DorisRon
Member
Here is a small guide for marketers on bots, fake web domains and other "dirty tricks" in online advertising.
Active automation has made the process of procurement and sale of advertising more informative and effective. But a fly is found in the ointment: the automation served as fertile ground for fraud.
In the end, if the machines enter into transactions which previously were agreed by people on the phone, it is fraught with consequences. The less human control is, the more Internet fraud we will see.
Ad teams, which are dazzled with the opportunity to buy a lot of traffic at a very low price, sometimes forget about the need to always be careful. In turn, the publishers are seeking to expand the audience, to ensure revenue from advertising. When they make decisions based on indicators of their traffic within their digital domains, they often collaborate with third parties, many of which are commonplace crooks that generate bot traffic.
Yes, the "non-human traffic" raises KPIs of impressions and views, but it significantly reduces the quality of the ads, because bot traffic does not lead to conversion. By reducing the quality of advertising, publishers have a need for increasing traffic - so there is a vicious circle in which profits are got only by crooks.
What is advertising fraud?
In simple terms, this is inhuman traffic or bots that generate advertising impressions or clicks on advertisements.
However, it should be noted that any official working definition of fraud associated with advertising, does not exist. And if you ask 10 different people, you can get 10 different answers.
Why is there mess around the definition of fraud?
Many experts disagree whether the substandard advertising placement is fraud. For example, video ads on display in the top half of the visible part of the site screen with mute sound 1x1 iframe, or banners superimposed on each other like cards. In these cases, there is no chance that the user will see the ads.
Therefore, some industry players may argue that it's just a bad placement. Others believe that this is the clearest example of fraud.
However, the important thing is that neither in the case of outright fraud or poor quality when placing, the advertiser does not get what he paid for. Obviously, this is cheating. And cheating is synonymous with fraud.
How serious can an advertising fraud problem be considered?
It depends on whom you ask. Pessimistic statistics show that about 40% of all ad impressions are generated by fraudsters.
However, many experts say that these figures are propaganda of suppliers of technologies designed to combat fraud.
As expected, companies that specialize in the development of high-tech advertising software products, talk about the high rates of falsification of online advertising. On the other hand, ad exchanges and ad networks say that the seriousness of the problem of fraud in the advertising industry is exaggerated.
In December last year, Association of National Advertisers (ANA) in association with White Ops, a provider of solutions to combat fraud, conducted a study on the degree of harm caused by industry robots. The study analyzed more than 180 campaigns of ANA members, including Walmart, Johnson & Johnson and Kimberly-Clark. As a result, the experts concluded that 23% of video displays and 11% of banner impressions are falsified. This means that the industry in 2016 carried the multibillion-dollar losses as a result of actions of crooks.
What are robots?
Bots are programs designed to create false impressions or generate hidden, unseen ads. At the same time, these programs aim to be left unnoticed.
Bots often take control of a user's computer and run in the background. A large number of computers infected with bots, is called a botnet.
The emergence of bots reflects the growing sophistication of fraudsters. A few years ago advertising fraud was relatively simple (only individual ad impressions were falsified), today bots allow deceivers to exercise sells on a large scale and in absolute silence.
Bots can do more than just forge impressions. Some do their dirty work behind the scenes creating such ads that none of the users can see - in this case the advertiser pays for the "real" impressions of their advertising. Other robots "arrive" on the website to simply "eat cookies", which would allow them to redirect yourself to another place.
In addition, every day bots better mimic the user's behavior after he allegedly clicked on the advertising banner. For example, the robot can go to the advertiser's website and view the video or fill out a form for lead generation, and sometimes even make "a purchase" by sending advertiser the appropriate code for conversion tracking.
Such activity is enough to fool the advertiser, who will live with the idea that his campaign gives a result. Further - more, the advertiser will increase the cost of optimizing their display campaigns, believing in real consumer interest in his products or services. In fact, he will only increase the funding of the scam.
Another way through which the botnet can be harmful to advertisers is "laundered" impressions. Bots clone behavior of real people whose computers are infected. The behavior of the robots in this case, is virtually indistinguishable from the behavior of the person viewing the ad, because the anti-fraud advertising system is powerless. And that means only one thing - these bots can get into the "white list" of advertisers, and their actions will be identified as the actions of real consumers.
Why is it so difficult to detect fraud?
