Vertical Search 2.0 -- Can it Kill Personalized Search?
It's been about five months since Search Engine Land reviewed the improvements that Eurekter made to their Swicki product, and about a month since I installed a swicki on my personal blog and forgot all about it. Since I'm preparing the long-overdue follow-up to my Viral Video Case Study, however, I thought I'd find out just how useful a Swicki can be.
What I discovered is that the future of vertical search has nothing to do with topics, and everything to do with social networks, the wisdom of crowds, and user demand. In fact, I can even see it eclipsing Google's Personalized Search product.
Now, if you're wondering just what a Swicki is, it's a kind of search-engine/wikipedia hybrid. It's a vertical search product that works on the premise of online social networks. Instead of defining its niche like, say, Mamma's health search engine does by choosing a specific topic, it's niche is defined by the interests of the online community that's using it. Granted, at first glance, this seems like Eurekster is really limiting themselves to a really specific niche market -- i.e. the geeks that pick it up first -- but Eurekster is actually using social media to let there be as many Swickis as there are online communities. As their website explains:
swicki technology automatically leverages the knowledge and behavior of online communities to add weight and vertical specificity to search results. [...]
With these tools, publishers can now harness their brand and audience to become vertical search destinations.
Without spoiling too much, the follow-up to my Viral Video Case Study was gonig to be on the Lip Dub video. Now, if you Google the term "lip dub", what you get is nearly 670,000 results. The first result is nearly almost famous Flagpole Sitta video that was produced by the staff of Connected Ventures in an effort to promote their online video sharing proerty, Vimeo. Indeed, the first page of results on a Google search of the company's name features this very video. Back on the first page of the "lip dub" search, however, is mostly buzz around the video itself.

All this tells me is that the Flagpole Sitta lip dub produced by Connected Ventures was a both successful viral video and guerilla marketing campaign. It raised the profile of both the parent company and its online video sharing property. But this is something I already knew. What I wanted was information how to leverage lip dubbing for the purposes of viral video without having to google every possibly combination of "lip dub" and every other marketing-related keyword. So I tried falling back on personalized search by logging into my account and running the query again, but I got identical results.
It was time to try something new, and I recalled my online marketing swicki over on my personal blog. So, of course, the first thing that I searched for was "lip dub". Mind you, what the query returned included two results that I wrote, but at least they were all that much more relevant than what Google gave me. More to the point, however, is that other results returned by the Swicki were similarly much more relevant that anything returned by Google -- i.e. they had to do with online marketing. The first is a short blurb on viral video that I posted on my personal marketing blog, and the other is no other than the Viral Video Case Study her on the SearchAnyway Blog that I am preparing a follow-up piece on lip dubbing videos for.

So my results are more relevant than Google's but they're still not ideal. Mind you, asking for ideal results outside an advanced personalized search seems like lot. However, with a Swicki, it doesn't have to be. As you hover on the icons next to each result, you can not only see how many people have voted for each result, but can vote for or against each result yourself. This way, you, as a member of my online marketing community, can help increase the relevancy of the search results.



All in all, this seems like something that can really render personalized search obsolete by capitalizing on its weaknesses. The problem with personalized search is that it might not be able to adapt quickly enough to your changing interests because it might make it difficult for you to find results that are relevant ot your changing interests. In a word, personalized search pigeon-holes your interests, and then fills the hole with concrete. As Michael Gray pointed out:
I don’t go to the library to see only the books the librarian thinks I should read. I don’t go to the video store to see only the movies they think I should rent. I don’t go to the supermarket to buy only the food the supermarket wants me to buy. I’m an intelligent human being and can make decisions for myself...
A swicki, on the other hand, allows you to choose the range of interests that you want to determine the relevancy of your search results. If I want marketing information on, say, "lip dubbing" I might turn to the swicki on my personal marketing blog, but if I want to simply watch great lip dub videos, I might turn to the swicki on my favourite vlog or online video blog. This isn't only from the user's point of view, but from the publisher's as well -- it is actually feasible that they'll be able to harness their brand and audience to become vertical search destinations.


















