There shouldn’t be any surprise in the “How Netflix is turning viewers into puppets” article but I still found it to be yet-another fascinating example of our new age of computing, one in which personalization and highly targeted experiences are the norm, one in which the big companies collect, combine, and analyze all sorts of data about our activities in apps, the Web, and the physical world.
In summary, Netflix has been collecting lots of data about what viewers do/like. They analyzed the data and made a $100M investment in a series, having high confidence that their customers will like it.
- They knew that their customers liked the BBC original “House of Cards” series from the 90s.
- They knew that the same customers also watched movies with Kevin Spacey or movies directed by David Fincher.
- So, they redid “House of Cards” with Spacey in the main role and directed by Fincher!
Again, there shouldn’t be any surprise. Loyalty cards were created for the same reason… to analyze consumer behavior. I still remember the story that Paul Watson (my PhD supervisor and good friend) told me about his time in the industry, building mainframe machines for data crunching. A large supermarket chain wanted to stop selling feta cheese because it wasn’t a very popular product. A data analyst told them, however, that they needed to keep feta cheese because the customers who spend the most always buy it. Another, more recent example, is the news about Target knowing about a teenager’s pregnancy before her father. Also, it’s been well-documented that Obama’s edge in the recent elections was his team’s ability to crunch big data (New York Times opinion: Beware the Smart Campaign).
It’s definitely a new world we live in. A major aspect of the device and service ecosystems from Google, Apple, Amazon, Microsoft is the aggregation of digital data about user behavior. We see it all around us. The more folks use Google’s search, the better it gets (remember Peter Norvig’s “we don’t have better algorithms than anyone else, we just have more data” statement?). The more people shop on Amazon’s properties, the better their recommendation systems become. The more we use Siri and Google Now, the “smarter” they become.
Apple might be calling their Apple TV product a “hobby” but I have no doubt that they are using it as a way to collect data, as a way to make educated bets for a future product, to quantify their decisions. It just makes sense. Google did it with their 411 call. Remember? You could call a free number, ask a question, and get information. Google collected lots of data in order to improve their ability to do speech-to-text recognition and also figure out how people ask local directory-related questions. I am sure Google Now benefits a lot from that data.
Back to the article about Netflix. There is some speculation of what additional data processing/inferencing Netflix might be doing: “It could make a lot of sense to consider things such as volume, colors and scenery that might give valuable signals about what viewers like.” There is no doubt that we are applying more “intelligent” techniques on the type of analysis performed on the collected data.
I just hope that in a world of generalizations and mass consumer targeting, we don’t forget to empower the small ones, the odd ones, the different ones who aspire to make a difference in the arts, sciences, politics, or any other aspect of life, in a way that doesn’t conform to what the big data analysis tells us.
One response to “Big data analytics determines the next big hit?”
The way people consume Netflix content is very personal, but with over 30 million users averaging around an hour a day we have a lot of data to look at. You can see some of this diversity in the original shows that we have coming out, we started with Lilyhammer, which is partly in Norwegian, and a second series is on the way. House of Cards has been a big hit, next up is Hemlock Grove, vampires for the Twilight tweens, Arrested Development for quirky comedy fans, Orange is the new Black for people who liked Weeds, Derek for fans of British satire, and Turbo for kids. These are all over the map, but we think we have an audience for them all. So far it’s a big experiment, as we figure out what works we feed it back into what we do next and also how much a particular show is worth.
The net effect seems to be that we can take on things that a prime time network looking for a hit show would pass on, we can build an audience more gradually, and we can cover more niche demographics. It’s a fun time to be working at Netflix 🙂