There are many, many components of a successfully integrated and automated Revenue Strategy system, each of which must work together to achieve optimal results.
Often times the most critical things to get right are the least sexy, such as ensuring the right data strings flow seamlessly back and forth between the right systems in the technology stack. But of all components that make up revenue management software, the part that gets the most attention — and the part that I get the most questions about — is the pricing algorithm.
It makes sense, right? An algorithm is a company’s secret sauce. It’s a self-contained, step-by-step set of operations that processes specific data and, based on rules set up by the user, spits out an educated answer. At Duetto, the answer our algorithm recommends is a room rate that has been optimized, based on a number of unique data sets, for the highest bottom-line hotel profitability.
We don’t use the same data or the same algorithm as other revenue management systems. In fact, since we came later than most of our peers, we took a radically different approach to room rate pricing. So here’s a little primer on how Duetto’s pricing algorithm came together and what kind of data is factored in to our pricing recommendations.Duetto's 'secret sauce' recommends rates for highest profitability, via @marco_benvenuti… Click To Tweet
Why is Duetto’s Algorithm Different?
I came out of Cornell’s hospitality school with a strong understanding of statistical analysis. Later, while working at both Caesars and Wynn, I was able to begin putting those ideas into practice and automating revenue management where I could.
It didn’t take long to get frustrated with the standard Best Available Rate pricing, where third-party and discounted rates change in lockstep with the BAR. It struck me as completely old-fashioned, and I knew it was alienating my customers and costing my casinos a lot of money.
The legacy RMS systems available at the time were no help either. That’s what drove me to want to develop something better for casino and hotel pricing. So when Patrick Bosworth and I founded Duetto as a consulting practice, it was important to each of us that we create a new way of pricing specifically for hotels. It had to be sensible.
Before launching the GameChanger application, we worked for two years consulting hotels and casinos on pricing best practices. During those two critical years, we spent a lot of time developing the algorithm in Excel. Fortunately, almost simultaneously, much more hotel booking data became available.
We began incorporating lost-business data into the algorithm, to help clients price their rooms based on unconstrained demand. I determined the new data sets and order of operations that would produce an optimal rate for any booking situation, instead of relying on the BAR and applying some inflexible discount. Open Pricing, which we developed in conjunction with our initial algorithm, allows hotels to yield each segment, channel and room type independently. It’s a no-brainer, really.
And we learned a very important lesson that became the foundation for Duetto: that hotels and casinos can be much more profitable if they move from a “calculate demand first, then price second” model to a “price first, control demand second” model. This allows hotels to run quick, real-time experiments to see if price increases change demand. When they don’t, hoteliers can feel more confident raising rates in certain segments and continue to test more aggressive pricing on days without unconstrained demand.
The Algorithm’s Evolution
To make sure the algorithm worked this way, our next step was to add some data science. We did not go the traditional route, looking for PhDs from other verticals who could adapt a pricing model for hotels. We didn’t want an improved version of the same BAR approach the industry has been using for decades. To create something different, we needed outside thinking.
So we hired Michael Skinner, a back-end engineer with several years of experience building large-scale data processing systems and extracting meaning from the data they spit out. Skinner was a senior engineer at Google on the network architecture team, where he spent time designing and developing query pipelines, analyzing and forecasting network usage patterns, and building automated design tools to assist in the buildout of next generation Internet infrastructure.
A little later we brought on board Araz Feyzi as director of product analytics to bolster Duetto’s growth by analyzing data coming from the product, marketing, sales and other divisions and turning it into actionable business intelligence.
None of these innovative analytical features would work without the unique architecture of Duetto, built by our co-founder Craig Weissman, former CTO of Salesforce.com and a widely recognized leader in the field of SaaS development and design. Weissman has a master’s degree in computer science and a bachelor’s degree in applied mathematics from Harvard University, and he knew scalable metadata-driven, multi-tenant design like no one else.
Essentially, the direction we took was to hire smart people who were technically sound and analytics-savvy and were aware of real-world challenges, but not necessarily what was standard in the hospitality vertical. It was important for Duetto to have both qualitative and quantitative approaches.
Why Pricing “Recommendations?”
To me, the qualitative side of pricing is just as important as the quantitative. I learned this at UNLV in my undergrad program and later at Cornell during my master’s education. But it didn’t quite hit home until I was working in the field and getting real-world experience.
An algorithm that learns from its user and adjusts in real time is a phenomenal tool, but revenue management can be just as much art as it is science. There must be people who can take the algorithm and its output at face value and then apply a human touch based on real-world principles.
At Caesars, for example, we would override about 80% of the recommended prices. You might ask yourself: “What was the point in having the system in the first place?” It certainly was the question I asked myself every day, which drove me to develop the algorithm that still stands as the backbone of Open Pricing.
Contrast that experience at Caesars against Duetto, where we’ve managed to reverse those numbers. Our customers accept 80% of our pricing recommendations and override about 20%.
We understand that, on top of our forecasts, our users might sometimes have additional insight into a market. So we recommend what we think is the optimal price, but we let you make the ultimate decision. And then we learn from our customers by implementing what we call “machine learning,” allowing the application to automatically adjust the algorithm after an override.
If a partner hotel is overriding more than 20% of Duetto’s pricing recommendations, it’s time to do a deep technical dive to find out what’s happening in that market.
Where Are We Today?
Hotels — and casino resorts that include hotels — is the only vertical we play in. We’re firmly specialized in room rate pricing and Revenue Strategy, and we’ve built cloud-based software powered by an algorithm that’s specific to this vertical.
As Duetto continues to grow, its team will continue to reflect these two foundations of SaaS engineering and hotel Revenue Strategy. We have more than 100 team members now, and most of them come from either the cloud computing or revenue management fields. Our Customer Success and Solutions Engineer teams are made up of former revenue managers, so we speak the language of our clients and understand their daily challenges and opportunities. And our Technology and Product teams bring experience from leading cloud-based software and analytics companies.
We’ve also capped a long-term project of revolutionizing hotel loyalty programs, which I’m proud to say has finally come to fruition with the My Rate feature in the GameChanger app. Now Duetto can calculate the value of every guest and recommend a personalized price that will give guests immediate gratification, convert more business and drive more direct (and profitable) bookings.
While we’ve accomplished a lot in less than five years to be where we are today, I’m even more excited about where Duetto is headed in the future. What’s contained within the algorithm will stay Duetto’s secret sauce, but we’re cooking up new uses for it all the time.
- Airbnb Loses Major Legislative Fight In New York
- Webinar: Use Duetto to Drive RevPAR (email required)
- How Will New Supply Affect Your Hotel In 2016?
Latest posts by Marco Benvenuti, Chief Marketing and Strategy Officer and Co-Founder (see all)
- Revenue Strategy Report: The Hotel Industry’s Big Bet on Personalized Loyalty - November 6, 2017
- Revenue Strategy Report: The Real Potential for Artificial Intelligence in Hospitality - October 23, 2017
- Stop calling it “Unconstrained Demand!” - October 20, 2017