Machine Learning at Disney: Solving Happiness

In the fiercely competitive entertainment industry, Disney is using machine learning to predict what will bring us joy

“What really gets us excited about machine learning’s ability, is to deepen the connection between consumers and our stories through personalization, optimization, and enhanced experience…Instead of using time-consuming, error-prone methods, we’re leveraging machine learning to analyze data, provide real-time recognition and experiences that increase engagement and satisfaction.”

– Michael White, Senior VP of Technology at Disney Consumer Products and Interactive Media[1]



For over 95 years, Disney has differentiated itself in the entertainment industry through unique customer experiences and delightful content. Its marquee brands and world-renowned theme parks generated over $55 billion in revenue last year, representing a steady 4.5% CAGR over the last decade.[2] But as competition for customer eyes and wallets has intensified, Disney has recognized the need to innovate to stay competitive – and they are leaning on machine learning to serve those customer needs.


History of Machine Learning at Disney

Disney approaches technological development using both empirical and practical methods. Since 2008, the Company has supported its own in-house research lab (Disney Research) that develops scientific and technological innovations to drive value across the company’s portfolio. The initiative is a collaboration between Disney and the Swiss Federal Institute of Technology Zurich and publishes on a number of computational fields including machine learning.[3]

One popular example of the implementation of Disney’s machine learning capabilities is the development of MagicBand – a park-wide pass that tracks customers’ movements, analyzes purchasing habits, and reports the real-time data to Disney. The purpose of this innovation was to streamline the customer experience by helping Disney ‘cast members’ anticipate customer behavior and make quick decisions around adding staff or incentivizing guests to head to another ride or attraction. This investment has been highly accretive for Disney: efficiency and capacity of the parks has improved, operating margins have increased by over 450 bps, and traffic trends continue to be strong.[4]


Machine Learning: Reinforcing Disney’s Competitive Advantages

Going forward, Disney will face two key competitive problems: The first is content creation: the proliferation and saturation of content means that Disney must become experts at predicting and adapting to consumer tastes and preferences. The second is content delivery: adoption of new technologies is shifting patterns of content consumption, so Disney must ensure they are delivering content through the most relevant channels and delighting users with every interaction.

Disney is working to extend its technological capabilities further by applying its expertise to content generation. In 2017 Disney Research published a study exploring computational models for narrative building. Using the ‘Q&A’ social media site Quora, Disney collected over 50,000 answers and used machine learning to build models for ‘good’ answers. Using good answers as a proxy for good narratives, Disney has effectively created a program that recognizes patterns of good stories and predicts what will resonate with consumers.[5] The Company is also developing a variety of machine learning-enabled content projects: research on dynamic storytelling between robots and children, programs that can analyze viewer facial expressions to predict engagement with content, and deep learning speech animation processes that can be integrated into existing production pipelines.[6]

Disney is also striving to optimize content delivery using machine learning. Over the next few years, Disney will launch two direct-to-consumer streaming services (multi-sports and Disney-branded film/TV content) with machine learning-driven recommendation engines to thoughtfully push content to consumers.[7] Additionally, the Company invests in smaller projects to enhance and learn from customer interactions, including chatbots for characters such as Miss Piggy, Judy Hopps of Zootopia, and Guardians of the Galaxy.[8]

In addition to the projects Disney has already undertaken, there are several competencies that Disney should build using their machine learning capabilities. One is a command of demographic segmenting and positioning – understanding how to best develop content that serves a spectrum of consumers. Alternatively, given the new focus on digital content Disney should develop innovative testing procedures for their DTC platform to make sure they are consistently producing and pushing the best products.


Looking to the Future

It is imperative that Disney innovate to stay ahead of their competition. However, the Company’s shifting competitive strategy may inhibit their technological progress. Disney recently received approval to consummate the $71 billion acquisition of Twenty First Century Fox. This merger may prove to be a distraction, as it will dramatically increase Disney’s international footprint, expand the content and distribution for its DTC offerings, and increase its ownership stake in Hulu. Furthermore, the Company is undergoing a significant reorganization to prepare for the integration and strategic realignment.

These developments raise several key questions: How should the company allocate its time and resources between the Fox integration and R&D? And if there are dollars available for R&D, which areas of technological exploration and development are most relevant to Disney today? (780 words)




[1] Magento. (2018). What can Disney teach us about Artificial Intelligence? | Magento. [online] Available at: [Accessed 8 Nov. 2018].

[2] The Walt Disney Company. (2017). Form 10-K 2017. Retrieved from SEC EDGAR website

[3] Disney Research. (2018). Research Areas. [online] Available at:

[4] The Walt Disney Company. (2018). Disney’s Q2 FY15 Earnings Results Webcast – The Walt Disney Company. [online] Available at:

[5] Wang, T; Chen, P.; Li, B. (2017). Predicting the Quality of Short Narratives from Social Media. working paper, University of Massachusetts Boston,

[6] Disney Research. (2018). Machine Learning Research Areas. [online] Available at:

[7] Woodie, A. (2018). How Disney Built a Pipeline for Streaming Analytics. [online] Datanami. Available at:

[8] Rayo, E. (2018). AI at Disney, Viacom, and Other Entertainment Giants. [online] TechEmergence. Available at:


Apple – Transforming the future of Healthcare


SoundCloud’s Open Innovation Approach to Reinventing Itself: Survival Instinct?

Student comments on Machine Learning at Disney: Solving Happiness

  1. I’m really interested in your point about how Disney can leverage innovation and machine learning to develop content for a wider spectrum of consumers. Interpreting “wide spectrum” as “more diverse”, I wonder what data Disney already has access to that it could use as inputs for this model. Perhaps it could use data from sibling networks in the Disney family, like Freeform, that have been piloting multicultural programming for the last few years. I also wonder if there is an opportunity, through customer loyalty accounts that power “magic bands”, to connect their off-screen behavior in parks with on-screen behavior for their new programs & delivery channels.

  2. MagicBands are a great innovation that clearly improves the customer experience and benefits Disney with rich data. One additional machine learning application of MagicBands that came to mind was if Disney could use them to make recommendations to guests to further improve their park experience. One example would be to notify a park guest if the line for a certain ride is particularly short, and that the visitor would benefit by visiting that ride at the present moment instead of waiting until later. Another traffic control-related recommendation could be to implement “pop-up sales” at certain gift shops if lines got too long, as a way to encourage visitors to “spread out” and decrease congestion throughout the day. In this example, Disney’s software could monitor traffic around the park and offer perks at non-congested areas in real-time.

  3. There are a number of companies working on learning algorithms to predict when people will be hungry. I wonder if there is space for Disney to work on this and have food freshly prepared for its guests as they walk into its various food shops across the park. This could be a nice way to enhance the customer experience and reduce congestion in the park.

  4. As technology improves, I wonder if we could also use the MagicBands to gather bio-metric data from consumers. Obviously the collection of this data would require Disney to jump through massive legal hoops, but if we could capture heart rate or temperature we may be able to predict a customer’s emotional response as they interact with the park. Disney is all about creating a “magical moments” – if we could define what makes a moment special by observing bio-metric data we may be able to design a more emotional experience for park goers.

Leave a comment