Netflix : Leveraging the Power of Analytics

To Bid or Not to Bid : House of Cards

In 2011, Netflix CEO Reed Hastings, was about to undertake the biggest gamble in the company’s history. Netflix was to participate in the auction for Media Rights Capital’s drama series House of Cards, produced and directed by David Fincher and exec produced by and starring Kevin Spacey. Behind the decision to participate was his bold vision to transition Netflix from being a channel which distributed video to becoming an original content producer. In the process, Netflix would be competing with the likes of industry veterans like HBO & AMC.

During the auction Netflix won the rights by bidding an astounding $ 100 million for 26 episodes of House of Cards (13 episodes per season & 2 seasons). In an industry where the fate of the second season depends on the success of 1st season, the terms of the deal were unprecedented to say the least. To understand what gave Reed Hastings the confidence to make this bid, it is important to understand the fact that Netflix is first and foremost a software company and therefore data analytics play a very crucial role in the DNA of the company. Reed Hastings was already a successful tech entrepreneur before he started Netflix and he brought all his knowledge of managing a technology company to Netflix.

One of the ways most well known ways that data analytics helped Netflix is by improving the predictability of titles. Netflix launched a competition, The Netflix Prize to develop the best collaborative filtering algorithm to predict user ratings for films, based on previous ratings without any other information about the users or films, i.e. without the users or the films being identified except by numbers assigned for the contest. Through this competition, Netflix could improve the accuracy of its predictions by more than a 1000 basis points (1).

The Netflix team embedded the use of A/B testing to successfully design its website & test its streaming services. What makes this achievement notable, is the fact Netflix is experienced through an extremely diverse set of channels – Broadband, 4G wireless, 3G wireless etc.

Netflix was able to capitalize on its deep analytics strength and its direct feedback from millions of users around the world while it contemplated the decision to bid for House of Cards. Its analytics team had 3 key data points which were instrumental in making the decision:

  1. The British version of “House of Cards” which was already available for viewing on Netflix was well watched.
  2. Those who watched the British version “House of Cards” also watched Kevin Spacey films and/or films directed by David Fincher.
  3. A lot of users watched the David Fincher directed movie The Social Network from beginning to end.

Netflix ended up winning the auction and House of Cards became the first of many blockbuster series that have so far premiered on Netflix. Data analytics was also a key factor when Reed Hastings planned to switch resources almost exclusively to the streaming part of the business from the DVD rental business.

These incidents highlight the competitive advantage that Data analytics has brought to Netflix. It also emphasizes the fact that companies with superior analytical ability and superior access to customer data will become very hard to beat.

(1) Wikipedia : Netflix


Bikes, Data and the Crowd

Student comments on Netflix : Leveraging the Power of Analytics

  1. Netflix is an interesting example of data analytics, thanks for sharing. So far algorithms, like the ones used by Netflix, are key to suggest the right content for a particular consumer, based on previous watching patterns. I am curious how successful data can be in predicting hits in creative content. When it comes to bidding on shows, in entertainment it often proves more prudent to bet on best-in-class talent pool. What would have happened if the data suggested betting on House of Cards even if Kevin Spacey and Fincher were not in the mix? Netflix would probably not bet because Hollywood is increasingly dependent on stars, both for talent acquisition and promotion. Names like Fincher bring groups of talented professionals with them and open many doors. Nevertheless, it will be interesting to see if more content distributors decide to be guided by data analytics when making bets on content.

    1. Hi Ola,
      Thanks for your comment. You are absolutely right that their success rate with a completely new series is still a question, on one hand House of Cards was a big hit, on the other hand Marco Polo was a big flop.

      What is without doubt, is that they’ve got analytics and computer science in their DNA, so the long term success of Netflix will be a good indicator of the long term competitive edge that Big Data provides.

  2. Great Piece! I agree in that the $100M bid for house of cards rights was a significant investment for the company. The fact that the decision was supported by data is astounding, as is the success of the show. Yet I wonder if Netflix if capturing equivalent value from the series. Have their revenues increased as a result of their original content? are subscribers hanging on to their subscriptions for longer than they would were original shows not available? How is Netflix measuring their return on investment?

    It would be interesting to know whether Netflix original content strategy is about new member acquisition, current member retention, or both.

    1. Hey Madrid,

      Thanks for your comment, you correctly note that while House of Cards was successful, its profitability is questionable. In fact Netflix as a whole is earning very little profit compared to HBO, but it is valued many times higher than HBO. In my view, with the current strategy, Netflix is unlikely to earn huge profits, what they have proven to the investors is that they have great execution skills, which should attract the long term investors ( similar to Amazon)

  3. Great article – Netflix is a great example of a company successfully leveraging analytics to improve its business.

    In addition to extensive data crunching prior to its House of Cards bid – Netflix tracks a vast number of events when you are watching a show, such as:
    – When you pause, rewind, fast forward
    – Browsing and scrolling behavior
    – When you leave the show
    – Where you watch from

    The “Danger” of this is that you are giving the customer exactly what they want. If e.g. 80% of users fast forward during the monologues in House of Cards, they might be skipped in the next season.

