Skip to content

Idealab's new search engine,, offers advertisers a cost per action model and is presenting itself to would-be search-engine marketers (and those using existing PPClick models) as an antidote to click fraud.

The notion of "action" varies by business model — it can be a conversion, a registration, a download or any other event defined by the advertiser (question: does this raise the issue of self-reporting and advertiser integrity?).

Snap CEO Tom McGovern also argues that it offers more transparency and greater relevance to users.

McGovern previously told TKG, €œWe think the problem with search is not breadth, but relevance." Snap seeks to improve on the €œlink analysis€ methodology for determining search relevance by factoring additional criteria into the relevance calculus. €œWe look at page views, transactions and downloads. We also use human behavior and clickstream data to influence results.€ All these factors seek to avoid what McGovern characterizes as the €œgaming€ of search results, also known as search engine optimization in the industry.

I co-led one of the search sessions at the recent AD-TECH show in Chicago with the inimitable Kevin Ryan. The topic that got the most audience attention during the session was click fraud. There were two themes in the audience comments:

1. As long as paid-search is cheaper than other media, it doesn't matter all that much
2. I have no way to capture how many clicks are fraudulent

Still there was an atmosphere of general concern about the issue.

In 2004, the Search Engine Marketing Professional Organization (SEMPO) conducted a survey that asked, among other things, about click fraud. The following are top-level findings:

  • Most respondents (58 percent) using paid search expressed some level of concern about click fraud but admitted they hadn€™t tracked it.
  • Roughly one-third of interactive ad agencies or search-engine marketing firms identified it as a €œmoderate€ or €œsignificant€ problem.
  • Fewer than a third of respondents said they were not concerned about click fraud.

This suggests a lack of solid information about the reality of click fraud but a growing concern about the existence of a problem. As we've discussed before, the would-be class-action lawsuits and relative lack of anti-click fraud PR coming from the engines contributes to the perception of a growing problem.

Yet Snap has a long, hard climb if it's going to gain any meaningful share with consumers in an extremely competitive marketplace. Its real differentiator is on the advertiser side. But without meaningful consumer traffic, the advertisers won't show up — the proverbial chicken and egg problem.

The coverage that click fraud is getting may wind up giving Snap a boost with advertisers. However, that remains to be seen. Still, Snap is doing something different and interesting and will be worth watching.


Related: The Snap Blog, including a new US$10 million funding announcement.

This Post Has 2 Comments

  1. SME self reporting/integrity also comes into play in the Pay Per Call model; Calls may be logged, but is business generated tracked/reported? The SME instinct to under estimate/report business generated through directional advertising helps fuel directional properties' quest to develop tracking systems, such tracking systems, however, may not always be in the complete long term interest of SMEs. The Pay Per Call system, for example, does not always display SMEs’ real phone numbers.

  2. PPCall isn't going to display an actual phone number — it requires a unique phone number for tracking purposes. Billing would be automatic and based on call duration; merchants can argue with vendors later about whether charges were justified. But the system isn't based on whether the leads actually turned into business.

    The merchant temptation in the PPCall scenario is "call me back on this number" to beat the system; however, they generally won't have any way of knowing where the call came from until the tracking data are delivered later or reviewed later by the merchant online.

Leave a Reply

Back To Top