From Revenue Manager to Hotel Data Scientist

We are living in the data-driven era where most innovation in business, medicine, education, and government will come from the application of mathematical models and algorithms to large volumes of data points. Sadly, progress will probably come very slowly as every industry is suffering from a long term problem with no short term solution – a massive analytics talent drought. More specifically, there just aren’t enough people with a mix of math, programming, critical thinking and business skills to help companies unearth the profit hidden in their data. According to a study by McKenzie Research, by the year 2018 the United States alone will face a “shortage of 140,000 to 190,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use the analysis of big data to make effective decisions.”

This shortage will hit the hospitality industry especially hard as it tends to be at the bottom of the totem pole for attracting analytical and technical talent. Unfortunately, hospitality is still perceived as an industry where soft skills are overwhelmingly more important than hard skills. Other service industries, such as retail, have become a lot more data and mathematically savvy in the last decade through supply chain and in-store innovations, but not hospitality. In fact, driving innovation through technology and data analysis is seen by many in the industry as a luxurious exercise that is a “nice to have.” For most hotel execs, it’s better to have the right couch in the lobby than the right data server in the telephone closet, even though many data servers are now cheaper than most couches. Even in the hotel companies that have made substantial investments in Business Intelligence and CRM systems, little attention has been paid to the quality of people who are responsible for extracting insights from those systems. Many hospitality companies have been sold or are paying for expensive intelligence software and hardware on the promise that they are a panacea that begin to solve problems as soon as you install them. Sort of a “set it and forget it” promise.

Here’s the bigger problem, in the next ten years most university graduates with a statistical, data mining, machine learning education will go into academic research, manufacturing, high finance, technology, and life sciences. This will leave little or no advanced analytics talent to serve the needs of service industries, where the proper application of analytics can yield immense value. The big brand hotel companies have already recognized this trend and for the last five to seven years they have embarked on an effort to lure analytics talent to their Revenue Management function. Companies such as Marriott and Disney have realized that hospitality is a data-intensive business and that there is a wealth of creative strategies and tactics that can be found when the data is analyzed by professionals. Yet, even for these big brands the single biggest obstacle to making data-driven progress is their inability to find enough qualified talent to fill their analytics positions. Even at the property level, attracting RM talent with the right combination of technical skills is like pushing a boulder up a mountain. At least I can say confidently that when it comes to analytics talent, big brands do “get it” and they are doing something about it.

In the independent world the story is very different. Most non-branded properties do not even have the most simple BI capabilities, putting them already at least 10 years behind the analytics curve. Even when independents invest in intelligence software, it is an enormous task for them to find the talent to extract value from these tools. Hence the glut of Data Warehouses and CRM systems collecting dust in independent properties all over the world. Further, even if they could find the right person they probably would not be able to afford them as annual salaries for statisticians and data miners now start at $80,000 and average around $145,000. The most glaring manifestation of this talent drought in the independent world is the fact that most properties have relegated their Revenue Management, arguably the most data and mathematically intensive function in a hotel, to employees that are promoted from the front desk and reservations.

Rounding out this perfect storm is the fact that there are only a handful of university hospitality program courses dedicated to data analysis. These, however, are mainly focused on the basic statistical analysis of small data sets. There are no courses dedicated to the real-world problem of mining and analyzing PMS-sized databases. Therefore, the industry can not look to hospitality programs to supply any of the much needed talent for years and maybe decades to come.

I believe that the single best way to tackle this talent shortage in the industry is to take a page from the financial services playbook of the early 80’s. In the 70’s, some forward-looking investors in Wall Street began using mathematical modeling to analyze stocks and bonds. This led to a massive demand for graduates with a mathematical background. The profession was soon formally titled Quantitative Analyst or “Quant”. Later universities began creating programs to fill the specific needs of the finance industry by introducing degrees in financial engineering, mathematical finance, and computational finance. Now, even though there is still a shortage in Wall Street of this type of talent, at least there is a pipeline of trained graduates being supplied by colleges and universities from all around the world.

