There’s little doubt that search has improved over the past ten years. The jump from Alta Vista to Google in the late 90’s represented a seismic improvement. Yet, in recent years, the rate of improvement has slowed and with spam sites, affiliate marketing and improved SEO, you could make a strong argument that the quality of search has declined.
Many are awaiting the semantic web as the next major shift in search. But while semantic search holds promise, today its deployment remains limited to niche applications. Having worked with semantic technologies for years, I believe that they are many years away impacting mainstream use.
Jason Calacanis has released an alpha version of Mahalo, a new search engine designed to bridge the gap. Mahalo is a search engine that combines the best of search, user-generated content and old-fashioned editing. To bootstrap Mahalo, the first 10,000 search terms are being hand-written by guides. Mahalo result pages look more like a wiki than a Google search engine results page. These landing pages are designed to help you navigate content, rather than just listing the pages that scored the highest. Like wikipedia, search terms with multiple meanings are presented with options to help the user disambiguate the results. Mahalo adds social software capabilities to the mix. Users can comment on each page and Mahalo also offers a Digg-like “Top 7 bookmark” feature.
Whether or not Mahalo ultimately succeeds, the human-assist model makes a lot of sense. Tagging providers, like ClearForest, found much of their value came in combining human judgment with technology.
Using guides to build results pages is not a new concept; it was the basis of some of the earliest web search applications such as LookSmart. The challenge of this model is to build the pages on a cost-effective basis. Only time will tell if Mahalo can achieve that for the broad set of topics and search terms required for a general search engine. But for content providers with strong presence in niche markets, the Mahalo model could be a compelling one. While a general search engine will need to define results for hundreds of thousands of terms, within a vertical market, you may cover all the common searches with a fraction of that. I have long argued that vertical search is a great fit for content companies, who typically understand niche markets and have editorial skills which can be used to gather and organize information.
Internet search, driven by semantic analysis may be the holy grail of information retrieval. In the meantime, human-assisted search may provide a compelling user experience.
UPDATE: To accelerate Mahalo's expansion, Jason has created the Mahalo Greenhouse, which allows users to get paid for development of new terms. Further details are available on Jason's blog. SearchEngineLand adds their thoughts, including whether or not Mahalo is comparable to Seth Godin's Squidoo (along with Jason's compelling retort in the comments).
Posted by: |