Data Science in Gaming and Hospitality

Opportunities in Hospitality Data Science

The Promise of Big Data

Today, the hospitality and gaming industry has more data than it knows what to do with. This includes both:

• Traditional structured data (e.g., transaction records, website visits)
• Unstructured data (e.g., social media streams, guest comment cards, video feeds, etc.).

Thanks to cheap storage, mobile technologies, powerful processing and a host of other innovations, hotels and casinos now have access to a treasure trove of information. Insights from big data can be used to deliver on customer service, refine loyalty programs and hone marketing campaigns.

Providing, of course, that companies are willing to invest in the technology and IT expertise required.

Real-Time Revenue Management

In the past, revenue management systems were hindered by a dearth of data points. Not any more. Now data scientists have the ability to tap into a constant flow of real-time pricing data and adjust their offers accordingly.

Take Lisa Terry’s story of Marketspan. To help Kees Hospitality compete with global hotel brands, the revenue management provider LeisureLink created a cloud-based application to handle revenue management, merchandising and electronic distribution.

“Instead of historical views, Marketspan applies its algorithms to a steady stream of real-time pricing data from a variety of sources including OTAs, wholesale packagers, websites and the client’s PMS.” In 2012, the CEO of LeisureLink reported a sales increase close to 94% year-over-year.

The partnership between Denihan Hospitality Group and IBM is a similar success story. In addition to filling rooms during non-peak seasons, Denihan was looking to boost revenue at popular times. IBM’s daily reports enabled the hotel’s team to:

• Predict the most beneficial type of business to book at a given time
• Understand how far in advance it would do so
• Estimate what the room-rate trend would be, which booking channel would be used, and the length of the stay

IBM claims their tools created a 40% boost in productivity for revenue management teams and allowed one of Denihan’s New York City hotels to perform at double the room rate during United Nations Assembly Week.

Satisfaction Guaranteed

Customer satisfaction is huge in the hospitality and gaming industry. Hospitality data scientists are looking at a wide variety of customer touch points – website tracking, line item purchase details, dietary preferences, room temperature settings, guest surveys, mobile apps, social media comments – to create a 360-degree view of patrons.

• Partnering with SAS Analytics, SaskGaming employed predictive data analytics in order to increase demand for slot machine play. Insights from the study allowed SaskGaming to offer the right games, in the right locations, to attract loyal customers.
• Airbnb, in an effort to help their host’s set prices, released Aerosolve – a machine learning platform which helps to mimic hotel and airline pricing.

Keeping it in the Casino

The gaming industry is equally, if not more, involved in customer satisfaction efforts. Casinos have historical and real-time data on everything from restaurant, spa and golf course purchases to time spent on slot machines. Real-time data from the casino floor is constantly fed into databases. Historical data from loyalty programs grows richer every day.

• This allows casinos to individualize the incentives offered to each customer for spending and gambling. To match gambler to experience, for example, Caesars has developed digital interfaces on its games – what Eric Rosenbaum calls “Netflix for gamblers.” If things aren’t going well, the game or a staff member can direct a player to another machine. Casinos may also incorporate data and analysis of their customer data to improve layouts of their floors, increasing the customer relationship value, and foreseeing security threats.q

Fraud Detection

Casinos are also using big data to keep undesirables out. You remember the story of Bill Kaplan and the MIT card-counting team in the late 20th century?

They were partially brought down by Non-Obvious Relationship Awareness (NORA), data matching software that determines connections (e.g., a prior relationship between croupier and player) that aren’t immediately apparent.

Combine this kind of technology with facial recognition data, video analysis, on-the-spot background checks, RFID transmitters in casino chips and you have a recipe for constant surveillance. The boys from Ocean’s Eleven wouldn’t stand a chance.

SoLoMo Marketing

Let’s say a hotel has done its homework and is making guests happy. How can it best attract new ones?

The answer, as data scientists in the Retail Industry will tell you, is personalized marketing. In the past few years, marketers have become wildly excited by multi-touch attribution and the convergence of social, local and mobile (SoLoMo) data.

Twitter comments, geo-location data, mobile apps that monitor your daily behavior – all of these can be combined with offline data to create a detailed profile of your behavior and preferences.

