Tuesday, July 29, 2008
Unless you’ve been living under a gargantuan rock, Cuil needs no introduction. If you don’t know what Cuil is, look it up. But I recommend Googling it, not “Cuiling” it, if you want accurate results.
The Cuil (of course, we all know now that it’s pronounced “cool”) buzz was everywhere yesterday, which was understandable given that an ex-Google employee, Anna Patterson, led the innovation process. The debut was rocky; the site was down early in the day, presumably because of a larger-than-predicted influx of curious searchers (or experimenters like me) and, once the site was operational, complaints galore piled in regarding Cuil’s speed, efficiency and accuracy.
This morning, with the URL finally working for me, I set out to really give Cuil a chance. All the negative press out there gave me the impression that everyone wants to hate Cuil for whatever reason; countless writers testing Cuil seem to be molding every experimental search so that it “confirms” their gravest fears.
For example, this reporter claimed that (a) Cuil was slow as molasses and (b) he couldn’t even find Canadian Prime Minister Stephen Harper on a search. Well, said reporter may need to try “search engines 101 seminars” on his next query. Not only was Cuil lightning-fast on my relatively old and slow computer, it found Prime Minister Harper for me via drop-down menu before I even finished typing his name.
Determined to give Cuil a fair trial, I set out on my own experimental search. The results were mixed – not as resoundingly negative as some pundits would have us believe, but certainly loaded with, er, kinks.
As anyone working anywhere near the industry knows, Cuil’s hanging its hat on the fact that its index is “three times the size of Google’s,” letting it kitchen-sink its users with an unprecedented amount of results.
Cuil, Cuil, Cuil. Do you really think Google became what is today without exploring the search volume topic along the way? Do you really think the most powerful search engine in the world couldn’t dwarf your search index if it wanted to? Google understands the law of diminishing returns; once an index reaches a certain level, additional results don’t necessarily help anymore. They become redundant, cause clutter, and may even lead to red herrings. Danny Sullivan’s analogy was bang on; how can you find a needle in a haystack when someone takes the haystack principle to new level and dumps the entire haystack on you?
Case in point: Cuil struggled to give me credible results on the simplest of queries. Being the sports fan that I am, I searched “MLB single season records.” Cuil brought back:
- two different pages devoted to retired first baseman John Olerud
- a site claiming to be associated with the Sporting News that turned out to be a broken link
Consequently, the same search on Google brought the following top-two results
- “MLB single season records” – Wikipedia
- “MLB single season records” – Majorleaguebaseball.com
Cuil’s “trillion-site index” dumped a haystack of meaningless sites on me, whereas Google retrieved the most credible pro baseball site of all – the official pro baseball site, to be exact.
Such a search raises another one of Cuil’s supposed sources of pride – its uber-relevant search. I can’t begin to comment on whether or not it’s just the PageRank principle masquerading as something else, and there’s enough literature on the subject already. But I can illustrate a major oversight by Cuil. Assuming Anna Patterson and her cronies are right – assuming Cuil’s search does explore new places and doesn’t just retrieve the most popular results – is that necessarily what we want? Often, the most credible source is the most popular source. After all, why did it become popular in the first place? Industries that require lots of research and fact checking – namely, journalism and academics – need authoritative sites on which to rely. In this case, if Joe Journalist needs to check baseball records for a story, which search will satisfy him more – Cuil’s avalanche of John Olerud sites or Google’s quick retrieval of MLB.com?
Cuil’s other major claim to fame is its new search results interface, which divides results into thirds and displays pictures with the results. Having a look at it first-hand, it’s more impressive than I expected. The images aren’t too big and the tabs that pop up with results are usually relevant. Overall, though, I’m not crazy about the idea. For one, the photo-centric layout squashes a potentially lucrative advertising opportunity. All the photo-oriented, pop-up loving advertisers tired of being shut out by Google’s rigid standards could’ve flocked to Cuil’s space. But if Cuil’s results are loaded with photos and clutter, what impact will a flashy ad have on the page? Will it not just blend in?
The bigger problem with Cuil’s interface: the results page doesn’t look like a results page. It looks like a home page. Think about the average bounce rate of a given website’s home page vs. the average bounce rate of a results page. The results page obviously keeps users’ attention longer, as it displays what they’re seeking. In theory, the Cuil page should do the same, but anything that looks like a home page – full of pictures and links spread horizontally across the screen – could potentially confuse and turn off a reader.
I, for one, prefer the simple, clean Google format. Even if my desired result appears on page two, it’s still quicker to scroll down the page than it is to wander through Cuil’s vast swamp of pictures and maybe-relevant, maybe-not results.
An example of how the pictures can create confusion: I searched “Sopranos complete series box set” and was pleased to find lots of relevant results detailing where I can buy the product. However, the pictures associated with each result were not of Sopranos box sets! They showed other television shows like Heroes or ones I’d never heard of. How can users trust a source if they query apples and find articles of apples next to pictures of oranges?
I’ve tried my absolute best to give Cuil a fair shot. And it isn’t all bad – it looks pretty enough, it’s speedy, and it does have a few features that better the search experience, such as the tabs that accompany search results. Overall, though, it serves as yet another example of the “if it ain’t broke, don’t fix it” paradigm. If one business intentionally departs from a massively successful competitor’s system to be “different,” odds are it will be hard to be “different” without just being “worse.”
Anna Patterson says Google “has looked the same for 10 years and will look the same in a year.” So what? The hammer hasn’t changed much in the last century, and no one’s demanding a new, less boring way to drive nails into drywall.
View Posts by Category