Fighting Online Ad Fraud With More Sophisticated Metrics

While most marketers are familiar with the concept of online advertising fraud, the conversation spiked recently with an article in The Atlantic, “Welcome to the Internet of Thingies: 61.5% of Web Traffic Is Not Human.” The piece details how, for the first time, bot traffic has eclipsed human traffic, accounting for 61.5% of all website visitors.  While this fraudulent traffic may be considered a mere nuisance to most marketers, what they might not realize is how much it’s actually costing them.  According to a recent article in Adweek, the cost is estimated to be on the order of $6 billion annually.

An Industry Epidemic

Online fraud affects every stage of the purchase funnel, from clicks to impressions to conversions.  Click fraud is the most prevalent, occurring when a person, computer program, or bot imitates a legitimate user clicking on a link or an ad, for the purpose of generating a charge per click. These fraudulent clicks cost advertisers money, without producing any sales, signups or other conversions.

Impression fraud occurs when page views or online ads are delivered and counted as being served without having a chance to be seen by a human. Some of the ads are invisible; others are layered behind other ads. But some of the largest fraudsters use bots that take over computers and simply render ads on pages while the computer is inactive or when the user is not in front of the computer.

Interestingly, these ads meet all of the current standards for ad viewability, which determine whether ad content is actually visible on the browser, and are almost always counted as genuine ads by fraud prevention systems. As a result, advertisers that pay for media on a CPM basis end up paying for ad impressions that are never seen by users. More rare, but still significant is conversion fraud, which occurs when a lead is produced from a non-human source, such as when a fake online form is submitted.

Fighting Back

There are a number of companies actively trying to solve the problem of advertising fraud. For example, most of the major search engines have filters that can detect fraud activity, such as repetitive clicks coming from the same IP address or single geographic area, and various other click patterns. Companies like Spider.io, White Ops and Integral Ad Science are using bot detection algorithms to detect and mitigate suspect traffic. Other organizations, such as Nielsen and the IAB, are working to develop an industry-wide standard of what constitutes a viewable impression.

While all the players that have brought innovations to market to combat online advertising fraud should be applauded, it remains an ongoing problem that’s only getting worse. As advertisers, publishers and ad networks are getting smarter, so too are the attackers, and their malware and methods are only becoming more sophisticated.

A Proposal

Given the complexity and severity of the problem, the best – and perhaps only – way to neutralize the threat once and for all is to change the way advertising effectiveness is measured and compensated. Today, much of the advertising industry still reliies on metrics such as CPM and CPC, which are particularly susceptible and easy mimicked by malicious bots and fraudsters.

Instead, advertisers should focus on metrics that impact their bottom line, like revenue. Fraudsters can generate bots that simulate a human viewing a page, clicking on an ad, and even filling out a lead form requesting more information about a product or service. But a bot can’t generate real conversions like completing a purchase with a payment, buying a subscription, or taking other actions that can only be generated by real humans. Only when marketers use more down-funnel metrics like revenue to measure advertising effectiveness can they ensure that real humans are interacting with their brand and their marketing dollars aren’t being wasted.

In today’s digital advertising ecosystem, everyone is susceptible to fraud. But with its challenges comes opportunity. Change will come when advertisers stop pursuing impressions and clicks, and instead focus on the metrics that matter to the bottom line of their business.

Tags: metrics
Recommend (42)
3 comments about "Fighting Online Ad Fraud With More Sophisticated Metrics".
  1. Neil Mahoney from Mahoney/Marketing , February 25, 2014 at 3:55 p.m.
    It has long mystified and amused me how gullible so-called media mavens can be. Clicks are not worth measuring. A click is nothing more than the equivalent of a magazine reader pausing to note your ad, then turning the page. The only thing worth measuring is downloads, which pre-internet was a prospect requesting more information about your product or service. Neil Mahoney
  2. Doug Weeks from Jumpcurve Online , February 25, 2014 at 6:07 p.m.
    The problem with your solution is that you are saying display advertising is only for low funnel DR campaigns and I think that is limiting it should be able to function at all parts of funnel. I guess there are metric for brand as well, but harder to interpret. Also its still does not stop people paying for fraud and frankly publisher getting less revenue because display does not work as well as other channels. I have a solution what if some of the big DSP players like Google started treating the impression fraud like they treat click fraud on Adwords and issue credits when they detect it. They own spider.IO now they can do it. Problem won't get fixed while advertisers pay for a product that does not exist and the guy selling the product pockets the money without any impact to the bottom-line. Some part of the $6 billion in fraud is revenue for U.S. based public companies at some point if industry player don't address this lawyers will sniff it out and we all know class action awards don't trickle very far down certainly not back to advertisers.
  3. Anto Chittilappilly from Visual IQ , March 23, 2014 at 2:56 p.m.
    Doug, if the advertiser and their agencies optimize their media based on the bottom-line metrics, then fraud-media will automatically get filtered out because the fraud-media can’t contribute to the advertiser’s bottom-line. Once the money stops coming, the need to filter and remove fraud will trickle down all the way to whoever pays to fraudsters. That’s the most powerful weapon against fraud.