Ad Fraud—The Devil of Marketing Budgets in the Advertising Industry


By: Lakshmi Jayakumar -- Principal Analyst, Indirect Services

30 July, 2019

Ad Fraud—The Devil of Marketing Budgets in the Advertising Industry


Abstract/Business Case

1    Introduction

An estimated US$16.4 billion of advertising expenses was lost due to ad fraud in 2017. This is being greatly driven by a rise in digital marketing spend. It has been observed that over 64 percent of all video ads played on the Internet are fraudulent. Close to 10 percent of the total digital marketing budget is lost to fraudulent buys that can be avoided.

2    Main Points

A brief description of the different types of ad fraud that impacts advertisers. Insights on how ad fraud impacts the ad budgets of companies.

3    Recommendations

Guidelines to avoid major impacts of ad fraud and protect client interests.

Key Issues

Ad fraud was not such a huge issue in the era of traditional advertising channels like TV, radio, and so on. Then, marketers could access only a limited target audience, yet had 100 percent control over their campaigns, which reduced the chances of fraud. However, the digital era has shifted the advertising landscape to where the marketer is no longer the key controller of the campaign. Campaigns on digital media can be impacted in many ways, which is sometimes out of the control of brand owners.

Ad fraud is one major challenge that is currently faced by marketers. This has a negative impact on campaigns as segmenting real customer activities from fraud become a huge constraint while defining upcoming campaigns. Programmatic technology is one of the most preferred automated technologies used by marketers for promotions as well as purchase and sale of various product lines due to its ease of obtaining buyer behaviour and campaign preferences. However, fraudulent users can act as a barrier to this since fake website visits, impressions, and clicks can change the results of the analytic tools used to measure campaign impact. This results in marketers diverting their budgets to the wrong campaign.

Identifying and mitigating ad fraud may become a challenge, but understanding the various types and models of this could help to minimise the impact of these so-called bots. This will help marketers recognise their signs and activities and improve overall business outcomes, so marketers can distinguish fraudulent activities from real customers’ activities and allocate their marketing budgets effectively.


1. How Does Ad Fraud Work?

Marketers create ads to impact their end consumers. However, when such ads do not reach the end consumer but instead appear on fake sites, it becomes ad fraud.

Ad frauds are categorised based on four major types of ad spend on digital media:

  1. Impressions—CPM (cost per thousand)
  2. Clicks—CPC (cost per click)
  3. Lead generation—CPL (cost per lead)
  4. Affiliated Revenue Share—CPA (cost per acquisition)

Of all the main types of fraud, most occur through impressions and clicks, which has led to the creation of fake websites as well as fake users.

Key Ingredients of Fraud:

Fake websites

Fake users

  • Also called cash-out sites
  • Extract ad dollars through fake impressions and clicks
  • Non-human traffic or invalid traffic on the website
  • High scope for cybercrime with revenue generation through fraudulent activity
  • Also known as bots
  • Made from developer tools used to test websites and apps before launch
  • Used for illegal ad impressions, fake mouse movement, page scrolling, clicks, and so on
  • Not listed, so detection is challenging


1.1 Types of Fraud Bots

There are three major classifications of bots based on the level of sophisticated behaviour.


Developer Tool


  • Make webpages load
  • Types—server-side scripts and browser-side scripts
  • Revenue generation model—Programmed for repeated webpage loading, creating fake exchanges, and generating impressions
  • Headless browsers
  • Spawn into data centres in millions
  • Revenue generation model—Perform repeated tasks like page loading, generating impressions, and clicks
  • They run at root on devices
  • Copy, record, and playback real human actions
  • Appear from diverse IP addresses, increasing difficulty to detect

These bots are so well programmed that such impressions, clicks, and traffic are difficult to detect on the analytic platforms of ad agencies and marketers, which impacts their marketing spend.

1.2 Role of Mobile Phones in Ad frauds

Mobile phones are one of the most commonly used devices in which marketers heavily invest due to the increasing time consumers spend using them. They also present the ability to offer customised ads based on consumers’ history and live location.

However, this environment has also been tapped by bots to commit ad frauds by making use of fake mobile devices and simulators made at data centres. These simulator apps are capable of launching apps, creating fake location data, as well as interacting.

However, software development kits can lower the rate of fraud in the mobile devices where good and bad apps can be easily identified.


2. Role of Marketers/Ad Agencies to Track Ad Frauds

Although it is difficult to mitigate the activities of bots, marketers and ad agencies can monitor them to some extent using sophisticated analytic tools to understand these fraudulent activities. Consequently, they can invest their budgets in campaigns where only genuine human traffic is detected.

  1. Identify patterns: Fraudsters often use programmed bots that generate clicks and impressions in a certain pattern as opposed to that of humans. These patterns can appear periodically, which could easily be detected.
  2. Level of consistency: Bots often appear to be consistent in their bounce rate, time, pages per visit, and so on. The metrices may be too high or too low, which does not occur in a normal scenario when we are looking at real human activities.

2.1 Usage of Technology to Detect Bots

Fraud can only be detected using advanced technology that can continuously monitor, track patterns and URLs from which the behaviour was observed, impressions, and clicks from bots for different campaigns. Such technology helps in tracking the real impact that campaigns have on consumers, differentiating it from bot activities.

There are three major areas where fraud detection can be performed to ensure better results:

three major areas where fraud detection can be performed


2.2 Action Plan for Effective Marketing Budget Allocation

Although mitigating the negative impact may be a huge challenge considering how wide the digital platform is, marketers could look at investing in big data methods that would help them to reduce outcomes of fake visits, clicks, and impressions. This would lead to drastic alterations in the analytic reports generated to define campaign success.

Some of the action plans that marketers could focus on are:

  1. Identify the scale by identifying registered domains and fraudulent websites to compare the activity on the campaign.
  2. Identify abnormal clustering and anomalies, for example, constant traffic from a specific location or different sets of IP addresses.
  3. Employ analytic professionals to identify past behaviour of campaigns and segment these to understand lookalike behaviour; map it with shopping and campaign after effects.
  4. Ensure that all campaigns are segmented and constantly monitored using analytic tools to capture any kind of behavioural change.

These steps may not mitigate the overall impact that ad frauds have on the campaigns, but will minimise fraudulent behaviour to a huge extent to ensure that marketers have more unbiased results regarding the campaign. This will help them plan where to invest their money with the support of the advertising and channel partners.

Conclusion—Analyst View

Ad fraud is not something that can be completely mitigated by marketers due to increasing technology that is empowering fraudsters to become stronger in the digital world. Marketers need to prioritise their brands and ensure better consumer experience to ensure that there is less negative impact created by the growing number of ad frauds.

A key approach should be to invest in the right technology, such as the use of marketing tools and analytics that can identify and distinguish fraudulent activities from that of human behaviour. Marketers should also work with a team of experts who can monitor the variations in the results and identify fraudulent activities, which will show the ad agency and marketer genuine activities as opposed to those of bots. Investing in simple marketing tactics such as feeds will also let them view user responses, which will help in identifying fraudulent behaviour on the platform.

Although ad fraud is only going to increase in the coming years, taking actionable measures that can reduce the impact will help buyers understand how to use their marketing budgets effectively, thus strengthening brand value and shaping a positive image in their end consumers’ minds.





COVID-19: Assess impact on your suppliers and ensure business continuity with Beroe’s WIRE
(World Instant Risk Exposure)