Paul Pellman is CEO of Click Forensics, a traffic quality management company.
How is Click Forensics addressing the display advertising marketplace beyond clicks?
While most of our clients are focused on the CPC space, we do have a number of ad network and publisher clients with significant CPM and more display-oriented businesses that see unique value in our current approach. We've even seen a number of ad networks offering blended CPC and display ad solutions, which can deliver even better results for their partners.
These ad networks are using our click scoring insight to filter out invalid and poor traffic, better manage their various inbound traffic sources, and thereby make real-time optimization decisions to increase the overall CPC and display traffic quality they deliver to advertisers and partners.
We are also continuing to enhance our offerings to better manage traffic at the pre-click impression level, which is essential for ensuring display ad quality. As a result, ad networks and others will be able to use our technology in new ways to provide additional insight and value -- namely, impression-by-impression traffic quality scoring that can be incorporated into the ad serving or routing decision. This involves not just taking the existing scoring approach and delivering it faster/at the impression level (a significant technical challenge in and of itself), but also revisiting how to score impressions differently than clicks, by leveraging different impression-level attributes and heuristics. We're are already working with select clients on this and plan to have some interesting results we can share soon.
What special challenges does an ad exchange present from a fraud perspective?
It's clear that ad exchanges provide significant benefits and efficiencies in the buying and selling of online media. But they also present additional challenges from a traffic quality and transparency perspective -- mainly because ad exchanges typically add another intermediary layer between the fulfillment of media between publishers and advertisers. This creates more risk of masked, nefarious activity from fraudsters and cyber criminals, and makes it more difficult to identify good quality traffic sources from bad.
The resultant use of third-party ad tags, multiple redirects and overwritten referring URLs makes it difficult to uncover these traffic quality and brand reputation issues without a robust third party solution like the one Click Forensics offers. Fortunately, our efforts around impression scoring and our existing relationships with some of our clients who are actively buying and selling media on the exchanges allows us to provide this added insight and protection.
What trends are you seeing in online advertising fraud these days?
We've been tracking overall traffic quality trends for the last four years through our quarterly Click Fraud Index, which is the industry's leading independent click fraud reporting service. The Click Fraud Index provides statistically significant industry CPC data collected from online advertising campaigns for both large and small advertisers across all leading search engines and more than 300 ad networks.
During this time, we've noted a number of interesting trends. Lately what's very clear is the increasing sophistication being utilized by fraudsters and cyber criminals, most notably thru the use of botnets to perpetrate collusion fraud. This happens when a user's PC gets unknowingly infected with a virus or malware that can, behind the scenes (typically in a "Zero-iFrame" window), be scripted to click on ads from a collection of sites controlled or owned by the fraudster. In pretty rapid order, a fraudster can control a large network of PCs to generate fraudulent clicks from thousands of diffuse IP addresses across hundreds of different/rotating sites.
One other interesting trend has been an increase in scripted attacks against ad networks. Historically, we've seen higher incidence of attacks against the top search engines. But as these top players have continued to increase their advertiser protection and fraud detection capabilities, the fraudsters are more and more looking for less protected environments that can lead to higher returns on their malicious activity.
Is fighting fraud based outside the U.S. more difficult than domestic?
Fighting fraud is really no different in the U.S. versus internationally. It's becoming more difficult no matter where you are. We truly operate in a borderless world.
However, we definitely see significant differences in click fraud rates across the globe, as noted through our Click Fraud Index statistics and trend data. As you might expect, third world countries and other countries with less than ideal internet security tend to have higher rates of click fraud, as it's easier for fraudsters and cyber criminals to maliciously infect these PC's. In addition, we do still see Click Farm activity, which tends to be more concentrated in low wage countries.
When does the government and the legal system get involved in what you do?
We've not really seen much government or legal system involvement. That said it is interesting to see some of the media providers using the legal system as an avenue to fight back against fraudsters and scammers. Microsoft's recent Microsoft vs. Lam lawsuit shows that the industry is getting more forceful in its efforts to take action to protect their network and their advertisers. In this case, companies are using the legal system to send a clear message to fraudsters that they're not only being watched but that there is significant financial incentive NOT to commit click fraud. We're also seeing it as leading ad networks and publishers continue to look at third party solutions like Click Forensics to uncover fraudsters for possible similar prosecution in the future.
Where is the big opportunity for Click Forensics? Helping ad networks, advertisers or publishers? Who's getting hit the hardest?
Click Forensics believes it's important to offer solutions to all players in the online media industry. We've always said it will take a community approach through which everyone -- ad networks, publishers, advertisers, search engines and independent third parties -- work together to help solve the problem of click fraud and improve traffic quality and transparency. This is similar to how other types of cybercrime (malware, spam, etc.) are handled as well.
We have a significant advantage of being able to help all these parties because of the robust community data set we have, which includes data on traffic sources and conversion data from ad networks, publishers and search engines. This rich data set allows us to continually build and optimize our heuristics. In particular, the deep and rich conversion data we get from our numerous advertiser clients provides unique perspective and insight.
That said, lately the brunt of the responsibility for traffic quality typically ends up on the ad network. Advertisers tend to hold their agencies or ad networks accountable for quality traffic, and publishers want to make the most money possible from their content/visitors. It's the ad network that has to balance these competing goals and we help them do just that. Ad networks can improve their top and bottom line by proactively managing their publisher networks and screening new traffic sources effectively to deliver higher quality traffic to advertisers. Their efforts seem to be paying off. Over the past year ad networks have made significant progress helping to solve the traffic quality problem by using tools like ours to better control the traffic they deliver to advertisers. And it's showed up in the recent drop in the industry click fraud rate we reported for Q2 2009.
What types of media are the biggest offenders? Why?
Invalid traffic is typically generated from clicks on text ads, so no specific type of media is immune to attack or more susceptible, for that matter. The kinds of botnets or scripted attacks that I described previously can (and do) occur against all types of media. Obviously, there is more incentive for malicious fraud in the content network, where publishers share the CPC revenue, so those networks are more prone to invalid activity. But it's not specific to any type of media.
What is the revenue model for Click Forensics?
We offer a SaaS solution that's priced based on the monthly volume of traffic that we process and score for a customer.
How will the increasingly real-time nature of online advertising affect Click Forensics?
We're big believers in RTB and the positive impact this will have as more and more traditional media is transitioned online and as online media campaigns focus more on buying specific, targeted audiences. It's also important as an avenue to improve overall media buying efficiencies as the existing costs to execute an online ad buy are just too expensive as a percentage of total media spend.
Click Forensics is already taking this into account as our real-time click-by-click scoring API allows for improved, real-time filtering and traffic routing. And our early efforts around true impression scoring will allow us to provide traffic quality scores as a component of the ad serving decision.
10. How is the consumer faring these days when it comes to fraud and online advertising?
Consumers are only impacted by fraud in two ways:
- When advertisers pass along the cost of fraud in their goods and services, or -
The impact of the first point is hard to measure. But the second can have all sorts of strange side effects such as search results being manipulated, or destination sites being substituted, or even having your machine blacklisted because it's the source of so many invalid clicks.
We've noted that these botnet and click fraud attacks are many times related to other fraudulent activities, such as identity theft or SPAM attacks. So in this sense consumers not only play a role in helping to stop fraud and poor quality traffic, but they also have a vested interest to ensure their computers are protected from such attacks.