Thursday, May 31, 2007

Aloha, Mahalo - A New Twist on Human-Filtered Search Results

The native language of Hawaii has given us words like "aloha" that have slipped into general use as well as more other terms like "wiki" that have been appropriated for new uses. Add to that list of appropriations the Hawaiian word "mahalo," which means "Thank you" in everyday conversations and now refers also to Mahalo, the new user-driven search portal under development by Jason Calacanis. "Mahalo's goal is to hand-write the top 10,000 search terms," goes the boilerplate on its page templates, an objective that's being lead by ex-Anchors from Netscape and like-skilled guides. Visitors to Mahalo can suggest links for inclusion in the service. How does this all work? As an alpha-level product you have to give Jason some slack but in truth it's not something that you're going to figure out as a user in a few seconds. Thank goodness for the FAQ.

On one level Mahalo is quite simple: type in a search term, get either a page of information and links that's been largely edited by a Mahalo guide or something that's been generated automatically for terms that they haven't populated as of yet. Being day two there are lots more pages that are misses than hits, but a listing of the top 20 searches appears on each search results page to give Mahalo visitors a sense of who's looking at what. You can also enter questions in a natural language style, which will provide results that look a bit like an amateur's version of (partnership, anyone?). An example of a topic page more fully populated by Mahalo guides is Apple, which lists a "Mahalo Top 7" links for the term, disambiguation (Did you mean: "Apple, the fruit? Apple, the Beatles' record label?"), financial information, products, news, blogs and fansites, information and reviews, upgrades and support, photos and videos, competitors, and "culture". Items that Mahalo guides really dig get a little icon. In theory users can make comments on Mahalo pages, but in my short tour I haven't seen any yet.

Well, this is certainly...innovative. Or utterly derivative, depending on your point of view. I know from personal experience that there is one huge brain between Jason's ears and it seems as if every idea he ever had or absorbed about the content industry exploded all at once from his noggin onto the pages of Mahalo. From one angle what we have here is with user input: docents put together some light content that surrounds links. Okay,we know that works. Kind of. From another angle we have a dot-com era version of Hoovers, a light assemblage of business and product info to guide the initially curious. Interesting, but who is this aimed at? From yet another angle we have Wikipedia, a catch-all encyclopedia format that tries to catch a wide variety of facets about a given topic. Digg and other social bookmarking services enter into the picture with Mahalo Top 7 bookmarks, but there's not a strong sense of how useful the first seven results will be: social bookmarking services don't rank relevance all that well. And of course there's the analogy to, one-stop answers to questions from the best sources available. Except we really have to trust someone called a "guide" as to his or her judgment on sources.

Finally, there's the question of when I will know when to go to Mahalo. Will it be when I have a question that's one of the top 10,000 search terms? Oooh, is what I want to find maybe number 15,000? I dunno. Try "most popular," Jason, people will be able to get their heads around that more easily. Do I go there to get the latest news? Hmm, they have news feeds from Fox and other partners but but why would I get them here rather than other places - and why aren't the guides lending a hand with filtering and updating the news? While Wikipedia may be in the hands of "those darn users" I have a fairly high level of confidence that information on almost any popular topic will be updated within minutes, if not seconds, of something happening in the real world across a huge array of topics. I also know that Google will insert hot news at the top of my search results and that user-generated sites will help me to find the really cool news pretty quickly. I don't know how true that's going to be of any well-intended editorial staff covering tens of thousands of topics every day - even with help from users. Will I go there for shopping? Probably not, services like eBay and Google will scrape together the information that I need more effectively. Will I go there for reference information? Maybe, but with such a generic approach to content organization I'd probably prefer to type in a term on Google and branch off to Wikipedia,, Hoovers or other key sources that it finds so easily. Will I go there to browse their taxonomy? Probably not, I've gotten too used to getting information on any topic level with one phrase and a click.

So, when DO I go to Mahalo? That's something that Jason needs to work on a little more. There are a lot of very interesting individual features and there's definitely a need out there for something between algorithmic search engines and the chaos of social bookmarking, but I am wondering whether this is more about a product vision or more about what to do with all of those ex-Netscapers who were inspired by Jason. If it's more the latter then it's not clear that a fairly limited and relatively anonymous editorial staff is going to have the horsepower or the respect within a given topic arena to build relevance creds. It gives Jason the control over writers that he desires, but in specific topic domains it may take more editorial talent to pull this off than he can afford.

There are so many ideas forming at once in Mahalo that it's far too early to write it off as a mish-mosh of interesting concepts - especially since people are growing tired of the "gaming" of search results. Calacanis could put initial feedback to good use, form more useful partnerships and come up with a tool that really stands out for an increasingly sophisticated online audience. But at this point my bet's against it. With Google's "Universal Search" capabilities beginning to phase in and more pure user-generated content plays becoming more disciplined and deep it's not clear that the features in Mahalo will ever mature to the point where they'll gel into a useful product in comparison to more established search and reference plays. At the same time there's far too little a sense of online community in Mahalo to make people passionate about online content feel that this product is really "theirs" in any strong way. In between these approaches there's probably room for a product that combines the best of search, editorial skills and user input to create marketable context for popular topics. But for now I don't think that people will be saying "thank you" to Mahala for its attempts at filling that need.
Post a Comment