Science Based Hotel Revenue Management Surpasses Fiction of Rules Based Models
“Science is a way of thinking much more than it is a body of knowledge. ~ Carl Sagan ~
When it comes to developing pricing strategies, dealing with the competition, and forecasting future demand, the job of a hotel revenue manager can feel a lot like a kayaker who is continuously trying to paddle upstream.
NB: This is an article by Dan Skodol, VP Revenue Analytics at Rainmaker
Over the past few decades, increasing access to data and advancements in technology have led today’s hoteliers to a notable inflection point when it comes to choosing revenue management systems (RMS). There is an array of RMS along with other less-automated approaches to revenue management available to help hotels improve profits, and they tend to fall into one of two categories: rules-based or science-based.
In today’s competitive market, hoteliers need a little more science and a little less fiction – because selecting the right RMS can mean the difference between profitability and insolvency.
WHY RULES-BASED RMS NO LONGER RULE
Several hotels still apply traditional rules-based revenue management. Rules-based approaches allow hoteliers to establish a standard set of “rules” that generate strategies for forecasting, pricing, and yielding different room categories and situations. While a rules-based system worked adequately in the past, when hotels were dealing with limited data, these approaches poorly leverage the abundance of data that is now available. Using a rules-based approach today leads to suboptimal pricing and yielding, and results in missed revenue opportunities for hotels.
Many rules-based RMS providers promote their systems as offering dynamic pricing “optimization.” In truth – and by definition – heuristics-based approaches lack the capability to achieve true optimization.1They are neither mathematically optimized nor statistically robust enough to account for today’s complex distribution landscape. In fact, most rules-based approaches to RM are simply geared towards automating the “gut-feel”-based decision making that would take place in the absence of any system. But how is a revenue manager to know if a $10 drop in rate on a slow night will truly drive the best possible revenue outcome for a hotel? This is a crucial gap that only a science-based RMS can fill.
An extremely simplified example would be if a hotel offered one room category at three predetermined price points for a given night: Rate #1, Rate #2, and Rate #3. In a rules-based environment, once the forecasted number of rooms reaches a certain threshold, the system automatically shuts off Rate #1 and begins offering shoppers Rate #2, and so on. This approach to pricing is not truly optimal as it fails to put all of the critical puzzle pieces, such as market conditions, competitive pricing, length-of-stay, the elasticity of demand, customer value, and shopping patterns, together in a way that drives the highest achievable revenue result.
Trying to include all this valuable data into rate updates that happen multiple times per day, or over long forecast periods, creates an impossible computational challenge for a rules-based approach. Moreover, as changes occur, and more data becomes available, hotels must engage in the time-consuming process of continuously adapting rules, dealing with conflicting rules, and creating new rules as circumstances dictate. As a result, rules-based approaches are primarily reactive as opposed to proactive, unable to take essential consumer insights and real-time data into account. Ultimately, the rules-based model exposes a hotel to missed profit opportunities and less than optimal decision making.
WHY SCIENCE-BASED RMS ARE THE PATH TO SUCCESS
Next-generation RMS are advancing beyond the rules-based model. While revenue management is still considered both an art and a science, today’s revenue professionals have access to more data than ever before. So in order to flourish, and have the desired impact on their bottom line, hotels must determine how to make the art more scientific.2
True Dynamic Price Optimization
When it comes to pure dynamic price optimization, science-based RMS offer a more sophisticated solution over the rules-based model. True dynamic pricing is a complex task that’s shaped by multiple factors such as elasticity of demand, market conditions, seasonality, and customer behavior. If you’ve ever shopped online at Amazon.com, and noticed prices fluctuate higher or lower after you’ve refreshed your browser, you’ve witnessed true dynamic pricing at work.
