Overview Generative AI Case Studies Partners Resources Blog

AWS for RMS: Modern revenue management in the cloud

Learn how to modernize, augment, complement, and advance your revenue management systems in the cloud using more data, scalable compute power, easy-to-use machine learning, and off-the-shelf artificial intelligence services to more accurately predict travel demand, set a modern pricing strategy, and provide the right traveler with the right product at the right price.

Revenue management, the cloud, and AI/ML | Attribute-based shopping | Dynamic pricing and AI/ML | Marketing and revenue management | Resources

65% of airline bookings

are influenced by airline revenue managers

10% of hotels

have to purchase revenue management systems

84% of airlines

want to use AI-based dynamic pricing strategies

2-5% gain in EBITDA

can be achieved from AI/ML pricing

2-8% room revenue increase

for hotels with attribute-based shopping

Overview

Revenue management (RM) is the backbone of the travel business and is quickly becoming more prevalent in the hospitality industry. Over the years, the travel industry has developed sophisticated systems for forecasting demand, managing inventory, and responding to competitors’ prices in the market.

Revenue managers across travel and hospitality have been using advanced mathematical models to predict demand and set pricing using historical data. However, the disruption in travel and subsequent changes in traveler behavior over the past 2 years have made those methods less effective. This has accelerated travel and hospitality companies’ desire to modernize their revenue management systems (RMS), to move toward sophisticated retailing techniques, and to give travelers the ability to personalize their experience. That level of personalization requires extensive data and sophisticated real-time models that can be recalibrated with each purchase made.

That’s where Amazon Web Services (AWS) and the cloud come in. To better understand the customer and to maximize profits, you need the ability to collect, store, and analyze troves of data quickly, make decisions faster, and easily connect the dots by using advanced analytics, machine learning (ML), and artificial intelligence (AI). The cloud can cost-effectively provide the storage and compute needed for such complex workloads, as well as the ML/AI tools to enhance them. In the guide, you’ll learn how travel companies are optimizing their revenue management strategies with AWS and our Travel and Hospitality Competency Partners’ solutions.

Enhancing the revenue management process with the cloud, AI, and ML

The process for effective revenue management involves five phases: data collection, segmentation, demand forecasting, optimization, and reevaluation. Each step can be enhanced and done more quickly with the aid of the cloud.

1. Data collection

Collect, store, and use first-, second-, and third-party data in the cloud in real time to improve your revenue management models’ accuracy and quickly update your models to meet market conditions.

2. Segmentation

Define products and the ideal customer for those products using models built with the data collected.

3. Demand forecasting

Predict demand for products at certain price points with sophisticated mathematical models, enhanced by AI-based detection functionality and ML-based notifications for real-time updates. By combining these forecasts with calculated price sensitivities and price ratios, an RMS can develop price-optimization strategies to maximize revenue.

4. Optimization

Use a linear programming model, an ML/AI model, a statistical model, or a combination to develop a list of the product offerings, inventory levels, and price points to sell to meet your revenue objectives.

5. Reevaluation

Continuously recalibrate models, adjusting for data changes (i.e., inventory) when a product is sold to optimize revenue.

For a total revenue management approach, advertising and marketing play a key role as well by using data and models to more effectively and dynamically market different combinations of ancillaries, value-based pricing, and experience-based loyalty offerings across channels to relevant customer personas.

Today, revenue management is central to digitization and personalization of products based on a customer’s unique preferences for competitive advantage. Effective customer segmentation, use of machine learning and artificial intelligence, and adoption of modern cloud technology are the key to success.”

Ben Vinod
Co-founder of Charter and Go and sitting member of the editorial board of the Journal of Revenue and Pricing Management

Case study: Cathay Pacific Airways modernizes passenger revenue optimization system on AWS

As a founder of One World Alliance, Cathay Pacific Airways (Cathay Pacific) serves customers all over the world with destinations across five continents. For its customers who are booking flights, Cathay Pacific uses a passenger revenue optimization system (PROS), which “basically processes all our bookings to optimize revenue,” says Lawrence Fong, general manager of information technology at Cathay Pacific. The airline has been using PROS for over 10 years.

Cathay Pacific ran into a number of challenges running the massive PROS solution on premises. It required difficult, costly hardware upgrades to the infrastructure every few years to keep up with demand. By migrating PROS to AWS, Cathay Pacific can now support more advanced analytics modules and automated security patching to improve efficiency. This modernization increased performance by around 20 percent, shortening backup windows from over 10 hours to just 6–7 hours.

“AWS helps Cathay Pacific to transform the business and stay ahead of the competition," says Lawrence Fong, General Manager of Information Technology, Cathay Pacific.

Download the AWS for RMS: Modern Revenue Management in the Cloud ebook to learn more.