Hospitality and gaming marketers then have the opportunity to supply personalized offers, promotions and services through the channels you prefer. For example, the mobile app Hotel Tonight checks centralized reservation records against a customer’s geo-location data and supplies last-minute booking offers.

This marketing tactic is not limited to newbies. Marketers are also using customer data and feedback to predict the value of returning guests. Denihan, for instance, provides customized offers. Patrons who spend more (e.g., tens of thousands) receive a different deal than those who spend less (e.g., thousands).

Beating the Odds

But it’s in the dynamic world of online gambling that number crunching is proving especially exciting. Data scientists have realized that big data can help online gamblers beat some rapidly changing odds.

For example:

• Claiming to be the largest online poker tournament database in the world, the poker stats firm SharkScope monitors millions of games and players per day. Those who use SharkScope can track their own statistics while avoiding the poker “sharks” who might gobble up their cash.
• The online startup Betegy is employing big data algorithms to calculate the outcome of soccer matches. It offers forecasts for 21 leagues and asserts that its complex mash of data (recent performances, team history, form, weather, player motivation) can successfully predict the outcome of 90 percent of all English Premier League matches.

In this realm, speed is paramount. SharkScope uses a MySQL database from TokuDB that is capable of producing insights for a variety of requests in just under two seconds.

Data Risks and Regulations

The Challenges Ahead

It ain’t all hearts and reservations. Hospitality and gaming companies face plenty of challenges in their push to become data-driven organizations.

For one, many are still grappling with outdated relational databases and with software that can’t handle large volumes of both structured and unstructured data. Not to mention the lightning-fast real-time analysis required.

For another, they’re suffering from a worldwide shortage in data analysts. Business intelligence and CRM systems are well and good, but you also need experts who know how to extract valuable information from them. The hospitality industry, could have a tendency to be on the bottom of the job wish list for data scientists. Finance, biotechnology, research and manufacturing tend to garner their attention first.

Excuse Me, That’s Private

The hospitality and gaming industry must confront another significant issue: individual privacy.

The massive online gaming industry, for example, routinely collects sensitive personal information, including data on behavioral patterns and personal finances. How is this data being stored? How is it being protected? It goes without saying that the security track record of these companies has not always been stellar.

Hotels and casinos are equally at risk. The more they collect, store and share data (e.g., with third-party business intelligence companies), the more they expose themselves to security breaches and abuse of information.

Global hospitality companies are required to comply with a dizzying array of data privacy regulations. The E.U., in particular, has recently enacted strict privacy rules.

Last but far from least is the issue of customer blowback. While some guests may appreciate the “personal” touches made possible by big-data algorithms, others are less impressed. Casinos have been repeatedly criticized for using loyalty programs and behavioral profiling to lure patrons more deeply into gambling than those patrons might have wished.

History of Data Analysis, Gaming and Hospitality

The hospitality industry – that hodgepodge of lodging, restaurants, event planning, theme parks, casinos and cruises – has been collecting data on its customers ever since man could scrawl his mark on a foreign cave. But if we’re going to talk about the birth of modern data science in this field, our first topic must be airlines.

The First Automated Booking System

Travel is to hospitality like Sinatra is to swing. You can’t have one without the other.

In the mid-1940s, American Airlines tackled the problem of booking flights with an experimental automated booking system called Reservisor. It was successful enough to interest Sheraton Hotels, who simply substituted hotel rooms for airline seats.

But it wasn’t perfect. For one thing, it was impossible for individual ticketing agents to directly query the system. Starting in 1953, Trans-Canada Airlines (TCA) began developing a reservation system with remote terminals. In 1962, renamed ReserVec, the new system was ready, complete with a transistorized computer. In 1963, it was in every TCA ticketing office. Queries and bookings could be completed in one second – no remote operators required.

SABRE

The U.S. took its own path. A chance 1953 meeting on a flight from Los Angeles to New York brought two Smiths together. C.R. Smith, the CEO of American Airlines, asked R. Blair Smith, a senior IBM sales representative, if something couldn’t be done about American’s clunky Reservisor system. A new project was born.

The result of the IBM/American partnership was the Semi-Automatic Business Research Environment (SABRE). Sabre had the power to create and manage airline seat reservations and make this data available to any agent at any location instantly. Completed in 1964, it was the largest private, real-time data processing system in the world.

