Most of the online ad industry sees fraud as an external threat, while it is mostly an issue with the supply chain of advertisers' campaigns.
A company that sells fraudulent web traffic for site monetization has won a contract to provide clicks to the US Government and is also launching their own fraud detection platform. This is an internal threat to the adtech environment where things are built on trust.
If taxpayer money is spent on fraudulent traffic and now the creators of fraudulent traffic are also the ones measuring its quality, this erodes the foundation of trust we operate on in our industry.
2. of2 19 The Dhar Method
Anyone that has read anything about ad-fraud has been exposed to the standard story of the
"cybercriminals" and "hackers" that are inļ¬ltrating the digital advertising ecosystem to syphon out
millions and billions of dollars in supposedly illicit funds. These types of press reports and stories have
helped fuel the idea that fraud is a problem that is external and a threat that is diļ¬cult to protect
against.
ā¢ What if the truth was entirely diļ¬erent?
ā¢ What if advertising dollars generated by fraudulent practices are the product of behavior by companies in
our own backyard?
ā¢ What if the problem is internal and can be dealt with as such?
Many would say that the internal problem of fraud would be more easily addressed; and they would be
absolutely right. Even if the "cybercriminals" and "hackers" are the ones generating the traļ¬c at the
source, there are hundreds of "publishers" and ad-networks that are funneling that fraudulent traļ¬c
into the mainstream advertising ecosystem. And those publishers and ad-networks have corporations,
names, faces and bank accounts. Those four factors make them easily addressable and removable
pieces from the larger digital advertising system.
What is Internal vs. External?
Now the question is, how do you know who is bad and who is good?
Our instinct is to trust the trade-press to do the vetting of these companies and to provide us
information and stories that are supported by facts and truth. Unfortunately, there are some in the
trade-press and larger publisher circle that don't know enough about the nitty-gritty of ad-technology
and ad-fraud speciļ¬cally to be able to discern what is legitimate and what is not.
Let's take the example of a company called eZanga. If you look it up online, you'll ļ¬nd a Wikipedia page
and many articles that mention their anti-fraud eļ¬orts and legitimate business practices dealing in
"qualiļ¬ed traļ¬c".
All of the articles and posts below lead up to what is the most important update regarding eZanga:
(2016, August 24) http://www.marketwired.com/press-release/ezanga-awarded-gsa-professional-services-
schedule-contract-2152997.htm
eZanga has won a contract from the US General Services Administration to provide PPC advertising
services to Federal Government Agencies.
The Federal government of the United States is now buying pay-per-click advertising from eZanga, a
company for which I will present evidence in the following sections to show, that generates, sources,
and sells fraudulent web visits and clicks.
This is a huge concern. Tax dollars are now potentially going to be spent on fraudulent advertising
activity.
3. of3 19 The Dhar Method
https://en.wikipedia.org/wiki/Ezanga
5. of5 19 The Dhar Method
(2013) http://www.adweek.com/sa-article/ezanga-154108
6. of6 19 The Dhar Method
In 2014, eZanga posted a YouTube promotional video of their traļ¬c service. The video has
approximately 1700 views and 0 comments. In the video at 0:38 the following statement is made by the
narrator: āDave now buys pre-ļ¬ltered traļ¬c from eZanga. And we ļ¬nally got the third-parties to agree to
something: eZanga's traļ¬c score's well.ā
They are not purporting that the traļ¬c is human nor oļ¬ering an explanation of how the traļ¬c is
generated or where it comes from; they focus on one aspect, that it āscores well.ā
(2014) https://www.youtube.com/watch?v=OGzYT5es0Is
A blog post on WebSite Magazine in 2014 state the following: āWhile other traļ¬c providers require clicks be
paid for before giving feedback on its quality, eZanga traļ¬c is pre-ļ¬ltered to eliminate fraudulent clicks before
they are directed to clientsā websites. eZanga spends its own money to test each new source of traļ¬c before
exposing it to their client base."
(2014) http://www.websitemagazine.com/content/blogs/posts/archive/2014/11/08/ezanga-puts-click-fraud-in-its-
sights.aspx
In 2015, a post on LeadsCon.com, the website for a prominent aļ¬liate marketing and performance
advertising conference, wrote about how eZanga was unique in that they weathered the storm of the
recession in the United States after 2008. This post in isolation is inconspicuous but given the nature of
selling click traļ¬c, it forced me to wonder how their business was so sustainable despite market trends.
