The hotel industry, long dependent on traditional methods, is now facing transformative change. As digital channels reshape hospitality, it’s important to grasp the role of artificial intelligence (AI). AI can improve guest experiences, personalize interactions, optimize operations, and refine revenue strategies – shifting manual tasks and spreadsheets to AI-driven decisions.

NB: This is an article from Flyr for Hospitality, one of our Expert Partners

Subscribe to our weekly newsletter and stay up to date

Our AI-first revenue optimization platform is changing the game, introducing new pricing strategies and enhancing competitiveness.

But how do we compare and measure AI-based revenue management outcomes? Which methods are most efficient and effective for real revenue gains? Should revenue managers interact heavily with an AI revenue management system or trust automation? What is the impact of overrides on revenue? How should we strategically influence? How do we monitor performance?

Technology leaders like Amazon, Netflix, Airbnb, and have a long history of enhancing their decision-making processes through modern causal methods. By leveraging internal software platforms and dedicated teams of data scientists, these companies aim to obtain quantitative insights into the impact of their decisions and potential future outcomes. They utilize a diverse toolkit that includes A/B testing, quasi-experiments, natural experiments, and machine learning causal inference methods, and strive to refine their strategies and make data-driven choices.

What are all these methods? And can they be used in hospitality to enhance our AI-first RM platform? The short answer is yes, they can, but we need to understand better what these methods can and can’t do.

Let’s explore.

The shifting dynamics between human and machine

Strategic reasoning and inputs into the RMS are multifaceted practices that hinge on historical data analysis, competitor benchmarking, and customer segmentation to forecast demand and set prices. Key elements include adjusting prices dynamically, working with other departments, scenario planning, and regular performance reviews.

To stay competitive, revenue managers monitor sales data, market trends, and competitor activities. They also work with marketing and sales teams to ensure pricing supports business goals, using what-if analyses to anticipate market changes and refine their strategies. This process is inherently complex.

FLYR’s AI-first revenue management technology harnesses the power of advanced machine learning algorithmic systems designed to quickly detect intricate patterns and trends that human analysts might overlook, facilitating more precise and prompt decision-making. The platform seamlessly combines data from diverse sources, encompassing historical sales data, market trends, competitor pricing, and external factors like economic indicators and weather patterns.

This comprehensive data integration empowers the engine to construct a deep view of the market environment and anticipate future demand. The dynamic entity continuously learns and adapts to new data inputs, ensuring prices are always in sync with the current market environment, capturing maximum revenue and minimizing waste.

By harnessing modern AI and automation to empower teams, organizations can optimize business processes, streamline operations, and unlock unprecedented levels of performance.

Modern Causal Inference at FLYR

Despite the recognized value of causal inference amongst data-driven organizations, it has been very scarcely deployed in hospitality settings. The hospitality sector faces challenges due to its complicated distribution and customer acquisition processes. These issues, fragmentation, and the lack of strong testing systems have made it historically hard to run effective experiments and put causal inference methods into practice. The good news is that many of the elements needed for modernization are now available. There are enough tools and techniques for causal inference to manage distribution problems that have been proven to work in industries like airlines and retail. FLYR also uses causal ML as part of its relentless and fast innovation cycles, thus possessing the technology to lower the cost of running experiments for clients.

Picture this: Controlling variables in a highly dynamic market is no small feat. Yet, by leveraging causal Machine Learning, we isolate the effects of human input from the constantly evolving market conditions. Through this meticulous process, we’ve charted new territories to enhance revenue management strategies.

One of the cornerstones of this innovation is the use of A/B testing, a method widely embraced in various industries for the reliability of the insights it provides. By comparing controlled scenarios—e.g. manual pricing interventions against automated engine decisions—we decipher patterns essential for robust revenue optimization.

At FLYR, we pride ourselves on being pioneers in applying these advanced methods and making unprecedented strides in revenue management. Through our A/B testing service, described in Cornell Hospitality Quarterly, we empower our clients to make informed strategic decisions. But we go further, and exploit the full toolkit that causal ML sets at our disposal.

As a case study, consider another technique out of the causal inference field: natural experiments. Easier to run than controlled A/B tests, they hold the potential to revolutionize traditional revenue management practices. At FLYR we use this methodology to help our customers understand the influence of past decisions. This process validates current strategies’ performance and arms them with data-driven insights for future endeavors. It also helps clients understand the ROI of the FLYR optimization platform vs. other options and continuously and demonstrably proves the value of innovation in our revenue engine.

For instance, when assessing the impact of strategic manual price adjustments by some of our most significant clients, the results spoke volumes:

  • In some cases, strategic manual interventions resulted in a revenue loss of up to 7%.
  • Conversely, other scenarios demonstrated positive revenue uplift through similar manual interventions.

Why does this matter? Simply put, it challenges preconceived notions and highlights the varied outcomes of manual versus automated strategies. Revenue management isn’t a battle of human versus machine. Rather, it’s about the symbiosis of human expertise and technological precision. However, it’s crucial to acknowledge that not all human interventions are equal, and some manual adjustments may not yield optimal results. By providing a concrete measurement of strategy performance, we inform teams to adapt their approaches and maximize efficacy.

Causal inference offers many tools that revenue management teams should have at their disposal. These tools unravel the intricate web of variables, leading to actionable insights. With a commitment to ongoing innovation, FLYR is proud to be leading that charge in making these tools available to clients.


One critical takeaway from our experiments is understanding the nuances that influence revenue performance. For instance, while some clients benefit from manual input, others do better by relying entirely on the automated engine. This heterogeneity underscores the importance of personalized strategies tailored to specific circumstances and market conditions.

Moreover, our causal Machine Learning in natural experiments doesn’t just stop at evaluating strategies. They are instrumental in helping our ongoing customers refine their approaches. By continually analyzing the impact of their decisions, we enable a cycle of perpetual improvement and optimization.

The future of revenue management lies in embracing these innovative practices. With the insights gained from field experiments and, more generally, causal inference, we can more efficiently navigate the complexities of market dynamics. As we continue to push the boundaries, the insights gleaned from these experiments will undoubtedly lead to smarter, more strategic decisions in revenue management that foster more substantial and sustainable revenue growth.

Read more articles from Flyr for Hospitality