Which Search Engine Has the Most Human Intervention?

The premise of Bessed is that the best search results are generated from a real person doing critical thinking about what a searcher would most want to see for a given search. In general, where appropriate, we feel a searcher would like a number of different types of sites put together in one quick yet complete package. Therefore, where appropriate, we try to offer in our top links these type of sites: an overview site, a link or two of recent news on the topic/keyword, recent blog posts, seminal historical events, videos of interest, photographs. Our thought is that if you are researching a topic, you could use our first 10 or so links and get a pretty complete picture.

In discovering sites, we obviously use the major search engines in addition to suggestions from site visitors. In doing so, I’m always interested to see which search engine I feel spends the most time manipulating results with the help of human editors.

While Google is the most accurate for the most searches, in my experience Ask does the best job of providing the greatest variety of sites that might satisfy a searcher’s desires. I have to believe that Ask uses a good number of humans in massaging search results, at least for more common searches. It’s amazing how good Ask’s results can sometimes be—and then even more amazing how devoid of links Ask can be at other times. The fact that their results for common searches are usually so good, while their results for more obscure searches can be so bad, is evidence to me that it’s not a purely robot-based operation like Google. (Google may do a bit of human intervention, but I think they’re much more focused on their algorithms.)

For example, Ask is the best at getting news headlines into their search results. Do a search on sports teams or politicians in the news, and Ask has jumped on it much quicker than Google. You might track it down via Google News, but one of Google’s weaknesses as an engine is that it takes longer to get newsworthy items into its main index and it often doesn’t give them much weight—unless, of course, they’ve been linked to a lot. But even with bloggers linking like crazy, a news item that would be of interest to a searcher often ends up nowhere in Google, while Ask will often pick it up. This again makes me believe that there is an editorial team at Ask doing some thinking about this stuff.

I could be wrong. It might just be that Ask gives a certain number of news sites high ratings in its algorithm and if they have a story that pertains to search term, it shows up. But Ask has shown me enough variety in search results and other idiosyncracies— for example, a search for Bill Gates brings up a 1994 Playboy interview in which Gates discusses the impending “information highway”—that it seems clear someone smarter than a computer is thinking about this stuff. Or maybe they’ve perfected artificial intelligence, and if they can get their robot working a little faster they might just be able to challenge Google.

I think Ask.com is the second-best search engine out there now, and wouldn’t be surprised if they challenge for the #2 spot in searches conducted over the next 5 years or so.

e-mail me: adam@bessed.com

Adam Jusko is founder and CEO of Bessed, a Web site promising “search without spam”, thanks to human-edited search results and ongoing visitor feedback. Do a search, offer your comments, submit your site–help create the “bessed” search site in the world.

Bessed Versus Wikiasari

So there’s been a big hubbub this week about Wikipedia founder Jimmy Wales getting ready to launch a new search engine, thus far dubbed Wikiasari, although that will not be its name in the end. It’s unclear exactly what Wales envisions for Wikiasari, in fact he doesn’t seem quite clear on it himself, but the obvious conclusion is that he wants to get volunteer users to help create search engine results in the same way that volunteers now create Wikipedia.

With Bessed having just begun and having a different but still-in-the-ballpark model, I thought I’d look at what we are doing in comparison to what Wikiasari may be doing, and the pros and cons of each.

Bessed is built on WordPress. We “seed” topics/keyword phrases with a few good sites and invite site visitors to add more sites, dis the listed sites, or just add their two cents via comments at the bottom of the search page. If we haven’t yet created a topic, the visitor can suggest it and we will go ahead and start the new topic. There is no charge to be listed.

Bessed is not a wiki; visitors can not physically make changes, but their comments are publicly made and publicly responded to. In the end, however, our editors make the final conclusion about what should be included and how the sites should be ranked.

Bessed editors are paid, either as employees or as freelancers. We feel this helps us maintain objectivity—no one is adding or not adding sites based on a personal interest in the search results.

On the plus side, Bessed offers: objective, human-powered search that encourages visitor participation but keeps spam out of results. Webmasters can get their sites listed free, and visitors can suggest additions, subtractions and modifications to the results, helping us improve the site continuously over time.

On the minus side, Bessed will have to deal with the issues of scalability. Simply put, a human-powered search site can not cover all the topics a robot-based search engine can, and it has the potential to be expensive to produce, especially in the ramping up phase. No one has been able to solve this challenge—Yahoo started with a Web directory and eventually gave it second-tier status while charging high inclusion fees. The Open Directory tried a volunteer-driven model, which has devolved to a state where no one can get a site listed and there are constant accusations of Dmoz editors holding topics hostage for their own personal gain.

For a startup like Bessed, the money/scalability issue is a huge challenge. Using paid editors versus volunteers is more expensive, even if it yields better results.

Now, on to Wikiasari.

Wikiasari will likely use a format in which visitors can manually make changes to search results, with the thought that this will help the cream rise to the top. This means that the site will rely on volunteers to create the vast majority of its service. As we have seen with Wikipedia, this can be done fairly well. And it completely solves the scalability issue, as you have thousands or even millions of people contributing to it. But can it be done as well with a search service as it has been done with an online encyclopedia?

Maybe, except for one thing that I believe is a huge factor. Wikiasari is a project of Jimmy Wales’ for-profit company Wikia. Wikipedia is non-profit. Are volunteers going to be interested in using their spare time to build a search engine that drives profits to Wikia, while they get nothing?

Jimmy Wales says that the recent cash infusion Wikia got from Amazon is not a factor here, but I have to wonder if Wikiasari is going to use some sort of mix of paid and unpaid editors to build the site out. After all, as a volunteer it’s a lot more interesting to help write an encyclopedia page about Lucille Ball (a la Wikipedia) than it is to make a list of Volvo dealerships in New Jersey (presumably a la Wikiasari). Trying to create a human-powered engine with unpaid volunteers who create and/or edit boring yet useful topics is going to be difficult, especially when it’s being done to build the revenue of a for-profit endeavor.

The hype this week has been about Wales building a competitor to Google. I don’t know if that is Wales’ goal or not, although he took some shots at Google in the original article discussing the launch of Wikiasari. At Bessed, our goal is not to compete with search engines but to complement them. We can’t do what a search engine does, but they can’t do what we’re doing, either. From where I’m sitting, Wales would be foolish to position a human-powered search service as competition to Google, but rather as a more intelligent yet less comprehensive alternative.

All in all, though, Wikiasari is a good thing from where I’m sitting because it gets more people talking about the possibilities of human-powered search, which gives me more opportunities to toot Bessed’s horn.

e-mail me: adam@bessed.com

Adam Jusko is founder and CEO of Bessed, a Web site promising “search without spam”, thanks to human-edited search results and ongoing visitor feedback. Do a search, offer your comments, submit your site–help create the “bessed” search site in the world.