In the process of advertising buying there are several points which are extremely susceptible to fraud. In addition, numerous persons are involved in this process, and some of them contribute to the fact that sales are not made transparently and honestly. Due to the fact that the counterfeit traffic passes through many intermediaries, it is extremely difficult to trace the money, which the botnet operator eventually gets.
Additionally, botnets have varying degrees of complexity. Some of them are easy to find, and some are virtually impossible. Finding and screening of scams are certainly cool, but do not get false sense of security, because more complex botnets are always in the shade.
According to the notorious Roger "Chatterbox" Kint in the film "The Usual Suspects": "The greatest trick of the devil is to convince the world that he does not exist."
If a danger threatens premium sites?
Yes. Modern fraud is no longer characterized with low quality. Even those advertisers who buy premium inventory, may suffer from the actions of bots.
For example, large publishers that use such form of business cooperation, as the affiliate program, often purchase traffic from third party which is just teeming with bots. Since these transactions are aimed at increasing the volume, publishers unwittingly open the door to the boat traffic. Advertisers using last-touch attribution, may be particularly susceptible to this type of fraud.
Are there other kinds of advertising fraud?
Naturally, though bots are considered to be the most common form of it.
▪ "Laundered" impressions that are redirected to sites with a large volume of traffic through web sites with fake content or pirate sites, which inventory isn't bought by the majority of advertisers.
▪ Advertising heap. Publishers may place several ads one on another, so that only the top one is actually in the field of view. In addition, a publisher can hold multiple ads within a single pixel and ads will be identified as viewed, even if they had not been seen.
▪ "Filling of cookies." Web sites use cookies to legally keep track of users activity. If a site visitor completes an action, such as a purchase or filling out a form, the site receives money under the terms of the affiliate program. However, fraudsters often use falsified cookies, to receive a commission.
▪ Fake websites are created solely for advertising purposes and does not provide any of the content that the audience wants to see. Many fake web sites are a part of a large network, and each collects a certain amount of commission, so as not to arouse suspicion. As a rule, such networks fraudulently earn millions of dollars.
▪ False domains. The scammers resort to high-end simulation of domains and popular sites. An advertiser may think that he buys a placement at premium publishers, but in fact he has to do with "Parody". This type of fraud is dangerous by the fact that it seems as if premium publishers have more inventory than they actually have. This greatly reduces their prices.
▪ Toolbars. It is a pretty sophisticated malware technology, which flourished in the 1990s and the beginning of zero. But they still quite actively create the illusion that advertisers are buying high quality inventory, although it is not so.
What about video ads?
Video is an extremely tasty morsel for fraud, as this format provides higher CPM, than banner ads or mobile.
Despite the fact that there are undeniable moments pointing to fraud (eg, clicks are usually generated by bots because people rarely click on video ads), fraud in this segment requires a higher level of programming knowledge. These programmers, whom experts call «badasses» of advertising fraud, create a more secretive, more modern and more sophisticated fraud mechanisms.
The video environment promotes fraud by itself. Simply put, the demand for video inventory among advertisers is much higher than the supply of publishers. The crooks who come to the market and fill that gap, use it.
What about mobile advertising?
While budgets for mobile advertising are trifling, the industry is not very worried about the fraud in this sphere. But investment in mobile is growing rapidly, and therefore impostors are activated.
Mobile environment has its own special problems, because the lack of cookies incredibly complicates the assessment of the results. In addition, a variety of mobile operating systems and applications face different technological challenges that are unique.
So if the developers of mobile applications and operating systems claim that they have not yet faced serious advertising fraud, it is possibly due to the fact that their instruments are not sufficiently matured to be interesting for romantics from the great online roads.
Best developers of ad fraud solutions
DoubleVerify, an integrated solution for fraud detection, is designed to verify the authenticity of the quality of the inventory.
Forensiq is the decision which does not focus on one aspect (bots), but analyzes the entire purchasing funnel of sales - from impressions and clicks to conversions.
Integral Ad Science is a technology, which inspects the order of 3.5 billion impressions per day and analyzes the behavioral model on the level of impressions to determine infection by bots. Also it mediates in disputes between advertisers and publishers.
White Ops is a technology designed for deep analytical "drilling", which allows to distinguish fake impressions from the real. The company mainly cooperates with ad exchanges and networks, which want to get rid of the fake web sites.