7. of7 19 The Dhar Method
(2015) http://www.leadscon.com/despite-recession-ezanga-enjoys-fairly-fast-growth/
2015 also saw the release of eZangaās WordPress plugin, which monetizes in-text traļ¬c from content on
WordPress sites. For those who are unfamiliar, in-text traļ¬c is when certain words in the content of a
site are converted to links to an advertiserās landing page, who is bidding on those keywords.
"Digital marketing ļ¬rm eZanga has released a new WordPress plugin called InText, which automatically places
advertisements sourced through eZanga on a blog or website. InText diļ¬ers from other WordPress advertising
plugins by working in real-time, identifying keywords and matching targeted advertising instantaneously where
other advertising plugins can take up to ļ¬ve minutes to register content."
(2015) http://www.wpmayor.com/ezanga-releases-wordpress-advertising-plugin/
11. of11 19 The Dhar Method
Although on the ļ¬rst page of Google results, there is this:
http://www.complaintsboard.com/complaints/ezanga-middletown-delaware-c263034.html
A complaint about the quality of their traļ¬c from 2009 with a series of comments back and forth
between their CEO/Founder and other complainants.
The CEO even responds with this statement:
"Now, letās looks at complaint boards, many users complain about many things. A quick google search
(as you request) for companies like AT&T, APPLE and Coke-a-Cola will also produce results like:
-> AT&T scams loyal customers with "Free Upgrade" SCAM
-> Apple scams us again - no FW connector w/mini - iLounge
Heck, Coke-a-Cola even made it to this complaint board:ā
The justiļ¬cation, which is seemingly quite legitimate, is that every company gets complaints and that
eZanga is no diļ¬erent.
Now, this is all what's publicly available and can be debated to no end.
So, I want to share with you my own experience with them in buying traļ¬c. I reached out to them to get
traļ¬c to my trusty old site www.eCelebNews.com.
I obviously would only want traļ¬c that passes the major fraud ļ¬lters so I chose to have access to all 5
that were available.
Integral AdScience, DoubleVerify, Moat, Forensiq, and Pixalate.
Let me be VERY CLEAR: None of these 5 actual fraud detection companies knew ahead of time,
that eZanga is selling traļ¬c with their name on it.
The way this happens is that eZanga owns their own "publisher properties" which they run traļ¬c to
and monetize in various ad-platforms. They sign up either directly or through a third party for access
to these ļ¬lters to scan the quality of their traļ¬c.
From the perspective of the fraud-detection vendors, they are just dealing with a publisher that seems
to occasionally have questionable traļ¬c to the site.
The way they create this "pre-ļ¬ltered" traļ¬c is by testing and exploiting the ļ¬lters they have access to.
See Dhar Chart Fig. 9.
12. of12 19 The Dhar Method
See my emails with one of their sales managers:
First I get accounts for various sources of traļ¬c, each one corresponding to a speciļ¬c fraudulent traļ¬c
detection vendor, along with corresponding pay-per-click prices.
She even goes as far as helping me with where to monetize this pre-ļ¬ltered traļ¬c.
13. of13 19 The Dhar Method
IntegralAdScience ļ¬ltered traļ¬c, she says, can be monetized on any banner network from āthe
exchanges.ā
Pixalate ļ¬ltered traļ¬c, she says, can be monetized on any search feed.
MOAT and Forensiq ļ¬ltered traļ¬c, she says, works well with video networks but not one in particular.
No suggestions on what to do with the DoubleVerify ļ¬ltered traļ¬c.
Main Point: It āall really depends on which ļ¬lter the advertiser on your side is using.ā
You might be thinking, āSo what? What does any of this prove?ā
To conļ¬rm the anecdotal evidence, I had a bot-detection company, Oxford BioChronometrics to scan
the eZanga traļ¬c coming to the page (eCelebNews.com) and provide the results.
If you had only seen the ļ¬rst few press mentions that I showed to you, it would be easy for you to pass
this company oļ¬ as a legitimate advertising company.
14. of14 19 The Dhar Method
Now, here is where the problem gets messy. The United States government, through its Government
Services Administration (GSA) has contracted out all of its Pay-Per-Click advertising services to eZanga.
