Businessweek Archives

Google is like primitive stone chisel


? What Do You Like? I Like That Too. |

Main

| No Olympic Buzz? ?

February 17, 2006

Google is like primitive stone chisel

Stephen Baker

Why is Internet search the industry of the future? Because today's versions are so excruciatingly primitive. The engines are implacable and undiscriminating, a little like the giant food processors in Upton Sinclair's The Jungle, which treated fallen factory workers as just another pig.

Case in point. I'm looking through Bruce Judson's online book, Go it Alone, which is financed by "contextual" ads on every page. (Thanks BuzzMachine) On page 89, he's giving advice to would-be entrepreneurs, and I see ads for lawnmower blades. Turns out that one of the startup examples he gives is for a retail gardening business.

"If Google's not good," my 13-year-old says to me, "then why are, like, 10 billion people searching on it?"

The answer, of course, is that today's search engines are extremely useful tools for us. We're like the sore-fingered cavemen who when handed a stone chisel consider it a gift from the Gods. But once we have search that does a better job understanding context, we're going to look back on today's search and wonder how we ever stood for it.

06:38 AM

search

TrackBack URL for this entry:

http://blogs.businessweek.com/mt/mt-tb.cgi/

I totally agree with your post here. According to most of the surveys I've seen, around 40% of searchers are still saying they are unsatisfied with the search results they are getting from traditional search engines. Since people can tell what information is "about" and computers can only analyze "content", taking advantage of a more human-indexed search is what I see as the trend moving into the future of search. I often use the metaphor of computers playing chess against human chess masters. Note that chess has only 64 squares and 32 pieces, and it has only been very lately that super-computers (not PC’s) have been winning these matches with any regularity. Google claims to have 90 billion items under index, so a software technology running on a personal computer being able to analyze that amount of data better than a human is decades away. Similarly, Looking for the information regarding “fisherman cleans bass” gives a human a perfectly good context in which to search, whereas a computer is including a person cleaning a musical instrument. Humans know the difference, but computers don’t, regardless of the processing power put into the equation. As seach evolves, and human indexing becomes more and more prevalent, we'll see search grow from a stone chisel to a laser pointer. The human mind is the key.

Posted by: Steve Mansfield at February 17, 2006 10:30 AM

The straight answer for your son is "Because it's good enough." And that is the criteria that we *should* use for most of our tools.

Search is by definition a difficult problem, not simply because the current tools are primitive, but mostly because we have such great difficulty formulating our own thoughts in our own minds and then expressing them clearly enough so that even an intelligent human can comprehend them with laser-focus accuracy, let alone a mere "tool".

Although Steve M. laments that 40% of searchers are dissatisfied, I would look at the same number and say that we should be quite pleased with the fact that 60% of searchers *can* be satisfied using such a simple and *free* service.

Yes, "search" can be improved, but once you've gotten past that 60% satisfaction level, the gains are difficult, expensive, and frequently not all that they are cracked up to be.

I participated in development of a web search engine back in 1998, and even today the "quality" of search results is hardly better than (and frequently not as good as) what we were doing back then, when Google was "new" and unproven.

Yes, we're going to see dozens of "new and improved" search engines over the next couple of years, not to mention the many new ones out over the past few months, but users will end up seeing very little net improvement, overall. Maybe we'll move the 60% up to 65% or even 70%, but we could end up moving it *down* to 45% or even 40% as the user interfaces get more cluttered and complex and confusing while the size and complexity of the Web continues to grow and become filled with ever-greater levels of noise.

The only truly significant improvement will come when we finally shift to a knowledge-based infrastructure which included a rich level of user context, but that could be a decade or more down the road. Unfortunately, the so-called Semantic Web is more of a detour than a path to that knowledge-based infrastructure and it will take years for this reality to sink in. Yes, lots of people will disagree with that assessment or timeframe, but it is up to *them* to "walk their talk" rather than blythely insist that "better search is near".

-- Jack Krupansky

Posted by: Jack Krupansky at February 17, 2006 12:38 PM


Best LBO Ever
LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus