The Shifting Sands Of Search

I had lunch today with Grant Ryan from Eurekster, whose swicki technology aims to enable the long tail of vertical search. Meanwhile, thanks to AltSearchEngines, I now know that there are search engines that find words within songs, or speak the results, or launch from within another application, or provide two-dimensional navigable tag clouds.

On Tuesday, InfoWorld published a piece by Ephraim Schwartz called "The Demise of Google." His basic premise? Verticals, with the capability of going into much greater depth on a given niche topic, could represent a serious challenge to the search giant.

So is Schwartz right? Will Google's downfall come not from some upstart building a better mousetrap, but from a change in the way people approach search in the first place? I've set out to paint a picture of a world of predominantly vertical search.



  • It would be user-generated. While substantial categories like job search may be able to afford to support commercially driven engines, the further you travel into specialization, the less likely the potential revenue will make a business venture worthwhile. If vertical search truly aims to support the long tail, it will have to be driven by the users. Eurekster's business model shows that the company agrees; it's looking to avoid becoming a search engine itself, instead providing the technology for users to build infinite small search engines.

  • It would limit discovery. We go to verticals when we know what we want, but the very nature of discovery includes the element of surprise. This isn't to say that discovery doesn't happen with verticals, but the possibilities are confined to within your pre-defined niche.

  • It would be repetitive. Infinite, highly specific search engines will vastly increase the chance that someone in Tulsa repeats all the work done by someone in Tallahassee to create the perfect Britney Spears search engine. There'll be a Britney group on FaceBook (thankfully, I haven't bothered to check if there is one now), a Britney swicki, a Britney Wikia entry, and 500 other sources for Britney love that haven't been invented yet.

  • It would require its own search engine.... And this is the key point. How can you hope to find the many niche engines for the topics you're interested in without a general search engine? You can go to Universal News in Manhattan and physically see pretty much every magazine available in the world. Online, though, you'll need a Google just to find the verticals you're after.

    Schwartz draws a parallel between the online and offline worlds, remembering that niche magazines caused the downfall of general-interest ones. But there are some major differences between the two.

    One is volume. Even a thousand or ten thousand magazines don't begin to approach the long-tail volume possible with user-generated vertical search. Wikipedia has nearly two million articles in English right now; imagine the same concept for search, with a different engine for every topic.

    More important, advertisers can still pick and choose who they want as viewers. Life magazine had to charge for every one of its readers, even if an ad only stood a realistic chance of being relevant to a few of them. Google doesn't have that problem. Yes, you might get ads for Jaguar the car when you're looking for jaguar the cat. But the economics of it still continue to work as they do now, no matter how large Google gets.

    Call me a fence-sitter, but in the end I think there will be a need for both. Social networking has taught us that people want to create small personalized worlds of highly and personally relevant content. The need for general search won't go away by itself.

    Google's sitting on shifting sand, but for now it's not quicksand.

  • Next story loading loading..