Click Forensics Malware Lab identified a new malware scheme targeting display banner ads. The program performs a pop-up or pop-under and rotates brand advertisers’ banner ads every 10-15 min in an effort to boost impression figures.

In Q4 2010, the countries outside North America producing the greatest volume of click fraud were Japan, the Netherlands, the Philippines, Sweden and France, respectively.

“While the overall click fraud rate dropped last quarter for CPC advertising, we saw the emergence of new schemes focused on display advertisements,” said Paul Pellman, CEO of Click Forensics.

“We are investigating the malware-driven attacks in more detail, but early evidence points to an impression inflation scheme. It’s something we will examine more closely and report on later this year.”

Click Forensics told WebProNews that the amount of click fraud traffic from botnets generally hovers around 33 percent, and it believes the sudden rise may be to due to the increasing sophistication and proliferation of botnets.

Botnets accounted for 42.6 percent of all click fraud in Q3 2009, more than doubling in the past two years and up from the 27.5 percent reported for the same quarter last year.

The overall industry average click fraud rate was 14.1 percent. That’s up from 12.7 percent for Q2 2009 and down from the 16 percent rate reported for Q3 2008.

In Q3, the countries outside North America producing the greatest volume of click fraud were the United Kingdom, Vietnam and Germany, respectively.

"The significant rise in botnet-generated click fraud lines up with recent findings of several well-known malware and online fraud tracking experts," said Paul Pellman, CEO of Click Forensics.

"Botnets perpetrating click fraud and other online schemes continue to grow in number and sophistication. Advertisers and ad providers need to be especially vigilant about such activity as we enter the competitive search marketing holiday season."
 

The dashboard is designed to help manage downstream publishers and affiliates and adjust online traffic sources and campaigns to improve quality and performance goals.

Paul Pellman
Chief Executive Officer
ClickForensics

"Our new dashboard provides a powerful and simple way for ad networks and publishers to figure out who’s sending bad traffic and who’s sending good traffic to CPC and display ads," said Paul Pellman, CEO of Click Forensics.

"Ad networks and publishers can use the information to increase the quality of traffic they deliver, resulting in improved advertiser retention and spend."

The new dashboard is accessible via web browser and can be used to identify, filter and block traffic that should not be billed or paid for.

Features of the dashboard include:

Click Type Identification - identifies 15 different categories of click types, including click fraud, out-of-geo clicks, double-clicks, volume spikes and sources of publisher collusion-fraud. In addition, it pinpoints sources of botnet attacks and spiders as well as traffic from co-location facilities.

Traffic Source Scoring - scores individual sources of online advertising traffic so publishers and ad networks can view the overall traffic quality they’re experiencing at any given time. Users can drill down into specific publishers, sub-publishers and advertisers for detailed information on click quality sources and destinations of high- and low-quality traffic.

Web Site Screening Tool – provides publishers and ad networks with a consolidated view of all publicly available information on a particular web site’s traffic quality, including WHOIS and registrant information as well as one-click access to traffic volume statistics from Alexa and Compete. In addition, it includes up-to-date traffic quality data collected from the Click Forensics community.

‘Movers and Shakers’ View - flags traffic sources with significant changes in volume or quality. The information can be used to alert publishers and ad networks of potential areas of concerns as well as traffic sources that have significantly improved or degraded in quality.

The new tool is called "Block List" which provides ad networks with a continually updated list of visitor IP addresses and publishers known to generate online click fraud. Block List can be used with the Click Forensics platform to block invalid traffic and to shift  advertsing dollars to other sources that offer better results.

The Block List tool could help marketers reduce the rate of click fraud between 15 to 25 percent according to Click Forensics.

Paul Pellman

"Our new Block List is a quick first step for ad networks looking for information on some of the best and worst online traffic sources," said Paul Pellman, CEO of Click Forensics.

"They can use it to help their clients make the most of their online advertising investments."

Examples of blocked traffic sources include:

–Publishers committing click fraud;
–IP addresses of visitors known to commit a high level of fraud;
–IP address ranges generating non-human traffic, such as co-location facilities; and
–Other traffic sources that generate hacker attacks and spam.

"The Click Forensics Block List feature is an invaluable tool to help us protect our advertisers and deliver better results," said Evan Balafas, Managing Director of Excite Digital Media, a search advertising network.

"Just as we proactively protect our own computers from viruses and malware threats, we need to continually do the same for our advertisers so they get the most out of the millions of dollars they spend online."

That’s down from the all-time high of 17.1 percent reported for the fourth quarter of 2008 and from the 16.3 percent rate reported for the first quarter of 2008.

Click fraud traffic from malicious scripted programs increased in the first quarter of 2009. Unlike botnets or malware, these new threats exist as simple JavaScript programs that execute upon a page view or site visit. Ad networks were found to be especially vulnerable to these attacks during the quarter.

In Q1 2009, the greatest percentage of click fraud coming from countries outside the U.S. came from Canada, United Kingdom and Germany.

"It appears that the drop in keyword Cost Per Clicks (CPCs) and the progress Yahoo! and Google made blocking click fraud from botnet sources contributed to the decline in the overall click fraud rate this quarter," said Tom Cuthbert, president of Click Forensics.

"However, we also saw an increase in scripted attacks aimed at ad networks, which are historically more vulnerable to such threats. Advertisers should pay close attention to traffic from these sources over the next year."

The data also found that click fraud schemes are increasingly complex, sophisticated, and more difficult to detect. One new type of fraud discovered this quarter was carried out by malicious scripts that execute when a visitor views a web page disguised as relevant content or search results.

