AI Will Eat Revenue Management

AI Will Eat Revenue Management

AI or Artificial Intelligence is currently dominating the news headlines.

NB: This is an article from Hotelsoft

Mark Cuban the well-known investor of ABC’s Shark tank has predicted that the world’s first Trillionaire will be someone who has mastered AI. One Mr. Bezos comes to mind. He is expected to be the first person to the Four comma club by 2030, just 12 years from now.

Artificial intelligence has been around since 1950s but never really took off as it was too difficult to scale. This was until three events happened, which are emergence of 1. big data 2. cloud business models and 3. advances in deep learning and parallel processing capabilities, which resembles most closely how the human brain works. Since then AI has grown, not sequentially but exponentially. It is expected that Singularity will be achieved by 2045, when machines become smarter than humans, which is only 17 years away, in our lifetimes.

AI is truly the future of hospitality, not just pricing and revenue management. I will discuss the reasons why AI will eat revenue management in hotels, airlines, cruise lines and other industries that have perishable inventories, and the reasons why it will not, at least not in the next decade.

Before we jump in, let’s set some expectations around time lines. Revenue management sector has been ripe for disruption for a while. The main challenge of disrupting pricing and revenue management is the bottleneck created by the fragmented technologies and existing enterprise solutions. It is a well-known fact as to how difficult it is to talk to the existing players and tech stacks, specifically the property management systems. For AI to be pervasive, the technology backbone of the industry needs to be fixed, which could take years.

While some AI algorithms will start determining pricing over the next few years and there are some players on top of that, including Hotelsoft, true AI replacement of revenue management is years away. The deep fragmentation lends itself to inefficiencies and thus a need for someone to police the system, in this case a Revenue Manager.

So, if one were to breakdown the tasks and responsibilities of the revenue manager into three broad categories, it will be as follows;

  1. Strategic – Data analytics, tactical, risk management, long term strategic
  2. Mechanical Turk – building promotions and rate plans, maintaining price parity (or not), managing inventory, forecasting
  3. Humanistic – human interactions and driving execution, business partnerships, goal setting etc.

An AI driven RMS can be seen holistically as an integrated function of an AI driven enterprise solution and not really a separate system. AI will be able to replace parts of Strategic and all of the Mechanical Turk, however it will not be able to replace Humanistic. You could imagine a system that truly analyses millions of prior outcomes and then comes up with the strategies based on these best past outcomes (in this case say maximizing the RevPar Index and guest satisfaction) and then implement these strategies across various channels to maximize these outcomes. It could be a mesh of self-communicating and transacting AI systems built on a singular goal of maximizing Profitability for individual entities.

Revenue management today is mostly rules based and driven by some form of regression modelling without any components of self-learning (note: AB testing is not self-learning as its constrained by A and B (or C) outcomes). Current algorithms make the forecasting better, but these are not self-learning systems.

The vision of a complete AI driven revenue management is a system that determines pricing based on the attributes of each request, create offers (optimizes 4 Ps – Place, Promotion, Product and Price) in real time, re-allocate online marketing spend, measure impact, maximizes outcome (an example could be RevPar Index, guest satisfaction scores) and then continuously iterate. The notion of an existing rate plan and offer could be done away with. Digital assistants (DA) like Google Assistant, Siri, Alexa are always “listening” to their protagonists, that is you, and know far more about you than you think. The AI driven RMS will need have the ability to communicate with these digital assistants, conceivably using some form of APIs, and make perfectly tailored travel itineraries based on your likes / dislikes, situation, needs – without a human being ever involved in either planning or executing the travel plans.

Therefore, an AI driven pricing will out-perform BAR based or rules based RM system. Today’s RMSs have varying degree of ability to forecast, change rates, open and close rate plans, but they don’t really have a built in intelligence that improves constantly and or writes new algorithms to improve performance. Think of this futuristic AI system as a very smart Revenue Manager’s brain fused with an RMS capable of processing unstructured inputs.

It is conceivable that this true AI revenue management system completely replaces revenue management, but it won’t be able to replace the humanistic elements of the profession. The AI will still need to be provided over-arching inputs, goals, short term and long term targets – at least in the near foreseeable future.

Read more articles from Hotelsoft

Related Posts

Comments are closed