Recently, Google started using its visitors and registered users as ‘experiments’ to find out exactly what necessary changes it should make to its main search website to increase efficiency of its Spartan search box and also to the results that it produces.

Marissa Mayer, vice president of search products and user experience explains the method as split A/B testing which involves Google introducing new versions of pages to users and measures their response level, at Google I/O conference. For example, the company was eager to know how many search results they should offer – the customary 10, 20, 25, or 30? When approached directly, users said they’d like more results, but testing proved otherwise.

At the conference, Marissa Mayer showed three slightly dissimilar versions of search results page that was tested with users. In this report, it was found that people searched 20 percent less when the results increased to 30 per page. After further analysis of servers, Google found this was because the speed was about twice as slow to display the longer results for the user, and in its business, speed matters. “As Google gets faster, people search more, and as it gets slower, people search less,” she quoted.

The same effect was observed for Google Maps. When the company cut the 120KB page size down by about 30 percent, it received about 30 percent more map requests. Split A/B testing also allowed Google to decide exactly how much white space to use around its logo and other elements on the search results page. Google stopped the industry practice of using a pale blue background behind ads to a pale yellow background. Users responded positively to this change.

As time passed by, Google has developed so rapidly and at the same time, kept pace with what its users actually needed then what they demanded. During its start-up, Google was a relatively unknown search engine. Google has also decided to present a beginner and an expert version of its search page, Mayer was quoted as saying. “The learning curve on search is really fast and people go from ‘Where can I get spaghetti and meatballs in Silicon Valley’ to ‘Italian food San Jose’ really fast,” she stated.

The complexity behind search

Although Google looks simple from outside, its search process is quite complicated. “A typical search will require actions from between 700 to 1,000 machines today” she said.  More and more complexities were found as Google moved to universal search which displays the regular search results mixed with that from its other search areas such as news, blogs, images, and maps. Mayer explained how because of the other narrower search services, Google missed sight of the simplicity its users need in a haste to bring the services to market.

Indeed in the long run, it is believed that universal search will grow far more sophisticated, with a page being filled with images, videos, and graphs with a holistic answer to your query, she said. Also, the search will become personalized, with results tailored for individual registered user. Google now looks on the bigger perspective, The Future, and plans to give more to its visitors and users.