Netflix and the art of data-driven decision making

Netflix and the art of data-driven decision making

For anyone involved in the book trade, one of the most valued skills is judgment. Is this manuscript worth publishing? Is there an audience for this title? What is the right format for this story? We might like to imagine that decisions on content are best made by those with the experience and instinct for it, but how might things be different in an age of AI, automation, and algorithms? 

As I’ve watched the rise of Netflix in recent years, and the way it has transformed global content consumption, I have often wondered how an old-school media mogul like Rupert Murdoch, John Malone, or Ted Turner might have run that business. What makes the CEO of Netflix, Reed Hastings, so effective? While other media leaders struggled with digital platforms, how was he able to navigate difficult transitions, such as when the company switched from sending physical DVDs in the mail to embracing broadband streaming? Is Netflix successful because it runs on algorithms, or because it is run by algorithmic leaders?

I had an interesting insight into that question when I met Andy Harries, the CEO and co-founder of Left Bank Pictures. Harries is one of the world’s top content creators, having produced titles including Cold Feet, Prime Suspect, Wallander, Outlander, and The Queen, which saw Helen Mirren win, among other awards, an Oscar for Best Performance by an Actress in a Leading Role.

Harries wanted to pitch a TV show about the British royal family, based on themes explored in The Queen. He met with all the major US TV networks, who liked the idea but, after lots of consideration and debate, couldn’t commit to moving forward. Finally, Harries decided to meet with Reed Hastings and Netflix’s chief content officer, Ted Sarandos.
It was the strangest meeting, Harries explained, as he handed me a cup of a coffee at his office in London. As soon as he walked into the conference room with Hastings and Sarandos, and before he had a chance to pitch the show, they told him that they were ready to move ahead. And not just with a pilot, but with a full season.

Unlike the other networks, the team at Netflix had already analyzed their audience data and had used algorithms to predict the show’s likely performance. They knew their audience and precisely the kinds of shows that would work. Furthermore, with an upcoming launch in the UK market, they believed that the proposed show would be a hit. And they were right. The Crown’s third season is now in production, and it has twice been nominated for an Emmy for Outstanding Drama Series.

Algorithmic leaders reveal themselves in the way they make decisions and solve problems. You don't have to work in a digital department to be an algorithmic leader. In fact, quite the opposite. How Reed Hastings and his team think about content, its relationship to their audience and their platform, and even how it should be presented and released is radically different from the way traditional leaders in media companies act and behave.

When you are capable of knowing precisely what any of your millions of global customers are doing or desiring at any point in time, how can you not see the world differently? How can you not seek to leverage machine learning, algorithms, and automation to fulfill those needs in a highly personalized way?

Not everyone has the scale and resources to leverage analytics like a global digital platform, but the lesson for the book trade is clear: ignore data and machine learning at your peril. The idea of machines making content decisions can be a confronting one, even for technologists. Reed Hastings himself didn’t always have that kind of perspective, he just realized that the availability of data and algorithms meant he needed to change his viewpoint.  

Humans will always play a key role in providing context and nuance to the content business, but in the near future,  designing and training an automated system to make a good decision, may be more valuable than you making the decision yourself.

The Algorithmic Leader: How to Be Smart When Machines Are Smarter Than You is out now.