By Olga Nalivaiko
28. September 2021

The ABC of Campaign Metrics - Part 3: Ad Fraud

The ABC of Campaign Metrics - Part 3: Ad Fraud

As advertising technologies evolve, so do fraud mechanisms. The higher the degree of digitization, the more potential for digital fraud emerges. With more and more ads traded programmatically and fraud detection processes not yet fully developed, a sizable share of advertising ends up being served to bots instead of real users.

Definition of Ad Fraud
As defined by Integral Ad Science (IAS), Ad Fraud is “any impressions resulting from a deliberate activity that prevents the proper delivery of ads to real people, at the right time, in the right place — resulting in financial or opportunity loss by the advertiser and/or publisher in a particular transaction”. To put it simply, Ad Fraud means advertisers are paying for impressions and clicks that never happened.

Table of content:

What types of ad fraud are there?

There are several distinct kinds of ad fraud. Let’s take a closer look at the most common ones.

  • Bot traffic: This is the number one source of ad fraud. Bots are automated systems set up to click or interact with sites. ad fraud using bots is typically referred to as Click Fraud.
  • Click farms: This is a non-automated form of ad fraud. Real people are hired to click on ads. Advertisers end up paying for clicks that will not bring any transactions or brand benefits. Another variation is when click farm agents download an app and interact with it multiple times, thus distorting the advertising results.
  • Ad and click hijacking: The ad that should be shown to the user is replaced by other content leading to the hacker’s website or the user clicks on one ad but his click is redirected to a completely different ad result. This type of ad fraud is performed by malware. To make it work, the attacker has to compromise the user’s computer, the ad publisher’s website, or a proxy ad server.
  • Hidden ads and ad stacking: Websites can display ads in a way not viewable to users, for example placing them into a 1 x 1 pixel square, displaying them outside of the viewable area or stacking them on top of each other. In this case the owner of the website will get a payout for the displayed ad but the advertiser will not get incremental traffic or response.
  • Popunders: Forbidden by most advertising networks, popunders are similar to pop-up windows with ads except that the ad window appears behind the main web browser window, rather than in front.
  • Domain spoofing: In this case advertisers are tricked to believe that their ads are being served on a premium website while they actually end up on a website of inferior quality.

What is the current situation with ad fraud?

According to Statista, in 2020 the worldwide economic losses due to advertising fraud amounted to USD 35 billion. The size of the global digital advertising market was estimated at USD 333 billion, which means the share of ad fraud is more than 10% of the total. The United States suffered the largest losses, followed by China and the United Kingdom. In Germany, the estimated figure in 2020 is USD 0.73 billion. Moreover, according to Statista’s estimates, the annual global losses from ad fraud will keep growing and reach USD 44 billion in 2022.

IAS’s Media Quality report for H2 2021 states that the average global ad fraud level for video content was 0.8% for optimized campaigns (but 5.1% for non-optimized ones) for desktop and 0.3% (6.9%) for the mobile web. The figures for display ads were 0.8% (8.6%) and 0.4% (7.8%) respectively. In Germany, the ad fraud values for video content in optimized campaigns were 1.0% for desktop and 0.5% for mobile web, which is close to the global levels. The same is true for display ads with 0.8% for desktop and 0.5% for mobile web in Germany.

MMT Ad fraud globally and Germany - IAS report 2021

How to detect ad fraud?

Detecting ad fraud has certainly been a big challenge for both advertisers and publishers. There are several specialized providers in the market who use proprietary algorithms based on machine learning methods to identify fraudulent behavior, such as abnormally high CTRs and poor or no campaign performance. Relevant URLs and IP addresses are added to a blacklist. Most popular among those providers include Integral Ad Science (IAS), DoubleVerify, Moat, Fraudlogix, and Pixalate.

What are the trends in ad fraud?

According to the Global Insights Report 2021 by DoubleVerify, the violation rates related to post-bid fraud and sophisticated invalid traffic (SIVT) were down 30% year-over-year in 2020, from 2.0 to 1.4%. The increased adoption of pre-bid avoidance solutions and platform certifications help identify and filter fraud and SIVT before it becomes an issue for advertisers.

"The margin-chasing of some of the intermediaries which just characterized programmatic advertising in the past – that time’s over. There’s a macro social debate, there’s a regulatory debate and there’s now also advertisers on the basis of this, voting with their money, meaning that you need to stand out as a clean vendor in this space to continue to attract budgets"


Daniel Knapp
Chief Economist, IAB Europe

Webinar “Driving Quality and Performance in Digital Media, EMEA 2021” (03.08.2021)

What can be done to prevent ad fraud?

There are several guidelines you can follow to minimize the ad fraud rates for your campaigns.

  • Choose trustworthy partners. Make sure your advertising partners and publishers use ad fraud prevention mechanisms and approaches.
  • Monitor your campaigns. Keep an eye on your campaign KPIs and traffic sources to identify any suspicious activity, such as a sudden surge in click-through rate without incremental conversions or multiple visits from the same IP.
  • Use blacklists. Suspicious domains and IPs can be added to blacklists and blocked. This is part of the standard solutions provided by ad fraud prevention partners.
  • Apply ad fraud protection solutions. They fight against bot traffic, click farms, and multiple clicks on ads from suspicious IP addresses. Examples include Perform, AppsFlyer, ADEX, Adjust, ClickCease, Intercepted, and Kochava.

Sources:

Receive 4 times a year a summary of our articles most relevant to you.

Subscribe to our newsletter


© Copyright 2021 | Mercury Media Technology GmbH