The Data Behind Streaming Success
When you watch streaming content, you are looking at the product of big data. It takes huge amounts of data to deliver and tabulate the content, and it takes algorithms to plan which content to deliver. The biggest success story to emerge from that big data model for entertainment is Netflix.
Though Netflix’s original business model was based on mailing out the DVDs people select for rental, its customers now opt for streaming. And that’s where big data and cloud computing become essential. Netflix explains its approach to evaluating content and deciding on renewal:
- We utilize detailed statistical models to determine expected hours of viewing for each piece of content over its license period.
- We compare cost per hour viewed against other “like” content deals (i.e. exclusive versus non-exclusive, TV versus movies, etc.)
- We look for high engagement and cost efficiency.
- For renewals, we look to renew content that performs well (based on hours generated relative to the cost) and do not renew content where the price doesn’t make sense relative to the value generated.
In the interest of improving efficiency on that end, Netflix transferred its holdings to Amazon’s cloud. It also started using Hadoop, “to run massive data analyses, such as graphing traffic patterns for every type of device across multiple markets.” That helps plan for improved data transmission and better understanding of the customer.
In addition to using big data solutions for delivery of content, Netflix applies algorithms to predict what their customers would likely want to watch next. This type of data mining technology makes Netflix confident that it can handle hosting original content. In fact, it bet more than $100 million on it when it first paid for the rights to two seasons of House of Cards, one of several original content series it planned to stream.
That bet paid off in more ways than one. Not only was the show itself successful, but it paved the way for other successful shows. That is not just a function of what people come to associate with the Netflix brand but of its own data on what works. Netflix accounts for it in this way:
When we started with original content we didn’t have specific data about viewing patterns over time for content that premieres on Netflix. We decided to use straight line amortization based on our experience with TV series from other networks. Now we have more specific viewing data for original content which shows more viewing in the early months of a show’s debut, so we are accelerating the amortization of such content commensurately.
Hits beget hits, so long as you have the data and know how to apply it to planning programming. That’s entertainment – big data style.
Comments are closed