Everything we love about civilization is a product of intelligence, so amplifying our human intelligence with artificial intelligence has the potential of helping civilization flourish like never before-as long as we manage to keep the technology beneficial.” – Max Tegmark, President of the Future of Life Institute
“In the organization of the future, there’s a human, a computer and a dog…” – Dave Roberts, SVP Revenue Strategy and Solutions at Marriott International.
In recent years, the technology adoption curve has increased steeply-the amount of data being generated is expected to grow exponentially over the next 20 years, the speed of decision-making is accelerating and new technologies powered by artificial intelligence (AI) are bringing the human and the machine ever closer.
As the cost of computing continues to decrease by 10 percent every year, the speed of innovation increases. Consider that Amazon made 80 million pricing decisions during the 2017 Christmas shopping season, most of them automated. As more data becomes available, the accuracy of the analytics will improve. Google has recently proven with Google Duplex that it is possible for AI to have a conversation with a human without the human knowing they are talking to a machine. Google’s machine-learning accuracy has exceeded 95 percent, which is the threshold for human accuracy.
At the same time, consumer expectations are rapidly evolving. Customers want to be treated as individuals. Expectations on personalization and the need for authentic experiences versus products is driving how companies market and sell products. As an example, fashion retailers are behaving more and more like tech companies. Using detailed data, as outlined in these articles (Zara: Technology and User Experience… and A New Kind of E-Commerce… ), allows them to fine-tune their product manufacturing and manage costs, in turn creating a competitive differentiator which has proven hugely successful in bringing them ever closer to Clarke’s law #2, which states that any sufficiently advanced technology is indistinguishable from magic.
While the hotel industry has not made the same advances as the retail sector, the same forces apply: changing customer expectations, exponential increases in available data and the need for faster decision-making. To provide critical insights with this mass of consolidated data and deliver the decision-speed required, the hotel technology stack will continue its rapid evolution. Silos must break down, and the future will be consolidated with cloud-based services that quickly scale.
What does that mean for the revenue management discipline in the hotel industry? Consider a midsize hotel with 150 room types, 10 rates trying to optimize 7-night length of stays for 365 days-that’s more than 3.8 million pricing decisions-and that’s just for one channel, during one price change. If prices vary by channel or segment, the number of pricing decisions can easily multiply into tens of millions.
How can humans keep up? The answer: they can’t. It is time to let go of the semblance of control that revenue managers still have today. Tesla founder Elon Musk states that it is not the cars that kill humans; it is the humans that drive the cars. Similarly, one could ask, is it the revenue management system (RMS) that is killing revenues, or is it the revenue manager who does not trust the system and overrides the decisions?
After more than 30 years of continuous innovation, revenue management solutions are smart enough and must be trusted to make the right choices for hotels to optimize revenues. Leading-edge systems, powered by the most trusted revenue science, are able to optimize price and inventory by room type, guest, product, channel and length of stay and are capable of making automated decisions. In the future, we will see the unbundling of the room product into attributes and experiences, and the increased integration of customer data to create more relevant and personalized offers, in real time.
Additionally, the discipline of revenue management is rapidly expanding beyond a focus on rooms. The most innovative companies are not only putting in place departments and initiatives to tackle the challenges of optimizing revenues in meetings and events, food and beverage and beyond; many are also looking to consolidate all commercial aspects under the umbrella of a chief commercial officer who understands how to appropriately enhance enterprise value.
One challenge which has become apparent is that the legacy systems the industry was built on are not always capable of dealing with the massive amount of data which needs to be transported between a property management system (PMS) or central reservation system (CRS) and an RMS. Old interface technology creates bottlenecks, the PMS systems are not able to deal with true attribute-based shopping (and pricing) and the channels still work on the premise of “tell me where you are going and when, and I’ll give you a long list of room types to choose from” instead of asking for what experience the prospective guest would like to buy. As the technology improves and legacy systems are replaced with the next generation of interconnected platforms, this will change.
Artificial Intelligence Plus Human Intelligence Is the Future
How will the digital transformation of AI and machine learning impact revenue management and dynamic pricing in the hotel industry? Will AI replace human intelligence in revenue management and pricing analysis?
