New research suggests that an elaborate cybercrime ring is responsible for stealing between $3 million and $5 million worth of revenue from online publishers and video advertising networks each day. Experts say the scam relies on a vast network of cloaked Internet addresses, rented data centers, phony Web sites and fake users made to look like real people watching short ad segments online.
Online advertising fraud is a $7 billion a year problem, according to AdWeek. Much of this fraud comes from hacked computers and servers that are infected with malicious software which forces the computers to participate in ad fraud. Malware-based ad fraud networks are cheap to acquire and to run, but they’re also notoriously unstable and unreliable because they are constantly being discovered and cleaned up by anti-malware companies.
Now researchers say they’ve uncovered a new class of ad robot or “bot” fraud that was designed from the ground up to keep its nose clean — running not on infected hosts but instead distributed across a vast, rented network of dedicated Web servers and computers.
According to White Ops, a digital advertising security company based in New York City, those rented computers are connected to a network of more than 570,000 Internet addresses apparently leased or hijacked from various sources.
White Ops dubbed the video ad fraud network “Methbot,” and says the individuals at the helm of this network are spending upwards of $200,000 a month just maintaining a fully automated fraud network that imitates real Web site publishers showing real viewers video-based advertisements.
Ryan Castellucci, principal security researcher at White Ops, said Methbot’s coders built many of the fraud network’s tools from scratch — including the Web browser that each rented computer in the network uses to mimic Web sites displaying video ads. Spoofing actual news Web sites and other popular video-rich destinations, Methbot requests video ads from ad networks, and serves the ads to a vast array of bots that “watch” the videos.
To make each Web browsing session appear more like one generated by a human, Methbot simulates cursor clicks and mouse movements, and even forges social network login information so that it appears the user who viewed the ad was logged in to a social network at the time.
“They’ve written their own browser from scratch in Javascript, and this allows them to arbitrarily control the information that gets fed back to the ad networks and to companies like us who try to detect this stuff,” Castellucci said. “This has allowed Methbot to scale to beyond anything the industry has seen before and putting it in a new class of ad fraud.”
Interestingly, the registration records for virtually all of those Internet addresses have been forged so they appear to be controlled by some of the world’s largest Internet service providers (ISPs).
For instance, one of the many Internet addresses White Ops says was used by Methbot — 196.62.126*117 — is registered in October 2015 to AT&T Services Inc., but the contact address is “
Another address tied to Methbot — 196.62.3*117 — is registered to the same
Whoever dreamed up Methbot clearly spent a great deal of time and money building the fraud machine. For example, White Ops says the address space alone used by this ad fraud operation has a current market value of approximately $4 million. A full list of the 570,000+ Internet addresses used by Methbot is published in the White Ops report page.
“Methbot operators invested significant time, research, development, and resources to build infrastructure designed to remove these limitations and provide them with unlimited scale,” White Ops said in its report. “They created dedicated data centers to support proxy networks in order to hide the single origin source of their operation. This is the first time we’ve seen data centers impersonating residential internet connections. This makes the scale of this operation virtually unlimited, with none of the typical durability issues of maintaining a constant base of infected user machines.”
White Ops said it estimated the earning potential of Methbot by looking at the number of phony video ad impressions it could serve up and the average cost to advertisers for displaying those ads. Assuming an average CPM (cost per mille, or per thousand number of impressions) of $13, the company estimates Methbot has the ability to serve between two million and three million impressions each day, with a daily revenue ranging from $2.6 million to $5.2 million.
WHO RUNS METHBOT?
White Ops’s report doesn’t delve much into the possible actors behind this ad fraud network, but there are a couple of tantalizing clues in their findings. White Ops found that the Methbot network originally used a program called Zombie to test the ad code in a simulated Web browser environment, but that later the Methbot team built their own Javascript-based browser. The report also notes that Methbot employs a program called “Cheerio” to parse the HTML rendered by the video ads.
Both Zombie and Cheerio show up in this October 2015 discussion thread on the Russian-language tech forum pyha[dot]ru. That thread was started by a developer using the nickname “adw0rd,” the same nickname listed in the phony ISP internet address ranges used by Methbot. A glance at adw0rd’s profile on pyha[dot]ru shows the user is from St. Petersburg, Russia and that his email is
The “contact” page for adw0rd[dot]com (again, with a zero) includes that same email address, and says the account belongs to a software developer named Mikhail Andreev. That page at adw0rd.com says Andreev also has the account “adw0rd” on Facebook, Google, Twitter, LinkedIn, Github and Vkontakte (a Russian version of Facebook). A look back at programming projects dating to 2008 for adw0rd can be found via archive.org. Andreev did not respond to requests for comment.
The “abuse” contact email address listed on many of the Internet address ranges that White Ops tied to Methbot was “
Tags: advertising fraud, adw0rd,
This entry was posted on Tuesday, December 20th, 2016 at 11:01 am and is filed under Other. You can follow any comments to this entry through the RSS 2.0 feed. You can skip to the end and leave a comment. Pinging is currently not allowed.