during its nearly 6 years of existence. About 40 percent of the company’s 1,000-plus employees were software engineers, a figure it wanted to increase to 50 percent; 7 of 11 executive-staff members were engineers or computer scientists, a reflection of Google’s identity as an engineering company. Talk of finance and other business topics tended to be crowded out of meetings of the executive staff by its members’ focus on “designing things.”
Dr. Rosing named four keys to the success that Google had enjoyed to date:
a brilliant idea—an algorithm called Page Rank that created a better way to organize, rank, and rate sites on the Web—which had been conceived by two Stanford students and out of which had come a better search engine;
a motivating mission statement, which put forth as the company’s aim organizing and presenting all the world’s information and trying to make it universally accessible and usable. “That’s one of the few corporate mission statements that I believe is actually realizable in, maybe, the next decade,” declared Dr. Rosing, who underscored his use of the word “all” in qualifying “the world’s information;”
a business model fueling an extraordinary level of reinvestment in the business, which was very capital intensive, owning large numbers of computers; and
a practice of hiring large numbers of engineers—“a feedback loop that we keep pushing on.”
Dr. Rosing launched his discussion of the panel’s topic, which he reformulated as “selective hiring,” by stating that Google was “working on some of the hardest problems in computer science” in search of its “Holy Grail[:] … that, someday, anyone will be able to ask a question of Google, and we’ll give a definitive answer with a great deal of background to back up that answer.” This vision, whose fulfillment the company was not predicting for the near future, was carrying it far beyond the technical problem of searching Web pages. In its pursuit, Google had designed and built, and was running, what it believed to be the world’s largest distributed computer. At the time of the conference, Google was beginning what was essentially the fourth rewrite of its code, which was constantly being rewritten. The company had two basic hunks of code: (1) Google.com, the familiar search engine, and (2) a set of code for its advertising system and monetization. The company kept the two separate because it had a policy of not selling the ability to get into its index; Google robots had to find a site or information, which was then ranked algorithmically. Employees working on these two code bases were kept apart as well, although they did at times need to interact.
The task on which Google was embarked was too difficult to outsource, said Dr. Rosing, adding, “Rather, what we need to do is to pull together all the best