Google's decision to fix what wasn't broken is a bold bet on the future of search and the way people use the Internet.
Google Instant, which the company unveiled Wednesday, is a fundamental shift: instead of search as an outcome, Google is trying to get people to think of search as a process in which you constantly refine your query without actually "searching," or hitting the button to produce a concrete result.
Google is betting that in a world of nearly instant communication that search is going to have produce an answer just as fast as updates are spat out from Twitter or other real-time Web services. It's a bit chaotic at first and will certainly throw a few searchers off their game as well as make those in the search-engine optimization game a little anxious.
Should it prove popular with users, however, Google Instant is also the type of search innovation that might be difficult for competitors to duplicate in a matter of weeks or even months, giving Google a distinct advantage heading into a new era of Internet search.
Why reinvent the wheel, some might ask? After all, Google has dominated the search market for years with its current approach to search results, and even though its competitors have joined forces against a common enemy, Google is still a verb that means "to search."
While Microsoft and Yahoo deserve credit for keeping Google on its toes, most of those innovations have happened on the front end, or in the way search results are presented. Google Instant is a combination of front-end user interface design and the back-end work needed to process results for the suggested queries on the fly.
And at the moment, Microsoft and Yahoo are very busy on the back end of the search technology process. Can the two companies invest the time and money needed to replicate the back-end work that makes Google Instant hum? Had either company made this kind of breakthrough first, they might have really made Google sweat.
But they didn't, meaning Google will either cement its reputation as a search innovator should this take off or incur the wrath of users who liked things better the way they were. Google executives were careful to emphasize just how much they tested Google Instant with members of the general public--ostensibly to avoid gaffes like the Google Buzz launch--as well as the amount of time they spent preparing for the load on Google's infrastructure.
During Wednesday's presentation, Google revealed that it had been working on this type of instant search results for several years but faced huge challenges trying to make it work on their computing systems, which by most accounts is one of the most sophisticated on the planet. The first iteration of Google Instant increased the load on Google's servers by almost 20 times the level at the time, but through constant tweaks and improvements, Google was able to get the increase down to a more manageable two to three times the current workload.
The early signs are good for Google: at one point Wednesday, users trying out one-letter queries with Google Instant turned Google Trends into alphabet soup, and later in the day Google Instant still ranked as the No. 1 result in hot searches of the day. And few performance problems were reported as both professionals and amateurs put Google through its paces, although some people on slower Internet connections reported problems.
The longer-term effects of this shift will bear watching. Almost immediately, the legions of search-engine optimization consultants struggled to figure out what effect Google Instant would have on their businesses.
For starters, Google Instant changes nothing with the ranking process of Google search results, according to Ben Gomes, a distinguished engineer at the company. Google's Webmaster Central blog also moved quickly to address such concerns, suggesting that Web sites could actually see an increase in impressions--the number of times a searcher sees the site in results--because of the ease of generating a new search results page. Google is now counting a three-second pause in typing a search query as an impression.
But will certain companies now want to optimize their Web pages to appear in search results for single letters, perhaps? What will come up when you type "a," Apple or Amazon? Right now, it's Amazon for many users.
As with the ranking of results, nothing will change with the ranking of ads under Google Instant, said Johanna Wright, director of product development at Google. However, like search impressions ad impressions are also changing: Google will now record an ad impression if the user stops typing for three seconds while an ad is displayed, and also if the user types a partial query and clicks within the search results page. Advertisers who pay by the impression rather than the click may need to change their strategies.
One thing seems certain: the old ways of measuring search market share are likely kaput, according to Search Engine Land's Danny Sullivan. There's no way Google should be allowed to count every query keystroke as a "search," meaning that those who measure this industry are going to have to figure out some new way to quantify Google's advantage over its rivals.
But events such as Wednesday's show why Google has that advantage. When Apple CEO Steve Jobs introduced the iPod Nano to an adoring crowd in 2005, he freely admitted the company--in a similar position within the music-player market--was killing off its most popular product, the iPod Mini, in order to introduce something it thought was even better.
Google certainly isn't killing search, but it's making a huge bet that even though people like things the way they are, it has come up with something even better.
Few people remember the iPod Mini now. Google said several times Wednesday that it doesn't think people will remember static search in the future.