Active automation has made the process of procurement and sale of advertising more informative and effective. But a fly is found in the ointment: the automation served as fertile ground for fraud.
In the end, if the machines enter into transactions which previously were agreed by people on the phone, it is fraught with consequences. The less human control is, the more Internet fraud we will see.
Ad teams, which are dazzled with the opportunity to buy a lot of traffic at a very low price, sometimes forget about the need to always be careful. In turn, the publishers are seeking to expand the audience, to ensure revenue from advertising. When they make decisions based on indicators of their traffic within their digital domains, they often collaborate with third parties, many of which are commonplace crooks that generate bot traffic.
Yes, the "non-human traffic" raises KPIs of impressions and views, but it significantly reduces the quality of the ads, because bot traffic does not lead to conversion. By reducing the quality of advertising, publishers have a need for increasing traffic - so there is a vicious circle in which profits are got only by crooks.
What is advertising fraud?
In simple terms, this is inhuman traffic or bots that generate advertising impressions or clicks on advertisements.
However, it should be noted that any official working definition of fraud associated with advertising, does not exist. And if you ask 10 different people, you can get 10 different answers.
Why is there mess around the definition of fraud?
Many experts disagree whether the substandard advertising placement is fraud. For example, video ads on display in the top half of the visible part of the site screen with mute sound 1x1 iframe, or banners superimposed on each other like cards. In these cases, there is no chance that the user will see the ads.
Therefore, some industry players may argue that it's just a bad placement. Others believe that this is the clearest example of fraud.
However, the important thing is that neither in the case of outright fraud or poor quality when placing, the advertiser does not get what he paid for. Obviously, this is cheating. And cheating is synonymous with fraud.
How serious can an advertising fraud problem be considered?
It depends on whom you ask. Pessimistic statistics show that about 40% of all ad impressions are generated by fraudsters.
However, many experts say that these figures are propaganda of suppliers of technologies designed to combat fraud.
As expected, companies that specialize in the development of high-tech advertising software products, talk about the high rates of falsification of online advertising. On the other hand, ad exchanges and ad networks say that the seriousness of the problem of fraud in the advertising industry is exaggerated.
In December last year, Association of National Advertisers (ANA) in association with White Ops, a provider of solutions to combat fraud, conducted a study on the degree of harm caused by industry robots. The study analyzed more than 180 campaigns of ANA members, including Walmart, Johnson & Johnson and Kimberly-Clark. As a result, the experts concluded that 23% of video displays and 11% of banner impressions are falsified. This means that the industry in 2016 carried the multibillion-dollar losses as a result of actions of crooks.
What are robots?
Bots are programs designed to create false impressions or generate hidden, unseen ads. At the same time, these programs aim to be left unnoticed.
Bots often take control of a user's computer and run in the background. A large number of computers infected with bots, is called a botnet.
The emergence of bots reflects the growing sophistication of fraudsters. A few years ago advertising fraud was relatively simple (only individual ad impressions were falsified), today bots allow deceivers to exercise sells on a large scale and in absolute silence.
Bots can do more than just forge impressions. Some do their dirty work behind the scenes creating such ads that none of the users can see - in this case the advertiser pays for the "real" impressions of their advertising. Other robots "arrive" on the website to simply "eat cookies", which would allow them to redirect yourself to another place.
In addition, every day bots better mimic the user's behavior after he allegedly clicked on the advertising banner. For example, the robot can go to the advertiser's website and view the video or fill out a form for lead generation, and sometimes even make "a purchase" by sending advertiser the appropriate code for conversion tracking.
Such activity is enough to fool the advertiser, who will live with the idea that his campaign gives a result. Further - more, the advertiser will increase the cost of optimizing their display campaigns, believing in real consumer interest in his products or services. In fact, he will only increase the funding of the scam.
Another way through which the botnet can be harmful to advertisers is "laundered" impressions. Bots clone behavior of real people whose computers are infected. The behavior of the robots in this case, is virtually indistinguishable from the behavior of the person viewing the ad, because the anti-fraud advertising system is powerless. And that means only one thing - these bots can get into the "white list" of advertisers, and their actions will be identified as the actions of real consumers.
Why is it so difficult to detect fraud?