    What does this do to creativity, just using data on consumers to decide what to produce? What if consumers do not know what they want before they have seen it? I hope Netflix’s data driven approach to TV shows does not kill creativity.

    1. Great comment! With creative stuff, one can never be sure, and therefore I am still not convinced that Netflix will be able to consistently hit home runs (shows like House of Cards), though they will be able to produce shows that are consistently successful. The other big question is their complete reliance on Amazon Web Services for infrastructure, which has become a competitor (a lousy one for the moment!!)

  4. Thanks for sharing this! I really love House of Cards and we actually had a case in SMICI on the series and the creative minds behind it. I would be interested to hear your perspective on how Netflix uses this model to generate content going forward as well. Is this data analytics model replicable repeatedly? What if the series doesn’t have an analog somewhere else for example. Or if they aren’t betting on superstar talent to start in the series? Just wondering if this is something that can fully be attributed to Netflix’s superior algorithms or if Kevin Spacey + the tried and true series model could be largely responsible.

    1. Hey, thanks for your comment!

      In my view Netflix has the following strengths:
      1. A great analytics team. Netflix has a computer science lineage which is reflected in their decision making process in entertainment & engineering – the algorithm that they used to predict which DVD you’d like based on just 3 previous DVDs is a landmark algorithm.
      2. A very relevant & extensive database to analyze and make predictions of of.
      3. They have a very astute management which takes decisions ahead of technological trends (their decision to switch from DVDs to video streaming is in the same league as the Steve Jobs ditching the keyboard for the iPhone)….TiVo/Blockbuster which faced similar challenges screwed it up big time.

      I would hesitate to say that their strengths will always translate into home runs. The creative nature of the business is inherently unpredictable! (see the related post on Relativity Media)
      Netflix has had its fair share of hits (House of Cards) & misses (Marco Polo). On balance, I expect them to do better than competition if they can sustain their strengths. The biggest challenge that they face is the threat of disintermediation by content providers and the rise of Amazon video, who is also their hardware supplier!

  5. Interesting article! I didn’t realize how much they had paid for house of cards, nor the way in which data was used to determine success of the series. I wonder if Netflix could utilize this software advantage to start their own film financing studio. Relatively Media was started about 10 years ago – they were successful in choosing and financing profitable movies. They determined the probability of a movies success using in depth data analysis. They were one of the first companies to utilize this predictive algorithm, as opposed to just a “gut feeling.” Netflix has much more data than Relatively, so they’re in a position to accomplish the same task – but better.

  6. Great post about a TV series that is definitely one huge success for Netflix. Although, I wonder to which extent data can help make such decision to bid so much money on a series. Indeed, I assume that although the British version of House of cards was well watched, it must not have been a majority of Netflix customers. Hence, the population from which is extracted the data might have a rather high error rate. Was Netflix extremely confident in its data analytics metrics, or did it simply take a risky bet? In any case, it definitely was a smart decision.

  7. Really interesting how they leveraged both crowdsourcing and analytics to make that decision. Though as othehrs have pointed out, I am unsure if it was good analytics and some luck or more luck than skill… It will be interestin to see how Neflix use analytics for more systematic decision taking on shows going forward.

  8. Thanks for sharing MaskOfZorro. I wonder if they used data analytics to predict additional subscriptions due to House of Cards being exclusively on Netflix. Then they could presumably calculate the maximum price they are willing to pay and also calculate payback time for the investment. Also, I wonder why the third data point was relevant if they were already considering films directed by David Fincher in the second data points.
    I agree that because they are an internet based streaming company, they have a competitive edge in the richness of data over companies like HBO or AMC which is on cable and can not collect the amount of data Netflix can. Now it depends on weather they can create and capture value through making the data insightful and executing.

  9. Interesting post. There are many in the creative industries that say that content production is a “gut feel” and not driven by empirical data. I think this is a great case to challenge that. Data analytics could be used in a number of ways – primarily to make the filter for content production and delivery more efficient and effective.

  10. great post,
    I think one of the challenges NEtflix faces is “being late” with the content presented.
    Their competitors are much quicker to display the latest episodes and seasons.
    This is something Netflix must resolve to become a true market leader.

  11. Great post. Do you think that this data algorithm is fully applicable to each bid that Netflix goes for? If so, how would you modify or incorporate variables to make sure you don´t get locked up in the factor that made this possible going forward? There could be a bunch of factors adding into this result. Definitely the approach proved succesful for Netflix, but do we know for example how much did the others offer.

Leave a comment