Automated Hotel Bookings

Success in the airline industry prompted hotels to take a closer look at computers. As Jennifer Riesselman notes in her history of electronic booking:

“In the late 1960s, hotels began to see the opportunities that electronic distribution provided and started to develop their own systems based on the airlines’ frameworks. Many hotel brands adopted hotel identifiers based on airline city codes. Some of those identifiers still are in use today.”

Holiday Inn, Marriott and Westin were just a few of the hoteliers that dipped their toes in the data of automated hotel bookings.

Marriott’s Big Data Explosion

Things got hotter in the technology boom of the 1980s. Desktop computers and hotel reservation systems appeared at the front desk. Travel agents learned to put down the phone and peck at the keyboard to make hotel reservations. In 1983, Westin became the first major hotel company to offer reservations and checkout with major credit cards.

But it was another chance meeting – this time between Robert Crandall, current Chairman and CEO of American Airlines, and J.W. “Bill” Marriott – that changed data history forever.

Crandall had been spearheading an effort in Yield Management, maximizing airline revenue through analytics-based inventory control. Marriott saw that Crandall’s big-data system was also a good fit for hospitality, where advance bookings, heated competition and wide swings in supply and demand were always an issue.

So Marriott came up with a Revenue Management System. As Jian Wang defines it:

“Revenue management is a methodology to maximize an enterprise’s total revenue by selling the right product to the right customer at the right price at the right time through the right channel.”

This data-driven system helped Marriott create a daily forecast of demand, offer targeted discounts based on that demand and, eventually, forecast guest booking patterns and optimize room availability.

Travelocity Hits Warp Speed

With the arrival of the Internet, attention shifted from company to customer. There had already been experiments in online access – as early as 1985, Sabre had introduced easySabre®, which allowed customers with personal computers to tap into Sabre’s system of airline, hotel and car reservations – but it was on March 12, 1996 that Travelocity was born.

Travelocity was one of the first websites that put bookings directly into the hands of the end-user. Now there was no need for a travel agent – customers could access Sabre’s fare and schedule information and make reservations for hotels, rental cars, cruises and packaged vacations in a New York minute.

In the process, of course, web administrators could collect large volumes of valuable data on customer behavior. In that same year, Travelocity was joined by Expedia and branded hotel websites.

As the industry crept into the new millennium…

• Valuable search tools appeared (e.g., searching for flights based on fare.
• Reservation systems became increasingly efficient.
• Customer loyalty programs fed directly into giant databases.

In 1998, Priceline launched it’s “Name your own price” model. Six years later, TripAdvisor provided an online forum for hotel reviews and recommendations. The tide of data was rising. Information was everywhere.

Casinos Invest in Data

Everywhere, yes, but not necessarily understood. Enter Gary Loveman.

In 1998, Loveman became COO of Harrah’s License Company. Harrah’s was a player – it was the first casino company to be listed on the New York Stock Exchange – but not a high roller.

• It lacked the ready cash to invest in hotels and shows.
• Customers were generally older folks with small incomes.
• Loyalty cardholders often gambled elsewhere.

So, as Effy Oz points out in his book, Management Information Systems, Loveman invested in data instead. He threw his resources into data warehousing, business intelligence systems and data analysts.

Through a series of experiments, analysts discovered:

• Gift shop discounts weren’t appealing, but deals on hotel rooms were – so deals were created.
• Customers who lived nearby valued chips more than any other incentive – so casino chips were offered to local patrons.
• Customers who enjoyed their visit in one year spent more in the next – so employee incentives were set based on customer satisfaction.

Using SAS software, Harrah’s could even determine when a gambler was about to pack it in for the day:

“When a Total Reward member is about to reach his or her pain threshold, the casino offers a free meal or a ticket to a show to keep her happy and keep her inside the building. Many people believe that winning probabilities are affected by previous winning or losing. Therefore, a staff member may also offer to lock the machines until she is back from the free meal.”

Harrah’s also shattered a long-held tradition. It began offering different incentives to different customers based on their preferences.

And it hit the jackpot. Between 2003 and 2006, the company spent $22 million on big data efforts. In return, Nuclear Research has estimated that Harrah’s earned $208 million in measurable benefits.