Although not directly proven with data from their campaigns, there is the imminent danger that tax
dollars will be spent on fraudulent web traļ¬c.
http://www.marketwired.com/press-release/ezanga-awarded-gsa-professional-services-schedule-
contract-2152997.htm
There are a variety of services oļ¬ered via this contract, designated as Packages A-G. The minimum
order for clicks is $100 and the maximum order is $1,000,000. There is an overall 1% discount from ālist
pricesā and a 1.5% discount for orders between $25,000 and $49,999.99. Additionally, there is a 2.5%
discount for orders between $50,000 and $249,999.99, and then a 4% discount for orders more than
$250,000. The payment terms are Net30, meaning that payment for March services must be received by
May 1st.
Please see detailed breakout of packages on the following page:
16. of16 19 The Dhar Method
They are oļ¬ering to āredirect a user upon clicking to any website of choiceā for $0.10 to $0.20 per
unique visitor from a text-based advertisement (see sample of eZanga ad-format below). This Cost-Per-
Click includes a dedicated account manager and the unique visitors can be targeted from anywhere in
the U.S, even from speciļ¬c states and cities.
The ad-sample above is taken from one of my accounts on eZangaās PPC platform, AdPad. I have 5
accounts there, one for each of the following āļ¬ltersā:
IntegralAdScience, DoubleVerify, MOAT, Forensiq, and Pixalate.
I have tested 25 diļ¬erent campaigns on eZangaās platform for diļ¬erent research initiatives and client
projects. The amount of clicks delivered was always in line with my budget speciļ¬cations, regardless of
what the text of the advertisement said. This is a clear indication of robotic traļ¬c because the
consistent click volume, even with gibberish ad-copy shows that no human on the other end is clicking
based on interest in the potential content of the landing page.
Also, there was never an āapprovalā process as referenced in the screenshot. Campaigns have gone live
immediately, whether I set it up at 2 PM or 2 AM.
One important distinction is between the pricing given to a customer like me and the pricing in the
public contract with the federal government. Out of my ļ¬ve traļ¬c options, the higher end of the bid
landscape was $0.01. The GSAās lower end of the CPC bids is $0.10 and the higher end being $0.20.
This is a 20X diļ¬erence!
18. of18 19 The Dhar Method
While the capabilities statement provided to the GSA by eZanga makes claims of the traļ¬c being
āscrubbed cleanā but I was given these claims as well on traļ¬c that ended up being over 70% bots. The
main government point of contact is listed as their CEO.
They attempt to claim that āreceiving PPC traļ¬c from a company that not only purchases clicks, but
also sells them, aids in your brandās safety.ā There is no logic in trusting a broker who also has a
ļ¬nancial interest in the sale of said good.
19. of19 19 The Dhar Method
I have excluded several points of data and screenshots that may be inļ¬ammatory, but I will make them
available by request. Please e-mail me if you have any follow up questions or want clariļ¬cation on a few
things.
I began my anti-fraud crusade in 2014 thinking that I could ācriticize by category and praise by name.ā
While this may have been idealistic, I still truly believe that would be the preferred way to operate.
Over time, I realized that there are certain organizations and individuals that need to be singled out for
particularly devious practices and this report covers one that I thought was becoming a concern
because of its contract with the U.S government and the possible spending of tax dollars on fraudulent
web traļ¬c.
Ad-fraud has been repeatedly justiļ¬ed, by the people committing it, as a small harm done to huge
corporations. That rationale does not apply here.
I implore anyone reading this report to please be vigilant with your advertising practices. Not every
company shows their true operation or is transparent about the quality of their good/service, and this is
especially true in digital advertising today. If you have a hunch about a company that you think is
proļ¬ting from the sale of fraudulent advertising or web traļ¬c, please feel free to reach out to me and I
would be happy to point you in the right direction or help you address the problem directly.
There are lots of disagreements on the rates of fraud in the online ad market. The one thing all can
agree on is that there is no base-truth. Fraud, in any arena, is something that lives in the shadows. It is
diļ¬cult, if not impossible, to accurately count something that does its best to not be counted. What I
think we can agree on is that ad-fraud is not an external force that harms the advertising world, but
rather most of the problem lives in our own backyards. We can take this as an alarming idea, or we can
realize that it makes the problem actually easier to solve. The choice is ours.
- Shailin Dhar
shailin@DharMethod.com