I guess there’s a way to make a living at being cranky. I’ve long enjoyed John Dvorak‘s columns, stemming all the way back to the early days of the PC, and he certainly has elevated the curmudgeonly column to an art form. But I guess I wonder whether it sometimes is more entertaining than informative, such as his recent provacatively titled “Google Must Die” column. He does the usual complaining about how search results aren’t any good and it’s insane to expect them to improve as long as we use this ridiculous PageRank thing Google invented. Perhaps I am too close to the business, but I am more likely to sit in wonderment about how good search is than to bemoan its very real failures.
Image via Wikipedia
I take with a grain of salt John’s rants about the poor search results, because I have over 20 years of experience with search, and I know where we came from. When I first started, it was considered miraculous to just be able to find the right documents if you knew the words in it–oh, and searching through several thousand documents was considered a large search index.
So, those of us that once sentenced ourselves to a trip to the library to scan the Reader’s Guide to Periodical Literature can appreciate that we find what we need a little bit faster and more conveniently now. But, to John’s point, if it’s so bad, why does usage continue to grow at an amazing pace? It reminds me of the Winston Churchill quote that “democracy is the worst form of government except all those other forms that have been tried.” Search is really awful except when you compare it to any other way of finding what you are looking for.
But all of this can be excused as just a back-handed compliment to Google’s success, in that we now take the magic for granted and notice only when it fails, like baseball fans who expect an umpire to start out perfect and improve with age. He makes the accurate point that popularity does not mean quality, but the real problem with John’s column is when he complains that the search results for two different searchers are different.
Personalized search is actually our best hope of moving beyond the merely popular to the more personally relevant. Search relevance has always been about the content, about analyzing the words better. But if relevance is the match between the searcher and the content, then why aren’t we analyzing the searcher, too?
John’s criticism that two people using the same search terms can’t find the same results is off-base. That is buying into the old model that we just want to match search words to content. We don’t. We want all searchers to find what they personally are looking for.
And this isn’t some futuristic notion. Personalized search is already here, on a limited basis. You should expect that to expand greatly in the coming months, in my opinion. I further believe that it is the single most important untapped relevancy booster–more personalization will make search results far more relevant than they are today.
If I’m right, then John might have to find something else to be cranky about.