The biggest mistake most companies make when choosing a strategy is "listening to the Hippo -- the Highest-Paid Person in the Organization," Neil Hunt, chief product officer of Netflix, told attendees at the Churchill Club in Santa Clara, California.

The club presented Netflix with one of its prestigious annual awards on Thursday. Netflix won its Game Changer Award for forever changing how the world watches TV and movies.

Hunt was on stage to accept the award and interviewed by Foundation Capital's Paul Holland. (Holland was a Netflix venture investor who had worked for Netflix founder Reed Hastings at his previous company, Pure Software.)

Hunt says the other mistake companies make is to listening to their customers, because "customers are poor at knowing what they need."

At Netflix, data rules the company. In today's online, cloud computing world, where this philosophy has given rise to "data-driven" companies like Google and Facebook, such advice seems common sense.

But in 1997, when Netflix was founded as a subscriptions service that delivered DVDs through the mail, it was a radical idea.

Hunt says that really committing to this idea means loads of A/B testing and a willingness to accept a high rate of failure.

For instance, a major design change involved 600 A/B tests, of which only "150 had a material win," meaning the other 450 bombed.

This kind of testing requires painstaking commitment. It also means that changes and new product rollouts would be "very incremental, driven by customer feedback."

Kicking out the breadwinners to change the company

But a company that only changes itself in tiny, incremental ways runs a different sort of risk: being put out of business altogether by a new idea that challenges the whole business.

Netflix came close to the brink of that, too, in 2011, when it realized that online streaming was its future -- and then bungled its move into that business so badly, it became fodder for a "Saturday Night Live" skit.

Netflix increased prices and spun out its DVD service into one called Qwikster. The change ticked off customers so badly that 800,000 of them immediately quit the service, and its stock tanked to about $42 from over $300 a few months earlier.

Netflix quickly reversed the Qwikster idea.

Internally, this wasn't just a huge A/B testing fail. The company still believed that internet streaming was its future and that it needed to spin out the DVD businesses. So it stuck to its plan.

"We knew streaming, the internet, was the disruption. Very few businesses get to survive a disruption. Usually the incumbent dies and the challenger succeeds. We didn't want to be the incumbent," Hunt said.

Streaming video required a much bigger investment in technology than mailing DVDs. So it had to raise prices.

And Netflix still split the company apart internally. It selected key people to "spend all of their attention focused on streaming business," building it as if it were a startup.

"We pushed out the DVDs guys. They went into a different office," Hunt said. "Me and my team didn't want to spend any time thinking about the DVD business at all. They went off and continued to produce a nice revenue stream."

"We didn't do it gracefully," he added. "It was a painful lesson, but it's how we were able to pay to attention to new customers and new revenue stream, and not just stay focused on the old business."

Data + vision = transformation

By combining a data-driven focus and streaming video, Netflix was able to respond to another potential threat: the possibility that networks would stop letting it stream their shows.

So the company was able to transform itself once again into an original studio.

With over 80 million subscribers today, Hunt says Netflix has an "unprecedented" level of data that "helps us decide the kinds of shows/movies we should make, and we can predict an audience for them with a surprising degree of certainty."

One more thing Netflix has over traditional studios: It is so good at cloud computing that it has put its "studio in the cloud," Hunt says, and uses cloud computing to do everything from managing logistics (like union drivers delivering cameras to a location) to film editing (uploading the footage to be edited immediately in another country).

And it seems to be working. Netflix shows have now been nominated for over 400 awards and won about 90 of them, including nine Emmy awards last week.

This story first appeared on Business Insider.