At the Google I/O developer conference in San Francisco on Thursday, Marissa Mayer, Google's VP of search products and user experience, urged the audience to focus on the ordinary and the everyday.

When you focus on the ordinary and everyday you solve people's problems, said Mayer. She cited the millions of page views garnered by simple widgets as an example of the revenue potential of the mundane.

Such advice may seem overly simplistic, but Google has made a business of simplicity.

To hammer that point home, Mayer cited Occam's razor for logic -- the simplest answer is probably right -- and insisted that the maxim applies to design, too.

It's worth noting that Google arrived at that insight accidentally. Mayer said she asked Google co-founder Sergey Brin how Google had settled on the clean, white design of its home page. "We didn't have a Webmaster and I don't do HTML," Brin replied, she said.

And initially, all that white space confused Google users. User tests early on revealed that searchers would load the Google home page and wait for upwards of a minute in some cases. Asked why, the testers, accustomed to pages chock full of content, said they were waiting for the rest of the page to load.

That's why Google.com has a copyright notice at the end of its home page, said Mayer, to indicate that the page has loaded and that searching can begin.

Nowadays, Google relies on data rather than serendipity. Mayer stressed that Google is data-driven. For example, she showed three Google search results page designs that were served to alternate sets of users to test the impact of white space. The amount of white space in the three samples was almost imperceptible -- it varied by only a few pixels.

Nonetheless, Mayer said that Google found the version with the least white space performed best in terms of its user happiness metric and in terms of ad revenue generated.

Similarly, Mayer said that Google tried running sponsored results against a light blue background and a light yellow background. The yellow background led to more searches and better revenue.

And when Google compared search results pages with 10, 20, and 30 results per page, Mayer said the company found that putting 30 results on a page led to one-fifth fewer searches. Analyzing the data, she said that latency -- the subsecond delay caused by serving more results -- drove the decline.

"Users really care about speed," Mayer said. "They really respond to speed."

A typical Google search, Mayer explained, touches some 700 to 1,000 machines in Google's data centers and can return a list of 5 million results in 0.16 seconds.

"As Google gets faster, people search more," Mayer said. "As it gets slower, people search less."

Mayer reiterated the importance of universal search to Google's business and repeated the company's view that search in the future will be more personalized. And she urged developers to have a healthy disrespect for the impossible and to pursue fanciful projects because they exercise the imagination.

"Search is an impossible problem," Mayer said. "We will never have the perfect solution."

Even so, expect Google to keep trying.