Google’s golden rule: Don’t be data-blind

by on
in Management Training

Within software engineer Sean Knapp’s first few months at Google back in 2004, he came up with an idea.

“Why don’t we put five ads on the top of the search-results page?”

Back then, Google ran no more than two ads. Knapp knew that some users might be alienated by the higher number of ads. But he wondered whether the generated revenue would offset the reduced usage.

He and a product manager went to Marissa Mayer, then Google’s user-experience chief, and dropped this question: Was it possible that Google was running too few ads?

Because Google puts such a premium on data, Mayer and other execs were intrigued. If there were a “definitive” answer for how many ads could run at the top of a search page, Google should find out what it was.

So Mayer gave the green light for an experiment. For every 10,000 searches, one request would return a page with three or more ads at the top. This went on for several months until a clear answer to the question emerged.

The answer: Users shunned pages plastered with five ads, but they did tolerate more than two ads.

Google now runs up to three ads at the top of its pages.

“No idea is a bad idea until the data prove so,” Knapp says, repeating what may be the company’s second-most-popular mantra after “Don’t be evil.”

Deciding questions by data is to Google what global supply-chain management is to Walmart. What’s telling about this experiment is that it shows the willingness of Google’s top executives, including CEO Larry Page, to reverse themselves if the numbers don’t bear out.

Lesson: Let data work as an effective check against defending the status quo. It works for Google.

—Adapted from “7 Ways Larry Page Is Defining Google’s Future,” Farhad Manjoo, Fast Company.

Leave a Comment