Traffic Of Good Intent: A Learned Perspective On Quality Traffic

I recently attended the IAB Annual Leadership Conference, where one of the main topics was media fraud, a subject of long-time interest for me that’s been receiving renewed attention of late, including an op-ed by IAB CEO Randall Rothenberg.

Two things dismayed me: to hear that no one actually knows the size, or the sources, of the problem; and to see technology vendors pitching a “solution” to the problem. Nobody wants to admit where the problem is, nor how large it is, so maybe this is the dirty little secret where most everybody is making money.

From our experiences, fraudulent/phony/nonexistent traffic almost always occurs for one reason: money! That is, someone perpetrates this action to make money, either independently, in cooperation with other parties, or by taking advantage of someone/company or technology. Years ago, we saw venture capitalists being the objects of fraud, then search marketers/advertisers, and most currently display marketers/advertisers. Suffice it to say that fraudsters will always be looking for opportunities and exploit illegitimate ways to make money.



We have seen that, most often, the fraudster is a “publisher” who manufactures traffic in a number of ways. Bot nets are one of the more nefarious methods used, but there are other ways as well to creates a large volume of inventory for unsuspecting advertisers, inventory that will not be seen by people or will be mis-targeted to an unintended audience.

Our investigations have concluded that in all cases, the fraudsters have developed techniques to perpetrate the fraud or have taken advantage of internal control deficiencies. When controls are improved, the incidence of fraud decreases.

Fraudulent/illegitimate traffic can be detected in a number of ways:

1.     An advertiser/agency can review reporting and observe impression/click volumes that are disproportionate to expectations.

2.     An advertiser/agency can review attribution reports and observe low performing or non-performing advertising.

3.     A network/SSP can observe reporting that reveals disproportionate impression and dollar remits to partners (usually publishers).

4.     An advertiser/agency or network could run pattern analysis software to analyze impression logs for indications of abuse.

5.     An advertiser/agency or network could employ some form of fraud detection service/technology which could detect nefarious activity.

One of the problems we, as an industry, face is that even “innocent co-operators” make money from the fraud. As such, unknowing IAB members are profiting from fraud, and many may be unwilling to ask questions from fear of impacting their bottom line.

I think it is worthy to point out that the industry will never be able to completely eliminate outright fraud. Also, the upcoming viewability standards might slow down fraud for a time, but will not deter it in the long term. For instance, there exists technology today that will emulate a browser and the actions that users perform. Many of these emulation packages are perfectly legitimate.

That is not to say that nothing can be done. There are several ways IAB businesses can address the problem. First, the industry can work to eliminate the fraud by instituting a number of standard controls (business/operational/financial) that will deter or detect fraudsters:

1.     Publishers must initially vet, and then monitor, all traffic sources to ensure high-quality traffic, particularly those sources that are contracted and compensated for traffic increases.

2.     Publishers must employ internal controls to prevent (or monitor) employees who are compensated/rewarded for traffic increases.

3.     Advertisers/agencies/networks/SSPs must initially vet all business partners (publishers) before agreeing to do business with them, and then monitor them for unreasonable increases, poor-performing campaigns, or erratic traffic fluctuations. Discrepancies can then be investigated.

4.     Technology companies can be certified to IAB/MRC/AICPA standards to ensure that adequate systems and internal controls are in place. The IAB certification standards are for both measurement and quality assurance guidelines.

5.     Technology companies can employ technical and statistical tools, both contracted and internally developed, that can assist in uncovering fraudulent users and traffic sources.

6.     IAB can play a role by incorporating strict language into its Code of Conduct, and then enforcing such a code.

In my opinion, this problem can be considered a measurement issue, and should be addressed with IAB members embracing the current measurement standards, combined with best practices for internal control (MRC and AICPA standards). Certification to IAB standards would go most of the way in addressing the problem.

4 comments about "Traffic Of Good Intent: A Learned Perspective On Quality Traffic".
Check to receive email when comments are posted.
  1. Scott Pierce from A6 Corp., March 12, 2014 at 2:27 p.m.

    Excellent job summarizing the standard controls the industry could and should put in place. Unfortunately, as you call out in your article, almost everyone is the advertising supply chain benefits from fraud. Only the advertisers are left holding bag with wasted ad spend. The industry can make a valiant attempt to detect and weed out fraud, but it may be a game of whack-a-mole. Rather than trying to eliminate fraud, a better tactic for advertisers is to only advertise on high quality pages. Technology now exists that can measure the quality of RTB inventory, allowing advertisers to target only high quality inventory thereby avoiding junk and fraudulent inventory. It’s easier to measure the quality of inventory from the top down, than it is to detect and eliminate fraud.

  2. Steve Plunkett from Rockfish, March 12, 2014 at 3:06 p.m.

    This is such an awesome list of those things that always come up as a problem later but nobody addresses first.. Thanks.

  3. Paula Lynn from Who Else Unlimited, March 12, 2014 at 6:25 p.m.

    Good luck.

  4. Eric Scheck from Cross Channel Digital, June 4, 2015 at 12:37 p.m.

    In the millisecond, prebid world, user agent lists are virtually worthless beyond a first-level screen when anybody anytime with mal intent can switch/mask their browser identity, which bots do. Statistical pattern recognition between NHT and human traffic combined with dual caching and subsequent clearing process (embargo period before payment release) may be a good step. Even if supposed NHT is a false positive, agencies and publishers can appeal through the clearing process to be compensated. I'm not aware of any trading desks/DSPs doing this, now, and it probably makes sense for independent third-parties with forensic capability to manage the clearing process, adding another layer of transaction security and transparency. My two cents.

Next story loading loading..