The development of this formal career path from a math/programming degree to a finance job took 25 years to be realized. In hospitality, we just don’t have that kind of time. The fastest way to engage the talent that we need away from other industries is to create the Hotel Data Scientist career path. Let’s call the title HDS to satisfy those hotel people who love acronym speak. For a definition of Data Scientists I defer to Harvard Business Review: “Data scientists are the people who understand how to fish out answers to important business questions from today’s tsunami of unstructured information.”

This definition comes from an October 2012 Harvard Business Review article titled ‘Data Scientist: The sexiest job of the 21st century’. In a nutshell, Data Scientist is a title that pulls together all the overlapping disciplines that are necessary to extract value from transaction data. In the hotel industry, a Data Scientist would essentially use data to help managers understand the needs and desires of their guests as well as how to manipulate the spending and behavioral patterns that lead to higher profit.

Here’s a short description of a Data Scientist profile once Hotel is added to the title:

A Hotelier at Heart – The Hotel Data Scientist (HDS) needs to have a profound desire to serve, just like the best chefs, concierges, housekeepers, front desk agents, and GMs. The desire to create experiences that WOW the guest must be innate in the HDS ethos. This will drive the HDS to think about and launch data-driven projects that actually add value to the guest stay and are not just academically interesting.
The Most Curious Person on Staff – An HDS never stops asking questions and challenging assumptions. The hotel industry is full of experienced managers who truly believe that their world has not changed much in the last decade and thus they continue to run tactics from the same old, tired playbook. The HDS will always be willing to ask “how do you know what you know?” and “why do you think that will work?”

A Librarian’s Mentality – Without “clean-enough”, “organized-enough” data a hotel will not make much progress with analytics. Unfortunately, most PMS systems are setup for agents to collect data for transactions and not analysis. This means that the hierarchies such as market codes, rate codes, and room classes were usually set up for easy data entry, rather than to provide rich information to managers about guests’ desires. As one example, few hotels in the world collect data on reservation rejections, an invaluable data point for RM. The HDS will define the data that is needed for deep analysis and will then standardize data organization and collection to facilitate information search and mathematical model building.

Silos Buster – Most hotels operate with distinctly separate departments that rarely make decisions together and sometimes their decisions are actually counterproductive. A typical example is when RM runs a discounted rate strategy to drive occupancy and Marketing is off communicating a luxury brand message to drive guest quality. The HDS will interconnect data from different sources to show department heads how their seemingly independent decisions affect each other and what is the financial value or synergistic action.

Statistical Soothsayer – The HDS will have the statistical background to be able to quantify to decision makers, the risk inherent in their decisions. Using historical data, they will be able to show when a decision has worked and under what circumstances it has not. Managers are paid to make decisions and they are rewarded when those decisions work. An HDS can show why a decision paid off or why not, thereby incorporating wisdom into any decision process.

Algorithm Ninja – The weapon of the HDS is the algorithm. The algorithms that make the most impact on the P&L are built on customized programming code. Using creative data manipulation and advanced mathematical computations, the HDS will show data to managers in ways that they have never seen it before. Hotel managers will be able to see the “what ifs” of different decision paths, the mathematically optimal point for any target, and 360 degree views of every guest, just to name a few.

Modeling Monster – Every hotel property is different. Each has its own physical layout, demand patterns, and guest profile mix that contribute to the flow of profit. The best HDS knows how to convert these seemingly random, real world processes into predictable statistical models that search for and suggest optimal decision paths in real time. The information from these models allows hotel managers to make profitable, data-driven decisions.

A Cool Cat – Above all, an HDS will know when to press forward for data-driven progress and when to defer to traditional approaches to decision making. A wise HDS will know when NOT to speak as much as when to speak. In that sense, they must be master communicators. A mathematical approach to decision making is hard for many people to grasp and those same people usually try to hide their lack of comprehension by denying the value of analytics. An HDS must have the tact to keep quiet when they hear a senior executive say that they “don’t believe in the math.” That executive will eventually come around as they all do.

All hotel companies should move to promote the Hotel Data Science profession as quickly as possible even if it is at a manager level within the organization. Doing so will have a big impact on the industry’s ability to begin the process of hiring away enough talented people from other sectors. Then the hotel industry can begin to leverage data to drive the innovations that will allow the industry to reach historical revenue and profit levels over the coming decades.