Man vs. MachinePosted: August 29, 2007
Read an interesting article about Mahalo in Fast Company recently. Mahalo is basically the un-search engine. They don’t use sophisticated algorithms like Google to rank order search results. They use good ‘ol fashion humans. Search for HDTV on Mahalo versus HDTV on Google and the result pages look significantly different. Results are nicely bucketed in Mahalo by human editors. Google throws results all over the page in comparison and leaves it to the user to parse results to find what they’re looking for be it reviews, manufacturers or general information. To be fair, search engine optimization (SEO) creates some level of havac in the Google universe. But from a user perspective, Mahalo results are far easier to digest.
The whole human versus machine bandwagon is not what I found interesting about the Mahalo story. It’s the premise on how the company was founded was interesting. An excerpt:
After a year as an executive, he [Jason Calacanis] quit corporate life, determined to do something big. The way to do that, he reasoned, was to pick something people use every day on the Web: email, instant messaging, or search. He chose search. Much of the inspiration for Mahalo came from Calacanis’s wife, who had put together a short email for friends and family with links to places to stay and things to do in Kauai, Hawaii, where she and Calacanis had their wedding. The list was neat, organized, informative. Calacanis wondered why search results couldn’t look that way too.
Late last year, he set up a meeting with Michael Moritz of Sequoia, the Menlo Park VC firm, and walked in with printouts for “iPod” and “Kauai vacation” taken from Ask.com, Google, Technorati, Wikipedia, Yahoo, and a few other sites. He taped them to the wall, where some unfurled for 16 pages or more. Next to them, he placed his own single sheet of handmade results. “Which ones are the best?” he asked.
I love the way Calacanis pitched his concept to VC. He took a very bottom up approach to how real people would likely want to interact with search results. You contrast that with the typical computer scientists’ approach where the focus is more on relevance and display is simply a method to output your results. Google is a perfect example of this. As the story goes, neither Page nor Brin were web developers in any sense of the imagination. Not that they couldn’t learn. They just didn’t care. Their research at Stanford was around algorithms to improve search result relevancy. Actual display of results? Not so important. So, when they got to a point of launching their new search engine and needed a display vehicle, they threw up a text box with a simple list of links for their results page. Google’s beautiful, simple interface was born. The ironic thing was the simplicity of the design was an accident.
My meta take away from the article, was that you need a diverse set of perspectives to build great product. Bank on only engineers or business types or designers, etc., and you’ll get a homogeneous perspective in product developent. Google could you some good ‘ol fashion common sense, a-la Mahaolo perhaps, to improve their results display. And for Mahalo to be able to get the type of search coverage across the web Google commands, it’s going to need computer scientists. The wisdom of the crowds, or in tech world known as the cross functional team, rules.