In today’s fast-paced, competitive, and omni-channel marketplace, a science-based approach accounts for multiple variables that influence supply and demand, and responds in real time
to changing situations. Rather than determining pricing from a finite set of previously defined options – as we saw back in our rules-based Rates #1, #2 and #3 example above – science-based models allow hotels to adjust pricing at a much greater level of precision. The system chooses rates from a continuous range of potential values that provide the greatest opportunity to capture demand at each point along the spectrum of customer price sensitivity. Studies have shown that even small variations in price can mean big differences in profitability.
Effective Use of Customer Data
The ability of science-based approaches to incorporate key customer data into a hotel’s forecasting and pricing strategies also makes a crucial difference in maximizing profitability. Consider how this plays out in the retail space, where two customers each purchase a Dean Martin CD from an online music store. In a rules-based environment, both customers are instantly offered a Bing Crosby CD. However, because the rules-based approach can’t take advantage of available customer insights hidden within the data, the retailer misses an opportunity to increase his sale.
Science-based technology takes into account that although one customer is a 60-year old who enjoys ’50s music, the other is an 18-year old who only purchases classic CDs as gifts. Her shopping history reveals that her past purchases have primarily been in the alternative rock genre. In this instance, she’s offered the latest Red Hot Chili Peppers release instead. She makes the purchase, even though that wasn’t her initial intent in visiting the site, and the retailer has doubled his sale.
Along similar lines, using data to predict a customer’s purchase behavior ahead of time allows hotels to segment that customer appropriately and ultimately ensure that at any given time, only the most valuable, profitable customers have access to scarce room inventory.
More Accurate Forecasting
Innovative science-based RMS also utilize machine learning (ML) algorithms in some key aspects of revenue management. ML offers the capability for a system to continually refine and recalibrate its algorithms based on the inclusion of new and real-time data.
Proactive science-based models utilizing ML maximize profits by performing hundreds of calculations at lightning speed. They adjust rates for rooms along with ancillary products and services, and accurately forecast different demand segments according to each segment’s willingness to pay.
Overall Profit Growth
With the growing trend toward hotel TRM,3 revenue managers must recognize the ultimate goal is not about chasing after occupancy growth, but instead, maximizing profits across all revenue streams. This is best achieved via a science-based approach where strategies are no longer based on static rules, but instead leverage complex algorithms and extensive data sets that allow hoteliers to make informed fact-based decisions.
In today’s constantly shifting hotel industry landscape, a rules-based approach is deficient in the science to forecast demand or optimize pricing accurately. Science-based RMS improve data integrity, predictive analytics, generate optimal pricing strategies, and better inform your revenue management decisions. As research shows4, when implemented appropriately, a science-based RMS leads to substantially higher profits.
1“Distinguishing Price Optimization from Rules-Based Pricing,” Revionics research white paper, May 2015,
2“Revenue Management for Today & Tomorrow: 10 Questions to Consider.” HSMAI, Hospitality Sales and Marketing Association International , 25 Feb. 2011, www.hsmai.org/knowledge/summary.cfm?ItemNumber=4635.
3Kimes, Sheryl E. “The Future of Hotel Revenue Management.” Cornell University, Cornell University School of Hotel Administration, 13 Jan. 2017, scholarship.sha.cornell.edu/cgi/viewcontent.cgi?article=1239&context=chrpubs.
4Aziz, Heba Abdel, et al. “Dynamic Room Pricing Model for Hotel Revenue Management Systems.” Egyptian Informatics Journal, Elsevier, 28 Sept. 2011, www.sciencedirect.com/science/article/pii/S1110866511000375.
Dan Skodol is Vice President of Revenue Analytics at Rainmaker. Dan came to Rainmaker with over ten years of Revenue Management experience in gaming, hotels, multifamily real estate, and airlines. He is responsible for researching and designing enhancements and innovations within Rainmaker’s hospitality product suite, as well as supporting thought leadership topics and studies via analytics. Dan previously held Director of Revenue Management roles for two casino organizations in Atlantic City, and Archstone Communities. He holds a BA from Yale University and a Master of Management in Hospitality degree from Cornell. Dan and his wife reside in Denver, CO with their two-year-old son and enjoy skiing, hiking, and travel.