Application industry can now be revolutionized by data technology, including the entertainment industry entertainment. Spotify and Netflix are examples. They are both successful examples of the entertainment industry succeeding through proper data analysis.
Netflix was founded in 1997, founded by Reed Hastings together with Marc Randolph. Hastings did his bachelor’s degree in mathematics at Bowdoin College and continued his master’s degree at Stanford University with a study in Artificial Intelligence. Based on his studies and experience, he is trying to revolutionize the business model from Netflix which used to be a DVD rental to become a streaming media service.
Through data and analysis, Netflix licenses its content based on organic viewing behavior vs standard pilot testing and reactionary paradigm. After a content, be it a TV series, documentary or film, is available and ready to be watched on Netflix, Netflix guarantees that the show will suit the audience’s preferences. This is due to the use of engine recommendations that they have. With more than 125 million subscriptions, Netflix has successfully processed the data into a blueprint. We can see the results from how Netflix beat its competitors.
Netflix Chief Content Officer Ted Saran dos and his team use data to guide decisions in different ways. Interestingly, compared to utilizing data based on existing content, Netflix looks at the data as a whole and analyzes how subscribers use the application. That is, Netflix curates and manages data in full, starting from the history of a consumer’s search to find the content he wants to watch, to in-depth understanding to find out user profiles, what shows are watched, to producers and also favorite actors.
Sebastian Wernicke, statistician and Chief Data Scientist of One Logic shares his thoughts on this. He identified patterns that distinguished successful data-based decision-making from unsuccessful decision-making using data. Wernicke suggested that the data and their analysis are only suitable for dissecting a problem and understanding its multi-variable components. To be successful, it is necessary to get industry experts involved in the process.
In the case of Netflix and Amazon, Netflix is first using data to understand its audience in such a comprehensive manner that would be impossible without in-depth analysis. The decision to license House of Cards cannot be categorized as purely data-driven. However, this actually pointed them in the right direction. Instead, Wernicke stated that Amazon uses data all the way through the process to drive its decision making. This shows why Netflix has soared success in pioneering an approach that differs from its competitors.
Another pattern to note is the kind of data that Netflix analyzes against the kind that Amazon does, and how that affects the overall result. Amazon saw the data as a reaction to a controlled subset of pilot episodes while Netflix analyzed organic consumption across its platforms.
There are two things that can be noticed from Wernicke’s analysis of Amazon and Netflix case studies. First, when using data in decision making, it is very important to incorporate industry experts into the process. Second, when predicting what consumers will want to watch or listen to, it is important to look at organic and uncontrolled consumer behavior within the larger ecosystem. According to a survey by financial services firm Cowen and Company, 86% of the 1,200 people surveyed were less likely to cancel their subscription after seeing House of Cards.
By understanding the early successes driven by Netflix’s big data, it shows how its strategy has contributed to the company’s overall success. So, as we have previously mentioned, data-driven also requires creativity, not only purely based on numbers.
Spotify was launched in September 2008 by Swedish startup Spotify AB. Spotify operates under a freemium business model, with two streaming music tiers: Spotify Free (160kbit/s) and Spotify Premium (up to 320kbit/s). The Premium subscription has the advantages of removing ads, improving audio quality and allowing users to download music for offline listening.
Sunita Kaur (Managing Director of Spotify Asia) explained that Indonesian people have spent 1.165 billion minutes enjoying Spotify during their three months of operation in Indonesia. He also added that Spotify users in Indonesia spend an average of 1.5 hours to enjoy the Spotify service every day. The majority of them do it from 12:00 to 16:00, and 20:00 to 23:00.
He could explain those statistics because they don’t just provide services for music streaming, but also pay attention to the characteristics of the listener.
It’s been a while since entering Indonesia, Spotify continues to pamper its subscribers with various creative features. One of them is Discover weekly.
Not only that, Spotify is also creative in sorting out the song genres that its subscribers usually hear in the Daily Mix feature. As for the Daily Mix playlist, here are the 3 main sources of data used:
- Playlists (information about what music is treated as similar),
• Individual hearing history (patterns in the order in which music is played),
• like/dislike/skip from station (information about how songs are joined together).
Spotify uses various collaborative filtering methods (artist-user decomposition, user-song matrices compacted by finding patterns in the data) that provide latent representations (vector numbers) of each user and each artist/song. This allows Spotify to create a comprehensive music map, where similar songs are positioned together and listeners tend to listen to music only in certain areas of that map.
Through such mapping, Spotify can find songs that are close, but new, relative to the preferences you indicate (e.g. Find Weekly, suggested songs to playlists) or to a given request (e.g. radio stations from playlists, songs, artist, etc.).
To summarize, most of Spotify’s navigation features are the result of data processing and creativity. Where innovations are born based on the individual data of each user.
These are the two companies that implement data driven in their business units so that they are more up-to-date and able to lead market share in industry.
An example of implementing Big Data in the entertainment industry – Disney is an international entertainment company that is well-known for their films and theme parks, the number of their employees who focus on developing big data analytics is estimated at around 1000 people, interestingly this organization uses an experimental approach in this big data analytics, Teddy Benson, Head of solution Integration Disney said that they use a proof of concept almost like a startup, Disney uses a small budget to do some experiments and see the risks.
Optimizing logistics using the Magic band
In 2013, Disney released the “Magic Band” which is estimated to have spent more than 1 billion dollars for development and implementation. This Magic band is made using RFID technology that can interact with thousands of sensors that have been placed throughout the rides. This magic band functions as access to hotel rooms, rides and attractions and is used for payment.
When entering a Disneyland vehicle, the officer gives a magic band to each visitor. This magic band can provide data to the company regarding their individual visitors such as visitor information, waiting time in line, length of time using the rides, as well as which rides and attractions are used by visitors , as well as various activities carried out by visitors.
With the data generated by the magic band, the management of Disneyland can set strategies for which rides need to be improved and also discontinued, add new features or attractions, add points where public facilities and logistics such as restaurants and toilets, predict which characters are the most popular. favorite among children as well as various other strategies that can increase visitor comfort and also improve operational efficiency.
Optimizing ticket prices: Lion King on Broadway
The unprecedented success of Lion King on the show, reflecting the importance of using data more effectively, the show has generated an estimated $8 billion more in revenue than any other show. Disney uses historical data on Broadway ticket sales to accurately forecast future demand for the Lion King show. The model they have developed can also predict the highest price for a ticket that a customer will accept.
Optimizing film production using Affective Artificial Intelligence (AI)
When Disney produces a film, an important step they take is testing for market fit. Before Disney implements Big Data, Disney will display film previews in a focus group and try to implement this feedback into the film before the film is released to the general public, this process is too complicated and prone to human error. Then Disney created technology to make the market fit testing process faster and more accurate.
Disney was one of the first companies to adopt affective artificial intelligence, Affective AI focuses on identifying and interpreting human emotions. In a research paper published by Disney and Caltech, explains that they put lots of cameras in cinemas that aim to monitor every face of the audience. With Effective AI technology, the system identifies and evaluates every response shown on the audience’s face to scenes in the film so that it can be used as input material to determine various scenes in the film.
In the future Disney can use Big Data and Affective AI to analyze the emotions of visitors when they visit Disneyland or even determine the ending of the film that best suits the audience in real-time based on the predictions and also the emotions of the audience at that time so as to make customers satisfied with the movies they watch.