Search engine Cuil (pronounce it however the hell you want to, but apparently they prefer “cool”) launched yesterday to a whole lot of “Google-killer” OMG vibes. And thus far, it really rather sucks.

The home page is Google-simplistic: Logo. Input box. Search button. Over-inflated index count. About link. Privacy link. On black. (Which is, I hear, the new black.) Start typing and a helpful suggest engine ala Google Suggest pops up. Cheers.

Try to search. One of several things will happen. Since we’re out of the “our servers are cooked” phase of things, chances are, you’ll get a results page. But if you were lucky to give it a shot early on, you’d just be flat presented with a “no results found” page. I searched “web application development” and “web development” and both came up with a 0-results page. This is apparently because the caching system isn’t able to retrieve results on the first request so instead places them in a queue. Except that Cuil doesn’t bother telling you that they’re still getting their shit together and that you’ll need to check back when they’ve actually pulled and cached those results. Not that you’d want to. Here’s why, after the jump.

An Index Bigger Than Google!

First and foremost, Cuil claims to be able to index pages faster and cheaper than Google. But the size of the index doesn’t drive relevance, except on the very edge of the very long tail. (Incredibly obscure search queries have a better chance of finding results if the index is incredibly large.) The question comes down to what the engines do with their indexed results. It’s not enough to simply cache and store a copy of every single unique page on the internet. You need to provide me the results that are most relevant. Google does this with PageRank and a giant swath of algorithmic goodness that looks at things like the weighted popularity of the sites linking in to you, Google Toolbar results, URL content, page titles and the like, with a lot of secret sauce thrown in. Presumably, Cuil uses some of the same metrics, since they just make sense. But a cursory glance at Cuil’s results set is frustrating to say the least and rarely what I’m looking for.

The Interface

The Cuil search results page is frustrating at first glance:
Cuil: Search That Sucks
Three columns of results if your resolution will support it, with a handful of pictures thrown in and the occasional breakout box suggesting alternate categories or like searches.

The issue with this is that people like to view search results-type data in list form. They like to click through multiple results, clicking a result and clicking back and trying the next result. When you have 15 results in a very roughly aligned grid on the page, the user has to define a strategy to select a result first. While the excerpts will typically drive my decision between the first and fourth Google result for a given search, I still put a good deal of trust in Google’s ability to rank their relevance and to present them in that order.

With Cuil, I’m not sure what ranks second. I’d guess that the top left result is first, but second is a mystery: Straight down and continuing down is confusing, since the fourth result is almost off the screen, while the fifth result is adjacent to the first. Ordering the results from 1-3, left to right on the top row doesn’t make sense either (though I think that’s how it’s supposed to read), since for some search results, the ninth result actually appears “higher” than the sixth result. This may be an effort by Cuil to lower the emphasis on “relevance rankings” at this level of granularity but search lives and dies by relevance and making that information confusing or scattershotting your results to a user and letting them pick is a risky game. Users are using your engine for guidance, not untethered choice.

The Suite

Cuil is a startup. But they’re a startup in one of the most difficult markets to break. Google isn’t just web search. It also looks up businesses in my area, maps the results for me, translates words, shows me the current weather, gives me the time in any city, converts my currency, shows me what my stocks are trading at and does basic math. These are mostly features that someone like Cuil can integrate in with a bit of time and some relatively simple programming. But mapping and directions are critical to me and Google does them very well. What’s more, they can integrate their map product in their basic web searches when they think you’re looking for something local. That kind of integration is incredibly useful and something Cuil won’t have for a good long time. We won’t even get into the fact that the remaining suite of Google products makes for an attractive consistency in look and feel across many different applications I use, from email to search to driving directions. Why would I choose another site for just search, even if the search is, say, marginally better? (It’s not.)

The Challenge

No one at Cuil is naive enough, I’d hope, to think that this is an easy challenge. Search is arguably the hardest market to break and one of the most critical: people rely on search in a way they rely on few other products to enable them to use the internet day in and day out. Microsoft has been tossing money, brains and more money at the problem for years now and their market share isn’t budging. And frankly, their search still isn’t that great compared to Google. They even have the benefit of something resembling a unified suite of products, including mapping. And they haven’t been able to win a battle, let alone the war. Cuil has $33M in funding. But just because it costs them less to build their index doesn’t mean theirs will be relevant enough.

And here’s the scary part: Cuil could build a genuinely amazing product. They could build the killer application for search, that would deliver better, more relevant results than Google and does so with a smile on its face. And that would be the least of their worries. Because Google users love Google. And everyone’s a Google user. Google doesn’t just work “well enough.” It works exceedingly well. Search is still their primary business and they’re not just standing still while others around them work to evolve.

If the fact that they are constantly working to better their mousetrap weren’t enough, the terrifying truth is that breaking people of their Google habit represents the single largest problem facing any search competitor. Google is a verb now. It’s the first place I go to search for something and it’s probably your first destination as well. The Google habit isn’t one people want to be broken of and giving them enough reasons to even consider your option, let alone how to spell it, is an immense challenge that will likely be met with failure.

There’s Room, Perhaps…

I’m not saying that someone can’t come along and bring something new to the way search is handled. Google did it to Yahoo, after all. It’s definitely a large internet and playing to a niche is a possibility that can drive you some market share. You don’t even need to play at Google’s level to be successful or to at least break even; far from it. But Cuil is looking to take Google head on. They’re looking to best them and deliver a superior product. And they may have bitten off more than they can chew.

It’s frustrating and telling to me that, despite the size or age of their index, the results that come back for my queries just seem… wrong. A search on Cuil for the lyrical phrase, in quotes, “by now you would think that I would be up” brings me the song as only the “third” result. The other 8 results on the screen focus on the word “that”, of all things. Meanwhile, Google delivers to me eight different sites with the song’s full lyrics, its name in the title of each of them. This is how I expect a search engine to behave.

Michael Arrington kept issuing the caveat that “they’re only an hour old” when he published his initial piece on Cuil. I’m not sure that’s an excuse. If their index is “larger than Google’s,” then what do they need more time for? If their search engine itself isn’t able to deliver relevant results to me with that large an index behind it, then it looks to me like the technology just wasn’t baked all the way through and that they needed to focus more on that before bringing exposing it to the light of day. His other claim that their breakout widget of alternate queries or search categories is useful is true. But it’s not a Google killer. If the component you’ve built to set yourself apart from Google could be executed by Google in less than a month, then I’ve got a bridge to sell you if you think you’re going to beat them at this game.

The worst thing you can do when trying to break into this market is to demonstrate your lack of ability to compete at this level early on. Anyone who might’ve given you a shot is going to summarily write you off and add you to the filter of also-rans that don’t warrant a second look. Get it amazingly right the first time and you might have a chance at really getting some of your launch momentum to work for you and build an attach rate. This isn’t going to happen with Cuil and frankly, it’s because it sucks.

Cuil: Search That Sucks (And for the love of all that is holy, don’t misspell it!)

Posted in: Rants