Hotels and resorts have used revenue management and price optimization for several decades to sell rooms, tickets and services to the right people, at the right time and for the right price. To determine the full benefits of dynamic pricing, the prices will have to be adjusted in real time, possibly after each transaction, using sophisticated techniques derived from a combination of sound mathematical models, predictive analytics and game theory.
The amount of available data and the ability to process it with machines is growing exponentially, whereas human brain processing capacity is limited. A team of software agents that rely on AI with big-data and machine-learning techniques will be needed to address this complexity. This however does not do away with the need for human agents with the expertise and intelligence to account for various external factors which the system might not be aware of but which would affect decision-making.
For example, hotels may find it interesting to consider automated pricing agents such as shop bots and price bots that reset the selling price at optimal intervals based on supply and demand with the objective of maximizing discounted cumulative profit or long-run average profit per unit time. As another example, machine learning can be used by seller agents to learn directly the probability of winning from a database of bid transactions with known outcomes.
It is conceivable that humans of the future will rely upon autonomous software agents that exchange information, goods and services with other software agents representing consumers, producers and intermediaries. Such software agents will help in all facets of electronic commerce based on up-to-date, real-time information, enabling consumers to be better informed about products and prices. Likewise, software agents will enable producers to be better informed about and more responsive to their customers’ needs. Such software agents may be programmed to be self-motivated and independent economic players relying on algorithms for optimizing the profit or other objectives on behalf of their human owners. However, this potential reality, while enticing, is flawed.
We envision a future in which human agents and software agents will work together to solve the problem. The future solutions will be based on an architecture that will enable software agents to work with human agents to provide the deep knowledge of the domain that may not be captured by software agents alone. Humans can also more readily adapt as circumstances or objectives of the problem to be solved evolve and change, helping the software agents to continue to deliver the optimal solutions.
Complex optimization problems, such as group or contract negotiation, or problems that arise in conducting a what-if analysis may have multiple objectives and constraints. Often, there is no single dominant algorithm that can produce a truly optimal solution to the problem because multiple competing objectives make it difficult to determine which solution is best. Humans can offer good initial solutions that software agents can use as a basis for further improvement. Conversely, human agents can refine alternatives generated by software agents to account for any special-case considerations.
From Artificial Intelligence to Critical Intelligence
In looking to revenue management of the future, it is helpful to ask: what are the key considerations for a revenue manager of today, and how can technology help?
1) Make sure that your data is in a place where it can easily be exchanged. Your “tech stack” should include systems which are cloud based, have an open approach to integrations (through open APIs) and are willing to work within and actively contribute to the ecosystem of hotel technology. Look for the latest in PMS/CRS and channel management technology.
2) Implement an RMS that doesn’t require manual decision-making, outside exceptional circumstances. Go for a decision system with the best revenue science available, not a recommendation system which is based on rules that you must control and manage.
3) Your RMS should also enable you to “blow up” your rate and product structure by understanding demand for each product in its own right and optimizing each while considering global optimality.
4) Consider adding a customer relationship management (CRM) system to the mix to move you along the journey of personalization and engagement. Understanding the consumer and being able to fine-tune offers for known and unknown customers will differentiate you from the pack.
5) Sit back and watch the magic unfold with your PMS and CRS capturing data, your CRM understanding your customers and your RMS making sure the right product gets put in front of the right customer through the right channel. Your leadership team will be astounded by your revenue wizardry.
As technology progresses, AI can assist in automating more and more critical tasks. However, human interaction and engagement with technology is still a key determining factor for a successful ROI of any system. We propose the term “critical intelligence” which uses AI to improve what humans cannot do as well, or as efficiently, and overlays it with human intelligence and experience.
Dave Roberts-senior vice president of revenue strategy and solutions for Marriott-is fond of saying that hotel revenue management of the future will require a human, a computer and a dog. The human’s job is to keep the dog fed, and the dog’s job is to make sure the human doesn’t touch the computer. While the dog may be a stretch, the message is clear: let the machines do their job so the humans can do theirs better.
Reprinted from the Hotel Business Review with permission from www.HotelExecutive.com.