Posts tagged as:

Mahalo

Intelligent Search

by nick on September 4, 2009

Google are the hottest company on the planet and they have well and truly won the war on search. That aint news to you. Fighting them directly is a bit like voting Labour in the next general election – a waste of energy.

You can’t be more of a lion than the lion himself, so throw in the towel. Move on. Fight another battle. Use new rules or change the game (even slightly). Semantic search is the future battleground where the engine understands more of your searching needs and the data it’s mined.

To the untrained eye this looks like the same ball game but it’s a much cleaner slate. Rather like Formula 1 this year where the cars appear the same as previous models but they’re inherently different. The new rules in both fields promise some new victors.

What’s so new about this semantic search?

Data has moved on exponentially since Google’s inception in 1997 (or so). Blogs, microblogs (e.g. Twitter) other social networking sites and book marking services stream and highlight more information than anyone could’ve honestly anticipated in the 90s.

Harnessing this data torrent allows for real time search results. If you’d searched for “Iran election” in June you probably wanted the latest news and insight on the troubles, not a standard bit of Foreign Office research.

Context is also increasingly important. Typing “Jaguar in London” could produce zoo or a car dealership results. Intelligence is needed to distinguish which you needed (known as disambiguation).

Making sense of search based on context and fresh data is the Holy Grail (closely followed by monetising it). Semantic search is craving to do just that.

Who’s playing in the hit-Google-from-another-angle game?

Bing (from Microsoft)
Mahalo (50% original content, 25% search and 25% knowledge exchange)
Aadvark (Vark.com asks your network for answers)
OneRiot (a real time socially-relevant engine)
Kosmix (a web guide with a dashboard)
Hakia (tabs results: web results, credible sites, images and new)
WolframAlpha (type a question, get an answer)
Twine (a bookmarking site on steriods)

Some are more semantic than others but that’s just eight players who are all in their relative infancy. With Yahoo’s open API code, Boss, anyone has access to a huge engine and can adapt from the basic Yahoo chassis. Google may have called game over on search 1.0 but there’s a whole new future out there…

UPDATE: since drafting this in early August, Google have announced a “fundamentally big change” via their Caffeine update.

Clearly, this threatens to put Bing et al back into their corner while Google blazes ahead with market share aplenty and more advertising than MadMen could dream of. We all know that’s not guaranteed though.

{ 0 comments }

Wikipedia + Google = better search?

by nick on November 23, 2008

Jason Calacanis shares a number of attributes, along with the initials, of Jeremy Clarkson. Both are tall, gregarious and outrageously outspoken in their quest for PR. (According to .net magazine, Calacanis called SEO ‘bullshit’ and a ‘wasted industry’ while speaking at a search engine conference, and said anyone from PayPerPost should kill themselves.)

But, unlike Clarkson, Calacanis is more than a one-dimensional critic: he’s a creator. This is the chap who started Silicon Alley Reporter and brought us Weblogs, Inc before selling up to AOL. His latest project is the human powered search engine, Mahalo.
mahalo
Paid humans review the Wikipedia pages Mahalo includes (opposed to volunteers) and humans also scrutinise the Google results to reduce the risk of gaming. The two results are mashed up to provide what Calacanis reckons is the ultimate in search accuracy with a massive dose of trust thrown in.

It’s just out of beta having reached 100,000 search terms and attracted 4.6 million uniques in August – no alternative search engine has broken the 1 million barrier. But like so many ‘new’ web projects (think Twitter) they aren’t even trying to make money yet as Calacanis’ profile has secured the project five years of funding.

It’s about building a useful service that people want, but Calacanis is also very ambitious. He wants to create the next Wikipedia or Yahoo and carve a 10% market share out of search. It’s early days but keep an eye on this guy; he’s far more that just a mouth.

{ 1 comment }