The interview with Steven Marder was what I consider an introductory course on social search. The interview with Grant is the advanced course. You can comment on the interview below.
Recently, we had the chance to speak with Eurekster’s Grant Ryan. One of the most interesting topics of our discussion was Grant’s description of the innovative contextual advertising model implemented by Eurekster. This model is designed to take advantage of the context information that is inherent in the design of each Swicki (Eurekster’s name for their vertical search engines). You can read the full interview transcript here. This post will discuss the impact of Eurekster’s advertising model.
What Eurekster has done is to allow advertisers to decide what Swickis they want their ads to show up within. In the interview, Grant provides an example about a jigsaw puzzle Swicki. When users know that they are in a vertical search engine focused on jigsaw puzzles, they begin to simplify their search queries – they type “dogs” instead of “dog jigsaw puzzles”. This is great for users because they simplify their search queries and get better results at the same time.
In addition, because advertisers have matched up their ads with certain Swickis, the ads are more contextually accurate as well. Unlike keyword based ad systems, which will see the search phrase “dogs” and present possibly irrelevant ads (about, say dog food, or dog accessories), Swickis match their ads very well. Furthermore, advertisers can focus on finding contextually relevant sites, instead of the rather low-level task of defining hundreds of keywords. See the interview for some more detailed information on how this works with Swickis.
When you start thinking about this, it begins to seem like an essential component of the process. For optimal results in vertical search engines, the embedded ads also need to be relevant. Without this type of context matching, the ads will in fact be less relevant. This is bad for advertisers (low click-through), bad for users (they learn to pay even less attention to ads, and so, of course, bad for publishers. The system Eurekster has put into place solves the problem rather nicely.
Since this is one of the first things users of vertical search engines see, and it’s integral to monetizing search, we think that other search engines getting into customized and vertical search will be compelled to solve this problem as well. They will need to get their advertisers to match up their ads with vertical search engines.
This may eventually lead to a somewhat “virtuous cycle”, where ads get more relevant, making them more useful to users, more valuable to advertisers, and more profitable for publishers. Interestingly, though, whether and how the incumbents do this remains to be seen. Such a new model seems like a rather radical change away from the immensely popular (and immensely profitable) keyword-based pay-per-click model — a move that search engines may not be quick to want to make.
While Eurekster remains a small player in the search engine space, other vertical search engine players will be looking at what they have done to solve these types of problems. They have built their service to a level where they have a serious volume of searches (500,000 per day according to Grant) by solving the types of problems outlined above.