Commentary

Opening A Can Of Bots - Defining Digital Fraud

Much has been written on digital fraud in recent months, but many industry players choose to remain in the dark. The problem is “invisible” -- why open a can of worms (or bots in this case) when you don't have to? But fraud is a problem that can't be ignored. 

Ad fraud takes place when a bot (online robot) poses as a human and either clicks on an ad or is served an ad impression. By serving an ad to a non-human or by having an ad clicked by one, advertising dollars are going to waste. 

Why does it take place? Because fraudsters have the opportunity to make lots of money. Botnet operators monetize networks they have a financial stake in by setting up ad-supported Web sites and driving bot traffic to them, increasing the value of the ad slots on their pages, as well as making money off the CPMs and CPCs from these bots. They can also supply fraudulent bot traffic to publishers that buy traffic in an effort to inorganically extend their audience and attract advertisers.

On top of this, premium sites unwittingly attract fraudulent traffic without any way to stop it. Bots are programmed to visit legitimate sites and pick up cookies that marketers find valuable. These bots are then targeted on expensive ad slots, such as in-banner video ads. In addition, they use pixel stuffing (stuffing an entire site into a 1x1 pixel) or ad stacking (placing multiple ads on top of each other in a single ad placement) to maximize their yield.

No matter how reputable the publisher or advertiser, everyone is subjected to fraud. As long as media is sold based on views (CPM) or action taken (CPC/CPA), there will be fraud. 

Not all non-human traffic is harmful. For example, Web crawlers are legal bots that systematically browse the Web for indexing and data collection. Web crawlers do not pose as human; therefore, they are not served ads, and they do not impact advertising campaigns. 

Other bad practices in online advertising are also often mistaken for fraud. These are ads that are legitimately served to humans yet are inefficient for advertisers, for reasons including poor viewability, ad collision, ads on non-brand-safe sites, auto-play/auto-reload videos, etc. 

Fraud, however, is an ongoing problem, and it’s getting worse. The bad guys are constantly updating their technology to try and outsmart the anti-fraud solutions. Today, 20 percent of the impressions on exchanges are fraudulent. In addition, 2 percent of direct impressions and 20 percent of network impressions are all fraud.

Now that we’ve opened the “can of bots,” we clearly see that fraud can’t be ignored. Luckily, there are different options for both advertisers and publishers to protect themselves.

Blacklisting has been the norm in the industry for a while. When bot traffic is detected, a site is added to a blacklist and remains on that list until it is updated again. There are several problems with this method. First, bots are more dynamic and sophisticated than the lists. They are a target in motion, and will just create new sites with non-human traffic. Also, as any legitimate site can unknowingly have contaminated traffic, blacklisting a whole site wipes out all the good impressions along with the bad, which hurts the advertiser's ability to scale while also hurting the publisher’s ability to monetize. 

The promise of fraud-free online advertising now lies with technology that has improved the ability to detect and even block fraud at the impression level. Industry solutions now go beyond the blacklist and analyze each impression for fraud signals along with dynamic historical data. Using this method, it is no longer necessary to wipe out sites altogether. Instead, fraudulent impressions can be “surgically” removed, allowing publishers to monetize their legitimate traffic, while protecting advertisers’ budgets from being wasted on bot traffic. 

It is an ongoing arms race between those who are trying to exploit the digital ad world and those who are trying to protect the victims of this war, and both sides are using technology to their advantage. As long as there is money to be made, this will go on. How are you protecting yourself?

2 comments about "Opening A Can Of Bots - Defining Digital Fraud".
Check to receive email when comments are posted.
  1. Jean Renard from TRM Inc., November 15, 2013 at 1:58 p.m.

    fraud is eliminated when the only metric that matters is ROI. Performance based results are the only thing that advertisers needs to look at to determine the effectiveness of a strategy. It is about time that we return to basics.

  2. Augustine Fou from FouAnalytics, October 17, 2014 at 6:40 a.m.

    However, many companies still define ROI as "more ad impressions for lower average CPM" which continues to incentivize suboptimal behavior (such as buying super cheap ad inventory to mix in, in order to increase the number of impressions while lowering the average cost). The definition of ROI needs to be updated to use meaningful business metrics that are based on actions in the purchase funnel.

Next story loading loading..