Windows Live Search Can Be Cuil, Too

Cuil is the new search service launched July 28 by several former Googlers. Given Google's overwhelming search share and Microsoft's search engine ambitions, Cuil should seem like a nutty concept. Why would anybody push into a market so dominated by a single vendor?

Some people might give online advertising revenue as an answer, and that's potentially a good one. Even a few percentage points of share could be lucrative to a niche newcomer. But Cuil isn't looking for niches, but breadth—and that means competing with the big three—Google, Yahoo and Microsoft. I'm not going to jump on the Cuil-is-cool bandwagon, as many other folks have today. Google shouldn't sweat about Cuil.

But Microsoft should wonder about Cuil. Microsoft executives need to ask themselves two questions: Should they buy Cuil? Should they imitate its approach? See, Cuil's concept potentially is Google's undoing and perhaps Live Search's, too, but less so. Cuil looks at search differently than does Google, which can somewhat be said about Live Search, too. Google's search approach is legacy, and it's about time someone came up with a better approach.

Google and the Legacy Question
Google's algorithmic search success is based on 1990s and early 2000s concepts about how the Web works, particularly linking, for ranking pages' relevancy to Web searches. Google also makes decisions on just how much of the Web it indexes, which, again, relates to concepts about how the Web works. But what if the Web changes how it works? Or, more disturbing: What if the Web would change but is hampered by the dominant search engine's legacy approach and technologies?

The Web has organically changed from its inception, and potentially bigger changes are ahead, particularly as more people use mobile devices to search for personally or locally relevant information. Already, many Web sites offer mobile-specific pages. Google is adapting, but is it enough? Call me a skeptic.

I'm increasingly disturbed by how Google works. I find search relevancy to be less over time—and there seems to have been a dramatic downturn during the last year. Search is not my area of reporting expertise, so I'll guess some reasons: Maybe the Web is getting too big for Google or maybe some of that recent "tweaking" has had a negative effect. For certain, I work much harder to get relevant results from Google now than a year ago and much more than about six months ago. The single exception: local search. How about you? Is Google making you type more keywords more often to find what you want?

GOT A TIP OR RUMOR?

Search should be about what's important to you. The problem with Google's page ranking, incoming links and other relevancy measuring metrics: The results are what's important to everybody else. Then there is the impact of the economy Google has created around search keywords. I've never been satisfied with Google explanations about how much or how little keywords affect search results. That said, I'm convinced that the keyword economy is a framework for legacy infrastructure that won't easily bend with the modern Web or, worse, is preventing it from organically growing.

Now there is Cuil, which has a stated philosophical approach that is what I would expect from a search provider more in touch with today's Web, not where the Web was in 2000 when Google started hawking search keywords. From Cuil's Web site:

Rather than rely on superficial popularity metrics, Cuil searches for and ranks pages based on their content and relevance. When we find a page with your keywords, we stay on that page and analyze the rest of its content, its concepts, their inter-relationships and the page's coherency. Then we offer you helpful choices and suggestions until you find the page you want and that you know is out there. We believe that analyzing the Web rather than our users is a more useful approach.

In a blog post late July 27, Om Malik said, "Cuil might be on the right track." More:

How it works is that company has an index of around 120 billion pages (which is a lot smaller than what Google claims) that is sorted on dedicated machines, each one tasked with conducting topic-specific search—for instance, health, sports or travel. This approach allows them to sift through the Web faster (and probably cheaper) than Google, which still enjoys a huge infrastructure advantage over its rivals.

Cuil's approach isn't so far from where Microsoft has been going with Windows Live Search. Relevance is increasingly about relationships—pages to pages, keywords to keywords, person to person and what you want to what you need. Microsoft's philosophical and technological approach lies somewhere between those stated by Cuil and Google.

Cuil, Google, Live Search Comparisons
This afternoon, I compared some searches using Cuil, Google and Windows Live Search. I started simply with "population, Caribou." My home town is Caribou Maine. The first-page results from Cuil, Google and Live Search were really surprising. Live Search brought back the population of Caribou, Maine—8,308 based on U.S. Census data—as the first item. Cuil did OK, offering links to two different pages with demographic data on Caribou, Maine. Google delivered nothing about Caribou, Maine, on the first page of results, but there were some links on the second page—and to some strange places. Why should USA Election Polls be the first page Google offers for demographic information about Caribou, Maine?

For a different query, "How to make ice cream," Google did much better than Cuil but not Live Search. Cuil brought back the strangest results, stuff completely unrelated, like "How to make a press pass" and "How to make egg drop soup." Not one of the 10 results led to an ice cream recipe. The only seemingly right one, "How to make ice cream the old fashion way," [sic] led only to fritod.info. That said, I now know where to look to find out how to make many other things, like jam, yogurt and—get this—a sleeping bag.

One more: "iPhone 3G cases." Cuil, Google and Live Search delivered surprisingly different results. Google's results were the most commercially oriented and led to vendors like Belkin and Griffin Technology that I am familiar with. Live Search led to places like cases.com. Google's results largely favored case manufacturers, while Live Search instead offered retailers selling differently branded iPhone 3G cases. If the searcher wanted to buy an iPhone 3G case, Microsoft gave more relevant results.

Cuil's first page results weren't at all commercial, just links to blogs or news reports about iPhone 3G cases. Four of the first 10 hits linked to MacNN. Overall, Cuil brought back the least relevant search results, but by far the most interesting. The presentation is more pleasing than either of the other two search engines and category listings help to refine queries.

Live Search is closer to Cuil than is Google, methinks. Microsoft offers category-like lists and in my testing does better at strictly informational searches. Cuil's stated philosophical and technological approach is sensible, and where Google comes up short because of its legacy strategy and technology. But Cuil's stated plan isn't the same as execution. Cuil won't in its current form replace either Google or Live Search. Meaningful relevancy is one reason. Local search is the other.

In several, simple local searches, Cuil failed worse than on ice cream making. For "camera stores, San Diego," I got great local results from Google, almost as good from Live Search and nothing relevant from Cuil. Why would I want camera stores in Buffalo, N.Y.?

Microsoft should closely look at Cuil's approach and execute rightly on it. Google's legacy approach is its weakness. As one of my eWEEK colleagues said this morning, no one could have guessed in 1999 that Alta Vista or Yahoo would decline. They were search giants. Google's dominance isn't assured. Microsoft needs to keep pushing the technological boundaries and more aggressively market, market, market Windows Live Search. Microsoft, don't just sell advertising. Buy it.

[Please send your tips or rumors to watchtips at live.com]