Netflix launches the “House of Cards” series. It’s not just a TV show. It’s the result of an in-depth statistical analysis and artificial intelligence (machine learning and deep learning) of customer movie preferences, which began a few years earlier when Netflix sent DVDs to customers, asking them to rate them from 1 to 5. Ratings were carefully recorded and catalogued for each customer. In 2007, Netflix even held a public competition to find a machine learning algorithm that could improve its ability to predict customer tastes by at least 10%. No one won the $1 million prize, but in 2009 the prize was re-run, and this time the BellKor’s Pragmatic Chaos team won with a 10.06% increase. 19 minutes later, another team, The Ensemble, submitted another solution, also just over 10%. Netflix therefore has a lot of data, knows how to use it, and has the courage to build its business on it. The first result is House of Cards: a story of lies, betrayal, and murder, just the right mix that Netflix knew would appeal to a large portion of its audience.