 The script initiates "Zero-iframe" or off-screen clicks that route the visitor session through an alias referrer website, and on to unsuspecting advertisers who pay for the phantom click. This occurs transparently to offending site’s visitor; they never see the ad or visit the advertiser, and their computer is not infected with any type of malware.

TechCrunch points out that over 30% of click fraud is via automated bots—another all-time high. This is a 14% increase over last quarter. I’m surprised the search engines aren’t doing a better job at keeping the bots down. This clearly continues to be a problem and needs to be addressed. I think the search engines need to release a state of the industry report on how they plan to cut these rates now ASAP.

Outside of the U.S., Canada is the largest driver of click fraud (7.1%), followed by Germany (3%) and China (2.3%).

TechCrunch also calls out that Click Forensics believes this may be a result of the economic state of the U.S., as old tricks like link farms are being used again. I think it is a direct result of a lack of focus, and perhaps effort, on the search engines’ side of things. What do you think?

Comments

The overall industry average click fraud rate grew 17.1 percent for the fourth quarter of 2008. That’s up from 16.0 percent in Q3 2008 and from 16.6 percent for Q4 2007.

The average click fraud rate of PPC ads appearing on search engine content networks, including Google AdSense and the Yahoo Publisher Network, was 28.2 percent. That’s up from 27.1 percent rate reported for Q3 2008 and down slightly from 28.3 percent for Q4 2007.

 

Click fraud traffic from botnets rose to its highest level, responsible for 31.4 percent of all click fraud traffic in Q4 2008. That’s an increase from 27.6 percent reported for Q3 2008 and 22 percent reported for Q4 2007.

The largest percentage of click fraud coming from countries outside the U.S. came from Canada (7.4%), Germany (3%) and China (2.3%).

"Based on the data we tracked in Q4 2008, it seems that the online advertising industry is not immune to the growing tide of cybercrime during this recessionary period," said Tom Cuthbert, president of Click Forensics.

"Both the overall click fraud rate and the rate of click fraud originating from botnets were the highest ever in Q4 2008. In addition, we’ve started to see old schemes like click farms reemerge. Advertisers should pay close attention to these types of threats in their online campaigns throughout the year."
 

ClickForensics Logo
(Photo Credit: ClickForensics)

 Yahoo is the first search engine to work with a third party click auditing company. The partnership will allow advertisers to share sensitive account information, such as click data with Yahoo through Click Forensics to eliminate click fraud and low quality traffic.

"Click Forensics will also provide additional advertiser feedback to Yahoo in an effort to help improve traffic quality," said Reggie Davis, vice president of marketplace quality at Yahoo, speaking on a panel Monday at the Search Engine Strategies conference in New York, where the partnership was announced.

Tom Cuthbert, president and CEO of Click Forensics, who was also on the panel said, "This means we’re going to work together on your behalf to identify and find these bad clicks. I applaud Yahoo in taking a really bold step in working with a third party."

A January Click Forensics report found that overall click fraud rate for online advertising was 16.6 percent in the last quarter of 2007, up from 14.2 percent in the last quarter of 2006.

Now, we know how this goes.  Click Forensics says click fraud is up, Google says it’s not.  There’s been a sort of running battle between the two for at least the past year.  So, to be blunt, we’re not suggesting that you keep a printed copy of this report beneath your pillow.

On the other hand, neither should you ignore the thing.  Tom Cuthbert, the president and CEO of Click Forensics, stated, "[I]t’s more important than ever before for advertisers, publishers, ad networks and search engines to cooperate and share data in order to stem what’s on target to be an even worse problem in 2008."  And suggesting cooperation and sharing is hardly the same as demanding money for a service.

Back to the stats, then.  According to Click Forensics, the overall industry average click fraud rate was 16.2 percent in the third quarter of 2007, and 14.2 percent in the fourth quarter of 2006, so a definite rise is taking place.

Also, "The average click fraud rate of PPC advertisements appearing on search engine content networks, including Google AdSense and the Yahoo Publisher Network, was 28.3 percent in Q4 2007.  That’s up from the 19.2 percent average click fraud rate for the same quarter in 2006 and 28.1 percent for Q3 2007."

Do what you will with that info, folks, but we’d encourage at least a day or two of hand-sitting to let Google respond.

Forbes has now taken Shuman’s criticisms and given Click Forensics Chief Executive Tom Cuthbert the opportunity to respond.

Here’s a snippet…

Are your estimates inflated by “fictitious clicks”–clicks that Google doesn’t count?

…All of our data is also based on Google Click ID, and Shuman knows that. He may be correct about other auditors in this space, but what he’s said and the way he said it is misleading.

How do you handle “click-backs,” situations in which a user reloads an advertisement’s landing page? Do you and other auditors count the page’s reloading as another click on the ad?

Absolutely not. Using Google Click ID prevents us from counting click-backs, and even when we have analyzed the number of click-backs compared to the overall problem of click fraud, it’s a very small percentage. Regardless, using Click IDs and other methodologies means there are no fictitious clicks in our data.

If I sound somewhat aggravated, it’s because I am. Shuman knows very well that we use Google Click ID. But he has a tendency to talk about fictitious clicks and then bring up our name to try to link us with that, even though he knows it’s not true. We’re a bit frustrated with that.

Forbes then went back to Shuman…

Shuman Ghosemajumder responded to Cuthbert’s comments by noting that while Click Forensics’ use of Google Click ID is encouraging, the auditor’s numbers were likely skewed by inaccuracies in how it counts fictitious clicks on Microsoft’s and Yahoo!’s network. He repeated his assertion that Click Forensics has access to only a tiny fraction–about 1%–of the data that Google uses to analyze click fraud, and again underscored that Google has no incentive to allow fraudulent clicks.

Can you figure out who’s telling the truth?

Comments

Tag: