Mickey Mouse in the 21st Century: How machine learning is making Disney World the happiest, and most efficient, place on Earth

A longtime staple for family-friendly vacations, Disney theme parks are using machine learning to add to the magic.

Disney has made its reputation for building its theme parks into “the happiest place on Earth.” Throughout the last six decades, park-goers have enjoyed the advances that Disney has incorporated such as the animatronic robots in Tomorrowland and the motion-simulated Star Tours. In turn, Disney has continued to invest in the customer experience.

With enormous volumes of visitors every year (Disney parks globally had ~150 million visitors in 2017)1, Disney faces a serious challenge in managing extensive back-end and customer-facing operations. At Walt Disney World (WDW) alone, this entails managing hotel accommodations, food service, and theme attractions for over 20 million annual visitors.1 Machine learning can help Disney simultaneously enhance the existing consumer experience, improve internal operations, and push the envelope on product development at a large scale. By integrating this technology, Disney has sought to make its parks among the most efficient places on the Earth.

 

Disney’s use of machine learning

In recent years, Disney rolled out the MagicBands wristbands and My Disney Experience platform to streamline the theme park experience. My Disney Experience helps WDW visitors plan itineraries, restaurant reservations and express passes. Visitors can then load their pertinent data (e.g. room access, ride passes and payment methods) onto the MagicBand, a sleek Disney-themed wristband. This integrated wearable not only simplifies things for consumers, but also gives Disney managers a massive trove of data on visitor behavior. The distributed network of sensors at entrances, food stands, and rides gives Disney an opportunity to use machine learning to drive efficiency and improve guest experience in its parks.

On the front end, Disney uses machine learning to help customers develop itineraries that minimize the time visitors are waiting in line.2 Complex algorithms also help predict key customer events like when and where a visiting family will sit for their dinner reservation. WDW control centers can then prompt kitchen staff to get a head start on the meals the family pre-ordered earlier in the day.2 By reducing idle time, machine learning helps to increase visitor satisfaction and to maximize the time visitors are free to make purchases around the parks.

On the back end, Disney uses advanced algorithms to better plan resource allocations. With continuous updates, park operators can better allocate the 80,000 employees across 240,000 shifts each week.3 Additionally, machine learning helps Disney actively predict and manage demands for costumes and laundry needed to make a magical experience for its guests.4 Lastly, Disney uses dynamic pricing models – a staple in other industries such as airlines – to match demand with park capacity throughout the year.5 Disney uses machine learning to drive operational improvements while delighting customers and maximizing revenue.

Disney also invests in machine learning with longer-term improvement in mind. For example, Disney recently filed a patent for a shoe recognition tool to track traffic patterns in the parks to finer resolution. This should improve the ability down the road to understand and manage visitor flow through the parks.6 Executives have also revealed investments in artificial intelligence to create robots that move around and interact with guests in the parks,7 driving visitor engagement.

Beyond the parks themselves, Disney’s massive investment in its own streaming platform Disney+ opens a new, massive frontier of consumer data that should reap major long-term benefits independently and with the parks. Development of facial recognition software will also help content creators better understand how viewers engage with film.8 Disney can use machine learning to capitalize on investments like these to provide meaningful insight into how to best connect with fans over the long term.

Potential avenues

Disney is setting itself up well for the future with its advanced analytics, and it still has opportunity to get more out of machine learning. Disney should offer customers cross-portfolio targeting between content and experiences. For example, if a family loves Star Wars on Disney+, Disney could provide catered packages to its new Star Wars-themed additions to Disneyland and WDW.

In the parks themselves, Disney can seek further engagement with feedback centers at attractions that are integrated with the MagicBands. This could provide data to support real-time and future park management.

Finally, Disney should make sure to keep focus of what it does best: connecting to fans. Disney’s long history of creativity, whether from Imagineers or filmmakers, has distinguished it as a cultural force for the last century. Machine learning can help Disney understand what works and where opportunities lie, but it will not replace the leaders and creators that have made Disney what it is today.

Open questions

As Disney continues innovating to develop its offerings, two questions highlight the opportunities it faces:

  1. How can Disney best use data sourced from the parks to inform content creation decisions across film and TV?
  2. What limits exist to using findings from machine learning between platforms (e.g. parks, film, TV)?

 

(799 words)

 

[1] John Robinett et al., Theme Index Museum Index 2017 (2018), Themed Entertainment Associates.

