Starwood’s early bet on machine learning solutions
The recent explosion in use of artificial intelligence and data analytics in the hoteling industry is no coincidence. In the past decade, the proliferation of digital technology has raised consumer expectations of convenience. Hotel brands face mounting pressure from data-driven disintermediaries – like Expedia or Priceline – and non-traditional substitutes – most notably Airbnb. Accordingly, revenue growth has slowed for the industry and hotel booking windows are shrinking.
To compete, hotels have had to lean into the new digital world. One could argue that no major brand has done that better than Starwood, which was acquired by Marriott in 2016. Starwood’s strategy has long included building a competitive advantage in digital expertise as a means of driving guest loyalty. It was among the first to introduce innovations such as smartphone check-in, key-less entry, and robot bellhops to its suite of hotels.
In 2014, Starwood invested over $50 million to develop ROS, a state-of-the-art automated dynamic pricing system. While dynamic pricing is now standard in the industry, ROS is unique in its comprehensiveness and machine learning orientation. It automatically calibrates pricing based on hundreds of variables, including booking/ cancellation patterns, competitor pricing, weather and climate, and booking patterns on other sites. And, as ROS learns, it prices more efficiently. Starwood has credited ROS with a 20% improvement in pricing since its development.
Starwood’s machine learning foray into customer engagement notably resulted in ChatBotlr, a chatbot that gives guests the ability to make service requests via text message. According to Marriott, “by leveraging natural language understanding and machine learning, ChatBotlr gets smarter the more it interacts with guests. Early findings show that 2 out of 3 Aloft guests are interacting or making requests with ChatBotlr and the service has [reached] a five-second response time.”
Marriott’s challenge and path forward
Since acquiring Starwood, Marriott has doubled down on the Starwood’s investment in machine learning and IOT (Internet of Things) architecture. Their strategy, according to then Senior Vice President of Digital George Corbin, highlights a mobile-first mentality to personalization of the guest experience and real-time engagement. Marriott’s short-term focus involves using artificial intelligence to anticipate what loyalty members want in their hotel stays and reacting real-time through a messaging system called mPlaces.  The medium-term focus appears to center around pulling in even more data to drive personalization and responsiveness. In 2017, Corbin highlighted what the future could look like for Marriot. 
[Marriot could pull in] information that … from your own interaction patterns with us, as well as some intent-based data from how you came to us, where you came to us from, what you were looking for there, and then leverage[e] that to dynamically adjust some of the content and the visuals and the messaging that you now begin to see in the app. … If you think about Google or Siri … the phone knows to listen to you. Our app could leverage that maybe and then direct us to a particular search or amenity or hotel. We would do that in those environments at the request of the member to develop the hotel room of the future. If consumers are being accustomed to talk to their homes they will expect to be able to talk to their rooms.[sic]”
A bright but cautious future ahead
In this author’s opinion, Marriott’s strategy is sound. By leveraging learnings from Starwood across its portfolio of hotels, Marriott is creating the revenue synergies I hoped to see from the acquisition. Further, research supports Marriott/Starwood’s thesis around the ability machine learning and dynamic customer engagement to create unique value in the hospitality space.  However, Marriott’s focus on personalization should not overshadow other ways in which machine learning can create value for customers; two areas come to mind. First, more can be done by Marriott/Starwood in using machines to optimize the use of time by its staff in everything from the check-in process to post-stay engagement. As Frank Reeves, co-founder of hotel technology company Avvio, notes “Little things, such as [dynamically] estimating when to refill soaps, can reduce time spent by staff knocking on doors and improve the guest experience. Secondly, they should explore using machine learning to expand how they think about their customers. Nascent research points the ability to use machine learning technology to create more dynamic clusters of guest types that evolve over time. Such capabilities would improve the ability of Marriot to segment and target customers over the long run.
The question remains for Marriott/Starwood: In a space that is increasingly developing these digital capabilities – where Hilton is partnering with IBM to introduce robot concierges, for example – how do you maintain a defensible competitive advantage in technology innovation? Only time will tell.
 CBRE 2018 Trends in the Hotel Industry. https://www.cbre.com/research-and-reports/2018-U-S–Hotel-Outlook. Accessed 13 Nov 2018.
 Boulton, Clint (2018). Starwood taps machine learning to dynamically price hotel rooms. CIO magazine. Accessed 13 Nov 2018. https://www.cio.com/article/3070384/analytics/starwood-taps-machine-learning-to-dynamically-price-hotel-rooms.html l
 Vondrasek, Mark (2015). Redefining service innovation at Starwood. McKinsey & Co. Accessed 13 Nov 2018. https://www.mckinsey.com/business-functions/operations/our-insights/redefining-service-innovation-at-starwood
 Boulton 2018
 Alexsoft (2018). How the Hospitality Industry Uses Performance-enhancing Artificial Intelligence and Data Science. Accessed 13 Nov 2018. https://www.altexsoft.com/blog/datascience/how-the-hospitality-industry-uses-performance-enhancing-artificial-intelligence-and-data-science/
 Boulton 2018
 Norton, Steve (2015). Starwood Hotels Using Big Data to Boost Revenue. Wall Street Journal. Accessed 13 Nov 2018. https://blogs.wsj.com/cio/2015/02/10/starwood-hotels-using-big-data-to-boost-revenue/
 Marriot International 2017. Marriott International’s AI-powered Chatbots on Facebook Messenger and Slack, and Aloft’s ChatBotlr, Simplify Travel for Guests Throughout Their Journey. Accessed 13 Nov 2018. http://news.marriott.com/2017/09/marriott-internationals-ai-powered-chatbots-facebook-messenger-slack-alofts-chatbotlr-simplify-travel-guests-throughout-journey/
 Marriot International 2017. Marriott Reimagines Its Mobile App To Meet The Needs Of Modern World Travelers. Accessed 13 Nov 2018. https://news.marriott.com/2017/02/marriott-reimagines-mobile-app-meet-needs-modern-world-travelers/
 Ting, Deanna (2017). What Marriott Learned From Starwood’s Loyalty and Digital Expertise. Skift. Accessed 13 Nov 2018. https://skift.com/2017/02/13/what-marriott-learned-from-starwoods-loyalty-and-digital-expertise/
 Ting 2017
 Aluri, A., Price, B., & Mcintyre, N. (2018). Using Machine Learning to Cocreate Value through Dynamic Customer Engagement in a Brand Loyalty Program. Journal of Hospitality & Tourism Research.
 Reeves, Frank 2018. How Artificial Intelligence will bring the human touch back to hotels. Travel News Daily. https://www.traveldailynews.com/post/how-artificial-intelligence-will-bring-the-human-touch-back-to-hotels
 Hilton 2016. Hilton And IBM Pilot “Connie,” The World’s First Watson-Enabled Hotel Concierge. http://newsroom.hilton.com/corporate/news/hilton-and-ibm-pilot-connie-the-worlds-first-watsonenabled-hotel-concierge