human traffic. Autoreloaders, for example, generate impressions with periodic patterns, something humans can’t replicate. Fast clickers are another GIVT type, where clicks originate from users who generate clicks less than one second after the impression is served.

Additionally, GIVT includes activity from IAB dummy bots, which use user agent strings that do not correspond to any known browser, making their traffic easily identifiable as fraudulent. Private IP traffic is also classified as GIVT when ad impressions originate from IP addresses within private network spaces rather than from public internet addresses.

By leveraging the expertise of top-tier ad and traffic checkers, DecenterAds platform ensures a robust defense against these varied forms of invalid traffic, maintaining the integrity and reliability of our advertising ecosystem.

Ad Fraud Detection Methods

To fight fraud effectively, you need to detect it efficiently. Each type of fraud requires a different approach. Let’s examine the approaches used by DecentrAds team.

Site/App Fraud (IVT)

Using automated algorithms and heuristic models to detect fraudulent sites and apps. Data scientists review multiple data points, publisher records, privacy policies, and other documentation to confirm fraud, assigning the IVT attribute as needed.

CTV Fraud

Involves collaboration among buyers, sellers, and platforms, with data scientists analyzing hundreds of data points per impression to detect anomalies and eliminate fraudulent traffic.

Adware/Malware

Utilizes real-time signal processing, machine learning, and fraud analysis to identify and block adware/malware impressions across browsers.

GIVT

Uses industry-supplied lists to identify bots, spiders, fraudulent sites, unknown browsers, and known data centers.

Non-Human Data Center Traffic

This case compares user IP addresses against a list of non-human traffic sources maintained by third-party and proprietary algorithms.

Bot Fraud

A deterministic approach for fast and accurate detection is better, enhancing business outcomes by eliminating fraudulent inventory. On the other hand, probabilistic tracking is less precise and can lead to suboptimal decisions. Speed and definitive identification are crucial as new bots are most active in their first 24 hours.

To Sum Up

Invalid Traffic (IVT) is a critical issue in digital marketing. It is created by bots, scripts, or human fraudsters, leading to distorted analytics and wasted budgets. It is categorized into General Invalid Traffic and Sophisticated Invalid Traffic. “Masked IP” was a prevalent IVT type affecting many apps in Q1 2024, so paying attention is critical.

Since IVT poses a severe problem in digital marketing, various strategies exist to detect and mitigate it. Contact our experts and schedule a call to learn how the DecenterAds platform safeguards traffic.