[2] Cliff Kuang. “Disney’s $1 Billion Bet on a Magical Wristband,” Wired, March 10, 2015, https://www.wired.com/2015/03/disney-magicband/, accessed November 2018

[3] The Leadership Network, “How Disney Creates Digital Magic with Big Data.” LeadershipNetwork.com, October 5, 2016. https://theleadershipnetwork.com/article/disney-digital-magic-big-data, accessed November 2018

[4] Pete Buczkowski and Hai Chu, “Corporate Profile: How analytics enhance the guest experience at Walt Disney World.” Analytics-magazine.org, October 2012. http://analytics-magazine.org/corporate-profile-how-analytics-enhance-the-guest-experience-at-walt-disney-world/, accessed November 2018

[5] Erich Shwartzel. “Disney Tests Pricing Power at Theme Parks,” Wall Street Journal, June 18, 2018, https://www.wsj.com/articles/disney-tests-pricing-power-at-theme-parks-1529331115, accessed November 2018

[6] Bernard Marr. “Disney Uses Big Data, IoT and Machine Learning to Boost Customer Experience,” Forbes, August 24, 2017, https://www.forbes.com/sites/bernardmarr/2017/08/24/disney-uses-big-data-iot-and-machine-learning-to-boost-customer-experience/#289bf6303387, accessed November 2018

[7] Dave Lee. “SXSW 2017: Disney ‘not in the business of scaring kids!” BBC, March 12, 2017 https://www.bbc.com/news/technology-39245946, accessed November 2018

[8] Zhiwei Dang et al., “Factorized Variational Autoencoders for Modeling Audience Reactions to Movies,” July 21, 2017, IEEE Conference on Computer Vision and Patter Recognition,  “https://www.disneyresearch.com/publication/factorized-variational-autoencoder/

Previous:

As Soon As You Publish A Map It’s Outdated: How Google Maps Uses Imagery And Machine Learning To Keep Their Maps Relevant

Next:

BlackRock: Is the Future of Investing in Machine Learning?

Student comments on Mickey Mouse in the 21st Century: How machine learning is making Disney World the happiest, and most efficient, place on Earth

  1. Great post! I especially like the potential of the shoe recognition tool to better understand visitor patterns and common walking routes. One really cool potential application here could be to identify the most commonly-walked routes, and then actively push visitors to other pre-determined walk routes to control site congestion. Disney could also prescribe a series of routes that the visitors should take once entering the park so that they could completely optimize for traffic at various sites.

  2. Really enjoyed your article, particularly the suggestion that Disney use the data it has gathered to link customer behavior across different platforms such as TV, movies & theme parks. That was my first thought as well, that if someone enjoyed the Star Wars ride they should get a personalized email to go see a new Star Wars movie coming out for example, or an invitation to a Star Wars-themed cruise. The challenge is that Disney’s data collection seems to decrease significantly with the removal of the MagicBands when visitors leave the park. What can Disney do to continue collecting data on its customers even after they leave the premises, specifically to give its AI the closed loop information it needs to improve its recommendations?

  3. This is a great example, and something that a lot of us can relate to! What I found really interesting was how Disney integrates their data collection into the customer experience, making it actually seem really fun for users to give Disney their data. At the center of this is the MagicBand, which visitors actually pay for prior to visiting a park. From a visitor standpoint, using this is really frictionless and, as the name implies, feels like magic. The visitor is incentivized to wear the band and use it throughout their visit, generating high-quality data for Disney at a fraction of the price of other collection mechanisms (especially given the revenue from the bands themselves). There are approaches (such as travel/hospitality companies placing apps on phones) which approximate Disney’s collection mechanism, but none are executed quite as well. Like with many other aspects of the Disney experience, this lesson should be applied to other industries, especially with the proliferation of wearable devices: if executed well, consumers can be highly engaged in the collection of their data, and might even pay you to do so!

  4. Great essay! I am curious about the shoe recognition tool. Also, Disney++ will definitely be a game changer for the Disney’s businesses.
    Especially, I like your thoughts on connecting online and offline through Disney++ (Disney is setting itself up well for the future with its advanced analytics, and it still has opportunity to get more out of machine learning. Disney should offer customers cross-portfolio targeting between content and experiences. For example, if a family loves Star Wars on Disney+, Disney could provide catered packages to its new Star Wars-themed additions to Disneyland and WDW.).

  5. I really enjoyed your post! I recently went to Disney World on a family trip and bought one of these magic bands. It actually syncs to a Disney World app which allows the company to collect even more data on what you do, even after you take off the band (which touches on the question JP raised). With regards to the questions you raised, I think content creation across TV and movies drives the demand for the parks and not the other way around. Additionally, I don’t believe park visitors would be a representative sample of Disney’s consumers, especially since many of Disney’s largest movies are not targeted towards children. I believe these factors are the main issues that would stand in the way of focusing too much on data collected from parks to inform the company’s broader decisions.

  6. Love your idea of Disney leveraging their data across the portfolio to upsell and identify customer types! While Disney has a wide portfolio of entertainment across segments from movies, toys, clothing and the parks etc, the target audience largely remains the same so it would make sense to tighten the connection between their revenue streams. One thing that the data can’t tell them is what the competition is doing. With these methods, they’re only able to gain insight on their own customers but not those you opt for other types of theme park or entertainment. While there’s a lot to be gained from leveraging the data of their own consumers, they should also try to get data on the customers they lose.

Leave a comment