Putting an End to Click Fraud

Every once in a while, I'll get a call from an analyst, a client or an industry peer looking to get some information that could signal the end to the search marketing boom. One buzz phrase that naysayers have embraced recently is "click fraud." Largely misunderstood, click fraud is by no means the David that can bring down the Goliath of search. As long as everyone plays their part, that is, and the industry does away with a very one-sided pricing model.

In looking at click fraud from a search marketing perspective, it becomes abundantly clear that there exists a large delta between the two juggernauts and the rest of the industry. The systems put in place at Yahoo! search marketing and Google do a comprehensive job of proactively mining referral data to identify any traffic that seems suspicious. This fraudulent traffic is then credited back to the marketer, all without the need for arbitration, disputes or lengthy ordeals. Since the majority of search marketers don't explore their options beyond Yahoo! and Google, most people do not need to worry themselves about this issue.

For marketers who embrace a more comprehensive distribution strategy, click fraud does begin to take form in a material way. Many of the smaller distribution engines, such as FindWhat and Kanoodle, have less sophisticated fraud detection systems in place. In the past, we have seen FindWhat drive the same volume of traffic, if not more, than Google on the same keyword portfolio. If Google represents roughly 55 percent of the market (including AOL), and FindWhat somewhere around one percent, a third grader could see that there might be something fishy going on. Once the difference in conversion rates is analyzed, it becomes clear that much of the traffic a marketer is paying for is not derived from genuine human beings earnestly clicking on your paid search listing.

Here is an example from one of our clients. This client runs a mature search marketing campaign based on a target cost per acquisition. In looking at year-to-date data, we see that their average conversion rate from search is 2.35 percent. For some of the Internet Yellow Pages (IYP) sites like Verizon Superpages, it is closer to a six percent conversion rate. However, the FindWhat channel comes in at just 0.26 percent for the year, making it a channel that we can no longer justify managing. It pains us to do this as there are some legitimate sales that are coming through this channel. Everyone loses in this situation. FindWhat loses revenue; our client loses some incremental sales, and our team ends up frustrated as they are unable to use all search distribution channels that are available to them.

Anytime there is an alignment of business goals such as the scenario above, a solution should be easy to reach. In the case of click fraud, there is an unfair assumption of risk that lives on the shoulders of the advertiser. The advertiser buys from a certain channel with the assumption that what they are buying are actual searchers and not an automated script or team of manual clickers somewhere offshore. It isn't until after the advertiser spends their marketing dollars and analyzes the data that they realize that the traffic is not quality traffic. A solution lies in the risk being shifted from advertiser to publisher, in the case of the second tier engines.

Let's stop for a minute and look at how our offline brethren tackle the risk issues at the television upfront.

During the annual boondoggle known as the upfront, a season's worth of television advertising is purchased in a frenzied week of late nights, endless parties and celebrity hobnobbing. Advertisers buy certain programs for their anticipated audience. Obviously, there is a high degree of guesswork being done when the audience numbers are applied to a given show. While the television upfront has its share of critics, the one aspect to it that does make sense is the concept of "makegoods." Makegoods are the networks' way of putting some skin in the game. Say, for example, next season's version of "The Apprentice" is a real stinker and only half of the anticipated audience tunes in. The network then has to deliver the promised audience that the advertiser paid for. This can be done by providing advertising time on other shows, or an extended run on the same show. Either way, the advertiser ends up getting the audience they desired. It is the one aspect of the television upfront that I think the online community should embrace.

Imagine for a moment that the Tier 2 engines all abandon their Cost Per Click (CPC) pricing model in favor of a Cost Per Action (CPA) model. Rather than sticking the advertiser with all the risk, the search engine would pick up the risk and put its money where its mouth is. This would fundamentally change the way in which search marketing is budgeted, forecasted and planned. First of all, the traditional model of initially launching with Yahoo! and Google to figure out a keyword portfolio -- and then migrating to the second tier -- would disappear. Search marketers would first launch with the Tier 2s to establish what phrases, categories and creative are the most salient for their audience. Only after some solid analysis of performance on Tier 2s would an agency recommend the higher risk (though higher volume) options of the larger engines. This would lead to a larger number of advertisers participating in campaigns with the likes of FindWhat, Kanoodle, Enhance and others.

To maximize revenues, the search engines could migrate to a dynamic pricing model, albeit on the CPA side. Rather than having click bid tolerance dictate who is number one versus number two, listings could prioritized based on what the CPA payout to the engine would be. Say for the keyword "airline ticket" there are four bidders, the highest of which agrees to pay the publisher a CPA of $12, versus $11, $10 and $8. This advertiser would appear first, allowing the search publishers to maximize their revenue opportunities, while taking the burden of click fraud from the advertiser.

The incentive for click fraud would disappear. No one would be getting paid for non-converting traffic. Trust me, if no one is getting paid for it, no one will waste their time building scripts and hiring people to click on paid search listings. It also eliminates the dirty tactic of competitors clicking on each other's listings to exhaust ad budgets.

For this model to work there would have to be some post-click data sharing between advertiser and publisher. Ideally, both publisher and advertiser would use a third party ad serving, bid management or analytics package to verify the conversion data. The usual precautions like mutual non-disclosure agreements would have to be established for data protection on both sides of the equation. It would seem to me that this level of effort would be well worth it for both parties involved.

So before everyone shorts their Google stock and writes off search marketing as the next spam, let us step up collectively and nip click fraud in the bud by shifting the risk assumption slightly off the shoulders of the advertiser. With no incentive for click fraud, the problem will go away and both advertiser and publisher will be better off for it.

Ron Belanger is Vice President of Search and Affiliate Marketing for Carat Interactive. Belanger is responsible for guiding the strategic vision of Carat Interactive's search engine marketing and affiliate marketing practices. Belanger has five years of search engine marketing experience coupled with nearly ten years of technology consulting and account management. Belanger has crafted successful search strategies for leading companies such as Radio Shack, Philips, Wachovia and Best Buy. He was elected to the board of directors at SEMPO (Search Engine Marketing Professionals Organization) in 2005, and sits on the MSN Search Advisory Council. He has appeared in publications such as B2B, OMMA Magazine, San Jose Mercury Daily News, Adweek and AdAge. He is a frequent speaker at Ad:Tech, Search Engine Strategies, and other events relating to interactive marketing. Ron Belanger earned a BA from Clark University.

Carat Interactive is a digital marketing agency committed to growing its clients' businesses by creating more profitable and enduring relationships with their customers across all interactive platforms. Carat Interactive is one of the world's largest interactive agencies with more than twenty offices around the globe. With North American headquarters based in Boston, MA, Carat Interactive has four additional offices in New York, Los Angeles, San Francisco and Atlanta. Carat Interactive is owned by Carat, the world's largest independent media agency with more than $15 billion in worldwide billings. Carat Interactive clients include Pfizer, Hyundai/KIA, RadioShack, Hyatt, UPN, and Palm, among others. For more information on Carat Interactive's services, visit its website at www.caratinteractive.com.

 

Comments