The Hunt for Search Engine Innovation, Part 2

I'm a bit delayed in posting this, but here's my Search Insider column from last week, continued in the extended entry.

The Hunt for Search Engine Innovation, Part 2

GOOGLE SHOULDN'T REST on its laurels just yet. Last week, we blazed through Charles Knight's Top 100 Alternative Search Engines and found many areas where innovation was lacking. This week, we'll visit some of the high-impact innovation categories and engines.

Sorting the engines into categories isn't a perfect science, as many engines are hybrids. URL.com is a meta-search engine combined with user rankings, Ujiko combines a graphic display with user ratings, Exalead combines category filtering with image search, and Polymeta is a metasearch engine with filtering based on keywords and categories that also includes vertical and multimedia search. Don't try too hard to sort it all out; by and large the most impressive engines have a clearer value proposition. Let's see what they're made of.

High-Impact Engines

 

Vertical niches: Goshme, discussed last week, aims to aggregate all vertical engines in one place. On the Top 100 list, the most innovative vertical search engine is Like.com. Microsoft's recent Medstory acquisition also signals that the major engines are watching the vertical startups.

After last week's column, Jessi Zambrano wrote about Indeed.com, a job metasearch engine not included on Mr. Knight's Top 100 list. Job search engines have been among the most successful innovators, and they've also been among the priciest search-related acquisitions. Searching for jobs is also one of the few search activities that truly matters to consumers' lives. Compare shopping search ("I want a good deal on something I plan on buying") with job search ("I want a new/better job"), health search ("I'm trying to diagnose/care for myself or a loved one"), and dating search ("I'm lonely"/"I want to start a family"). The latter three categories really matter, so expect search pioneers to emerge from them. I'd include some kind of food search in that bunch, but once you're online, searching for food is generally not a matter of fulfilling primal necessities but finding a decent takeout joint.

Multimedia search: Here's where there's the most need for improvement, and several startups have a leg up on the major engines -- for now. The Top 100 list includes just a few examples of multimedia search engines, and they focus on video. Blabline is simply a Google Custom Search Engine. Clipblast bills itself as the world's largest video search engine, though I'm not sure how it defends the title (my bet: the honor goes to YouTube, MySpace, or most likely Google). Blinkx has the most momentum, and it's a favorite of Search Insider columnists; Aaron Goldman recently wondered if Google should buy Blinkx, and I predicted back in January 2005 that it was a ripe acquisition target.

Semantic Web: It's getting harder and harder to write about any form of Internet innovation without factoring in the semantic web. John Markoff recently covered the topic in The New York Times, and by the time the Times gets to reporting on technology, you know it's old news. The one semantic engine on the Top 100 list is Swoogle, a database impenetrable to anyone without a computer science degree (I can, however, tell you the difference between a rondeau and a villanelle).

That Swoogle is hard to parse is in a way ironic, as the gist of the semantic Web, to oversimplify it, is to provide a way for all forms of online content to better understand themselves and each other. For instance, if a search engine or other content site were to index or link to this column through the lens of the semantic Web, it would know that this column has everything to do with search engine innovation and nothing to do with obscure forms of poetry. It would also realize that Aaron Goldman is an esteemed MediaPost columnist and not this septuagenarian who has jogged 200 miles in 72 hours, and it would surely never mistake me for the more infamous individual who shares my name. In the most utopian visions for the semantic Web, such as those shared at a DoubleClick Industry Insighters Salon earlier this month, the Web will be so adept at understanding your own interests and wants that you won't need to search for anything at all.

That's one of those beautiful ideals, to become so good that you make yourself irrelevant. Could we really get there one day with search?

It's unlikely. Even if any form of search became that good, we're still hunters and gatherers at heart. We'll always want the empowerment of thinking that searching is a skill, and if the right result is presented to us, we'll take the credit, even if a computer programmer or algorithm actually made it happen. That means that the ideal search engine of the future, the standard every engine should shoot for that truly gives consumers what they want, will be one step shy of perfection.