In the process of advertising buying there are several points which are extremely susceptible to fraud. In addition, numerous persons are involved in this process, and some of them contribute to the fact that sales are not made transparently and honestly. Due to the fact that the counterfeit traffic passes through many intermediaries, it is extremely difficult to trace the money, which the botnet operator eventually gets.
Additionally, botnets have varying degrees of complexity. Some of them are easy to find, and some are virtually impossible. Finding and screening of scams are certainly cool, but do not get false sense of security, because more complex botnets are always in the shade.
According to the notorious Roger "Chatterbox" Kint in the film "The Usual Suspects": "The greatest trick of the devil is to convince the world that he does not exist."
If a danger threatens premium sites?
Yes. Modern fraud is no longer characterized with low quality. Even those advertisers who buy premium inventory, may suffer from the actions of bots.
For example, large publishers that use such form of business cooperation, as the affiliate program, often purchase traffic from third party which is just teeming with bots. Since these transactions are aimed at increasing the volume, publishers unwittingly open the door to the boat traffic. Advertisers using last-touch attribution, may be particularly susceptible to this type of fraud.
Are there other kinds of advertising fraud?
Naturally, though bots are considered to be the most common form of it.
▪ "Laundered" impressions that are redirected to sites with a large volume of traffic through web sites with fake content or pirate sites, which inventory isn't bought by the majority of advertisers.
▪ Advertising heap. Publishers may place several ads one on another, so that only the top one is actually in the field of view. In addition, a publisher can hold multiple ads within a single pixel and ads will be identified as viewed, even if they had not been seen.
▪ "Filling of cookies." Web sites use cookies to legally keep track of users activity. If a site visitor completes an action, such as a purchase or filling out a form, the site receives money under the terms of the affiliate program. However, fraudsters often use falsified cookies, to receive a commission.
▪ Fake websites are created solely for advertising purposes and does not provide any of the content that the audience wants to see. Many fake web sites are a part of a large network, and each collects a certain amount of commission, so as not to arouse suspicion. As a rule, such networks fraudulently earn millions of dollars.
▪ False domains. The scammers resort to high-end simulation of domains and popular sites. An advertiser may think that he buys a placement at premium publishers, but in fact he has to do with "Parody". This type of fraud is dangerous by the fact that it seems as if premium publishers have more inventory than they actually have. This greatly reduces their prices.
▪ Toolbars. It is a pretty sophisticated malware technology, which flourished in the 1990s and the beginning of zero. But they still quite actively create the illusion that advertisers are buying high quality inventory, although it is not so.
What about video ads?
Video is an extremely tasty morsel for fraud, as this format provides higher CPM, than banner ads or mobile.
Despite the fact that there are undeniable moments pointing to fraud (eg, clicks are usually generated by bots because people rarely click on video ads), fraud in this segment requires a higher level of programming knowledge. These programmers, whom experts call «badasses» of advertising fraud, create a more secretive, more modern and more sophisticated fraud mechanisms.
The video environment promotes fraud by itself. Simply put, the demand for video inventory among advertisers is much higher than the supply of publishers. The crooks who come to the market and fill that gap, use it.
What about mobile advertising?
While budgets for mobile advertising are trifling, the industry is not very worried about the fraud in this sphere. But investment in mobile is growing rapidly, and therefore impostors are activated.
Mobile environment has its own special problems, because the lack of cookies incredibly complicates the assessment of the results. In addition, a variety of mobile operating systems and applications face different technological challenges that are unique.
So if the developers of mobile applications and operating systems claim that they have not yet faced serious advertising fraud, it is possibly due to the fact that their instruments are not sufficiently matured to be interesting for romantics from the great online roads.
Best developers of ad fraud solutions
DoubleVerify, an integrated solution for fraud detection, is designed to verify the authenticity of the quality of the inventory.
Forensiq is the decision which does not focus on one aspect (bots), but analyzes the entire purchasing funnel of sales - from impressions and clicks to conversions.
Integral Ad Science is a technology, which inspects the order of 3.5 billion impressions per day and analyzes the behavioral model on the level of impressions to determine infection by bots. Also it mediates in disputes between advertisers and publishers.
White Ops is a technology designed for deep analytical "drilling", which allows to distinguish fake impressions from the real. The company mainly cooperates with ad exchanges and networks, which want to get rid of the fake web sites.