People familiar with the entertainment industry are also familiar with the phrase “Hit maker”! So, who is a hit maker? A person who delivers one hit show after the other. But people familiar with the entertainment industry also know how difficult it is to deliver a hit. Who knows what will click with the audience? Directors, producers, playwrights spend days conceptualizing a content, only to end up receiving 1 star from the critic and no returns from the collections.
Data and Netflix
Then comes Netflix. And their revolutionary use of Big Data Analytics. From predicting the kind of content that would garner high viewership to recommending content to specific users, Netflix uses data everywhere. In fact, since its days of being a DVD-by-mail service Netflix placed prime importance on collecting user data and building a recommendation system. Cinematch was the first algorithm behind their recommendation system. After launching their streaming media service in 2007, it took them 6 years to collect enough data to predict the sure-shot success of their first original production “House of Cards”. Netflix set the biggest example of how analytics used in the right direction can literally spell success for a business, in a domain as unpredictable as content production.
How does Netflix gather and use data
The user base of Netflix currently stands at 99 million approximately. So, you can imagine the huge volume of behavioral data that it receives. To collect this data, they create certain data points known as “events” in the world of Big Data Analytics. E.g.
- When does a user watch a show
- Where do they watch it
- On which device do they watch
- Do the nature of shows vary with the device
- When do they pause a program
- Do they re-watch any portion of a program
- Do they skip the credits or not
- The ratings
- The searches
Netflix’s data scientists gather all these information and process them to reveal useful insights. These insights help Netflix in taking almost all of its business decisions.
The House of Cards Saga
Netflix observed that viewers who loved the original BBC miniseries of House of Cards were either fans of Kevin Spacey films or liked movies directed by David Fincher (most of them watched Fincher’s film “The Social Network” from the beginning to the end).
They dug up the numbers lying at the intersection of these watching patterns. Turned out that the numbers were pretty huge. Thus, they easily predicted that House of Cards would be a runaway hit. Now we know how right they were! Not only for content creation, they also used analytics to market the show. They identified different sections of audience and made 10 different trailers to target these sections. E.g. a woman who loves watching female-centric drama but is not a David Fincher fan watched a trailer with all female characters of the show. On the flipside, a mid-twenties film buff watched a trailer featuring David Fincher. This tactic served a big purpose very subtly – different sects of the audience felt a “need” to watch the show without being inundated with the show’s promotions. If you are wondering how Netflix gathered all the individual data, then think about this:
- The details you provided while signing up
- The movies/series you said you liked
- Are you searching for a particular actor or director?
- Are you repeatedly watching a particular genre?
- Are people in your age bracket or in your zip code watching a particular show, at a particular time on a particular day of the week?
Behind these normal actions of yours, lie the DNA of Netflix’s functionality – tons and tons of DATA!
The decision-making Data Framework
Data accumulated from numerous sources influence decisions regarding shows. Not only user data, Netflix also observe data generated by piracy sites. “Prison Break” is a hit show on that front. The image below gives an overview of the data analytics framework at Netflix.
Aegisthus is a platform built buy Netflix’s engineers to work on Hadoop MapReduce in order to convert Cassandra’s SSTables into query-able formats.
How does Netflix benefit from using all the analytics
Analytics has given Netflix an edge over all its competitors! In fact, Netflix is one of the few content generation companies who use data and analytics at this scale and that too for content production and acquisition. Most of its competitors use analytics for content promotion. The benefits that it draws from using analytics are
- It has been able to do away with the concept of pilot episodes. Pilot episodes are sample episodes for new program concepts. Focus groups and experts then watch these episodes to determine the show’s chance of being successful. But since Netflix uses data to analyze the likelihood of a show’s success, it doesn’t need to shoot pilot episodes. That reduces their expenditure by a good amount.
- The success rates for Netflix’s original shows are 80% as compared to the 30%-40% success rates of traditional TV shows.
- Using data to target advertisements and recommend shows to specific users has helped Netflix in lowering its promotional campaign budgets. It has also helped in providing the user with a more personalized experience. As they themselves say, about 75% of Netflix’s viewing is driven by the recommendation algorithm.
For a company which works in an out-and-out creative domain like content generation, it is extremely difficult to make a choice. Sometimes the gut feeling starts speaking and despite all the data indicating in one direction, you understand that the content is just not good enough. And, although you can have all the correct ingredients, the recipe for success is finally dependent on the cook. So without a good production, no amount of predictive analytics can save a show. And finally, not every analysis is correct! E.g. Analytics couldn’t predict that “Breaking Bad” would be as big a success as it was. Using analytics always helps a business in taking informed decisions. However, data can’t save a bad idea, it can only validate a good one.