I know.

Posts Tagged ‘Web search engine

Google in Jeopardy!

leave a comment »

Q: This company dominates search and faces extinction.
A: What is Google?

Google investors may want to brace themselves for a “permanently high plateau” as its flagship search engine and email service face the prospect of being marginalized or eliminated by a new competition. While the media focused on the televised game show Jeopardy! in which IBM’s Watson computer obliterated human opponents, they may have overlooked the real story: the birth of an ultimate search engine, not powered by Google.

Google is search, and so far, only search

Google has done well for its long-term investors. Since its IPO in late 2004, the stock has risen more than six-folds. But the business has stayed the same: 96% of its revenues in 2010 came from advertising, based on its market-dominant search engine. Pretenders have come–Yahoo, Ask, Wolfram Alpha, Bing, to name just a few–and most have gone. Soon, so may Google.

Google knows it must diversify and develop new income streams. It has developed or acquired services including Google Voice, Maps, Gmail, YouTube, Android, Chrome, and even a social networking site called Orkut that dominates in Brazil (and nowhere else). Yet, all of them combined contribute to less than 5% of the company’s total revenue. Profits are out of the picture altogether.

And now Google’s flagship services are under siege. Web-based email usage declined from 2009 to 2010. Users are increasingly turning to mobile services, including texting, and social networking sites like Twitter and Facebook to communicate. Mobile email access did surge 36%, but it was not clear how this is done (via Blackberry-like push or Gmail App-like pull). What is clear is that to the extent Google, Microsoft, or other providers want to consider advertising, screen real estate simply does not exist on cell phones.

ComScore reported that usage dropped 24% for teenage users between 12 and 17 years old. The only segment increases came from users over the age of 55. In technology, there is not much better leading indicator of success than the age differential. The evidence is clear: web-based email is shriveling.

Watson will kill the Google star

Even more troubling for Google is that the heart, brain, and body of the company–the lightly decorated search engine visitors see when they visit–will face an unprecedented and probably hopeless fight against IBM’s Watson, who recently obliterated human rivals in a competition that almost no one thought a machine could win perhaps just a few months ago.

IBM defeated the best chess player in the world a few years ago, but its new machine is even more ground-breaking. It was able to parse regular language, often shrouded in tricky puns and entendres, and find the likeliest answer to a question.

Ken Jennings, the record holder for the longest winning streak on the show, described how Watson did it.

The computer’s techniques for unraveling Jeopardy! clues sounded just like mine. That machine zeroes in on key words in a clue, then combs its memory (in Watson’s case, a 15-terabyte data bank of human knowledge) for clusters of associations with those words. It rigorously checks the top hits against all the contextual information it can muster: the category name; the kind of answer being sought; the time, place, and gender hinted at in the clue; and so on. And when it feels “sure” enough, it decides to buzz. This is all an instant, intuitive process for a human Jeopardy! player, but I felt convinced that under the hood my brain was doing more or less the same thing.

Translation: Watson was given a set of words, and had to decipher the question and find the answer. Exactly what a search engine ideally would do.

IBM has already landed a partnership with Nuance Communications, a speech-recognition leader, to develop an offering for medical diagnosis and treatment. And as the company’s legal counsel pointed out, “the amount of information available to lawyers is practically unmanageable.” Yet with Watson, a lawyer can analyze hundreds of millions of pages of content and mine them for facts and conclusions — in about the time it takes to answer a question on a quiz show. Consider the number of paralegals, associates, even lawyers this service can replace.

Searching for answers, not links

But the true path to glory is in search. Right now, a user has to figure out the mix of keywords for non-trivial searches. “Apple rotten” may turn up agricultural and other results, but is unlikely to find discussions of business practices.

Instead of finding pages that match keywords, Watson was expressly purposed to find answers to the question that those keywords represent. Others have tried before, most notably by, but invariably failed due to half-hearted efforts to understand the questions. Watson has come closer to any machine has ever done in replicating knowledge, and it will change “queries” into answers.

Where there is search, there is money, specifically in search-driven advertising. Instead of serving up unrelated and sometimes nonsensical ads, Watson can “understand” the nature of the query and pinpoint the products or services a user is looking for. Just like an effective salesperson can listen to a customer and propose solutions that the latter is not aware of or can imagine, Watson can do the same, 24/7, across the globe. Better matches mean greater returns for advertisers, and higher rates for the seller (Google, IBM, etc.)

While Watson can be used in a variety of fields that requires human interaction, its implication for Google cannot be more direct. Admittedly, it is far from being an economical alternative, at an estimated $3 million of hardware. But most queries will be easier than the typical Jeopardy! question, and economies of scale as well as Murphy’s laws have solved uncountable numbers of problems once thought to be too technologically complicated.

Besides, this is where the industry is going. Greg Sterling, founding principal of Sterling Market Intelligence, a consumer research firm focusing on Internet’s influence, points out that “answers not links” is stated objective of the likes of Bing and Siri (acquired by Apple). And as Google’s Marissa Mayer has said many times, “the ‘perfect search engine’ is one that ‘understands’ what we want.”

Remember the declining use of webmail due to the proliferation of mobile devices? Mobile search will also change consumer behavior, because the small screen will highlight the need to get to the answer as quickly as possible. How often have you seen someone whip out the phone to settle an argument? She certainly is not looking for ten or ten thousand results. As Mr. Sterling explains, “When I ask for ‘the best sushi restaurant in New York’, I want a very few of the best options, not 390,000 links.”

Even if Google tries and succeeds in emulating Watson, the effort may take a better part of the six years that IBM spent on the project, by which time Watson will be that much more refined. While Google is trying its best to develop Android and other ventures that result in a legitimate, new line of business, it needs to refocus on search to stand a chance of remaining relevant in the next few years.

Admittedly, it has a few advantages. With $36 billion in cash and counting, besting IBM’s team of 20-25 people would not be an issue. Google’s roots were already in search, and unlike IBM it is largely unburdened by desires to sell million-dollar hardware. But the effort to reinvent itself must start now.

Written by Gnostradamus

Sat, 2/26/11 at 1924