• by Aaron Goldman , Op-Ed Contributor, July 1, 2009

There are many ways to slice and dice search query intent: Navigation vs. browsing. Entertainment vs. information. Commercial vs. non-commercial.

Accordingly, there are many different types of search engines: General vs. vertical. Real-time vs. all-time. Assets vs. answers. Pages vs. people. Facts vs. decisions.

And there are many different ways to refine your search if your engine of choice didn't get it right the first time: asset specification (e.g., image, video, blog, product, desktop, etc.), related searches, timeline, even a wonder wheel.

Of course, for each type of asset, there are many ways to drill down -- e.g., color, size, price, rating, etc.

Now, within the various engines, there are many different ways search results are displayed:  links, images, one-boxes, multiple panes, previews, snippets, etc.

And sure enough, there are many different ways these results are ranked: relevancy, recency, authority, popularity, personalization, etc. There's PageRank, PulseRank, and it won't be long before there's PeopleRank.

 
It's All Semantics

Whenever a search startup goes out to raise money or a stalwart rebrands or advertises to steal share, it typically focuses on one of the aforementioned flavors of search and sprinkles on a dash of differentiation.

Cuil was to be the massive index that protected privacy. Ask was to be the search engine for women (that one is still mind-boggling). Wolfram Alpha aims to be the fact... er, computational knowledge engine. Bing wants to be your decision engine. Mahalo is the answer engine. Twitter wants to own real-time search. Meanwhile, One Riot purports to be the real real-time search engine. Facebook is after the social search crown. And the list goes on and on.

What Flavor is Google-Killer?  

Of course, the irony in all this is that Google already serves most, if not all, of the flavors above in some shape or form. And it serves them well. (Or, as Malcolm Gladwell would say, "well enough" -- I just finished reading "Outliers").

As many havediscussed in the wake of the Bing launch, whether or not any would-be Google killers have a better recipe is a moot point, as the Google habit is well-formed and its footprint is huge.

It would be like a new QSR looking to unseat McDonald's as the fast-food king because it figured out how to cook a better-tasting burger. I've got news for you, would-be McD killers, people don't eat McDonald's burgers because they taste the best. They eat them because they are good enough, there are locations everywhere and there's something on the menu that everyone in the family can enjoy.

And, no, rebranding as something they're not won't help the other big QSRs topple the golden arches. Too bad Microsoft (or, rather, JWT) didn't get that memo before rolling out its positioning for Bing. Calling Bing a decision engine is definitely overpromising. Personally, I'd have created different spots focusing on each vertical in which it trumps Google -- travel, shopping, health, etc. -- and show the features (eg, Farecast, Cashback, Medstory) in action, before bringing it home with slogans like "Travel Better with Bing," "Shop Better with Bing," or "Get Better with Bing." 

Back to my QSR example -- I'm not saying there's no room for something like In-N-Out that makes kick-ass burgers. Half a billion in annual revenue ain't a bad living. Ditto for search. So settle into your niche, Wolfram.

And, while Wendy's may not be waaaaaaaay better than fast food, it is a solid #3 in the category. And that's not too shabby. Ditto for search. Although methinks Steve Ballmer is a little more aggressive than Dave Thomas ever was.

Query Is As Query does 

Let's go back to the various flavors of search. Again, searcher intent manifests itself in many different ways. For argument's sake, let's say there are just three core types of intent -- commercial, information, and entertainment -- and walk through the various ways a searcher could satisfy his or her craving:

1.       Query an answer engine -- eg, Mahalo or Yahoo Answers

         What is the best selling hybrid? (commercial)

         What is the typical Chicago weather in July? (information)

         What is the most popular surrealistic comic strip? (entertainment)

2.       Query an asset engine -- eg, Google, Yahoo, Bing, Ask

         Prius picture (commercial)

         Weather radar 60614 (info)

         "The Far Side" (entertainment)

3.       Query a people engine -- eg, Twitter, Facebook or Aardvark

         If you own a hybrid, why did you buy it? (commercial)

         Does it look like rain at North Avenue beach right now? (info)

         Can someone lend me their "Far Side" book? (entertainment)

4.       Query a fact engine -- eg, Wolfram Alpha

         Prius MPG (Commercial)

         Average Rainfall Chicago July (info)

         Far Side last date published (entertainment)

5.       Query a personal archive engine -- eg, MyLifeBits or Mint.com

         What do I currently pay for my car lease? (commercial)

         Did it rain last time I went to the beach? (info)

         Do I own any "Far Side" books? (entertainment)

Catch-All If You Can

But what's the real intent here? Why would you be doing all this querying? Bottom line, you're trying to make a decision. 

Wouldn't it be nice if you could query a true decision engine as follows?

        What car should I buy? (commercial)

        Is today a good day to go to the beach? (info)

       What comic book should I read? (entertainment)

Behold the promise of Hunch.

No, Hunch is not ready for prime time today. It will take time to learn about you, learn from others, build its index and refine its algo. But if there were any one flavor of search I'd be working on producing artificially right now, it'd be decision-making. It would, of course, taste like chicken but smell like victory.