The move to attribute-based shopping in hospitality

A report by CarTrawler and IdeaWorks Company predicted that airlines would collect roughly $65.8 billion in ancillary revenues worldwide in 2021, an increase of 13 percent compared to 2020. Some of the largest airlines collected more revenue in ancillary sales than in base airfare. Many see it as a way to give travelers more personalized options while optimizing revenue. It’s no wonder that hotels want to adopt a similar strategy, known as attribute-based shopping.

There are several benefits of moving to an attribute-based shopping approach. One, accommodations and lodging providers can provide guests the control that they want to build the room and amenities that they desire. Two, you can differentiate your brand with a larger variety of room offerings. And three, attribute-based shopping offered in your own digital marketing channels can increase direct bookings and reduce distribution fees.

This is the direction many hotel brands want to head toward. However, doing so requires hotels to shift their revenue management practices and potentially upgrade their RMS. To bundle amenities and rates at this level, hotels need to collect and access the relevant data needed in a cloud data platform that integrates with the RMS. This approach requires elevating your data analytics approach to combine it with AI/ML. Hotels can use AI detection to monitor customer buying patterns and sales conversion points and use ML to automatically update revenue management rules and campaign management rules in real time.

Today, most RMS do not offer out-of-the-box capabilities to support this shift in pricing, but custom data models can be built using AI and ML and be integrated with RMS to take steps toward this goal.

Dynamic pricing and the role of AI/ML

AI/ML has been taking more of an active role in dynamic pricing. There is a need for dynamic pricing techniques that scale and adapt to market conditions but maintain the appropriate business risk mitigation controls so that the customer experience is not impacted.

Customers often ask what the strategy is for getting started with these types of techniques. It is important to consider:

What are your specific targets?

It is necessary to be very clear on establishing the overall goal of your algorithms—for example, is it to maximize margin or increase search-to-booking ratios? These will need to be baselined to determine if a model is effectively delivering your business value.

What is your search capability?

How do you calculate and present prices? These questions will determine the supporting infrastructure required—ranging from near real time to inter-day to daily price changes. These questions help to establish the frequency of change in your algorithms and help the measurement of the algorithm value benefit.

What are your risk controls?

Changing prices does not come without risk. The business controls are critical for full traceability of the impact of decisions made. These mechanisms can be approval stage gates, testing to a subset of the market to fully automated decision-making algorithms. Also, when building models, consider bias and solvability of training datasets so that there is no unintended market bias against market or customer segments.

View Foundations of Revenue Management for a deep dive on the evolution of these techniques and the AWS Revenue Management IT Buyer's Guide for solutions and reference architectures to help you adopt dynamic pricing using ML/AI.

Case study: TUI dynamic pricing in action

TUI is the world’s largest tour operator. In April 2022, TUI deployed the first version of its Reinforcement Learning Agent (RAL) as an endpoint in Amazon SageMaker, which can be used to build, train, and deploy ML models. RAL gets queried up to 200 times per second by the pricing engine in real time. The system decides margins in such a way that a previously defined target function gets optimized.

The RAL is constantly A/B tested against different flavors of the agent as well as against the traditional rule-based system, which also acts as the fallback solution. The system decides the margins in such a way that a previously defined target function (the profit) gets optimized.

“After a short adjustment time of the new algorithm, we were able to achieve higher margins. Over the last 3 months, the total booked margin strongly increased compared to the rule-based system. On top, we could prove the ability of the system to adopt to changing market dynamics. With Amazon SageMaker, we now have a stable platform for our ML models, which lets us focus on the model development,” says Dr. Lukas Schack, Principal Machine Learning Engineer at TUI.

Download the AWS for RMS: Modern Revenue Management in the Cloud ebook to learn more.

Marketing and revenue management

Modern revenue management has a key role in enhancing and monetizing customer experience, perhaps the strongest currency of business success. In fact, customer-centric travel and hospitality companies are swiftly migrating to the cloud, accelerating the rearchitecting of their microservice-based applications and data platforms to create a powerful flywheel between revenue management, customer experience, and marketing.

Customer experience drives consideration, conversion, and willingness to pay. Reviews and ratings are key but not the only expression of customer experience and subsequent behavior. Social media sentiment, clickstream patterns, real-time transactions or customer interactions with the various digital assets, or a contact center are equally critical real-time signals. Data ingestion, combining multiple sources, and real-time processing so that effective decisions are made based on AI/ML predictions, recommendations, and automation is required.

Travel and hospitality players are turning to the cloud to innovate, injecting technology in each of the customer journey stages, from awareness to consideration, decision, experience, or advocacy. The purpose is to improve customer experience meaningfully, knowing full well that modern revenue management could monetize this effectively and must be an integral part of this powerful flywheel.

Read more about marketing and the five steps to implement your modern RMS in the AWS for RMS: Modern Revenue Management in the Cloud ebook.

Learn more

AWS for RMS: Modern Revenue Management in the Cloud

This ebook explores how to modernize and optimize revenue management with the cloud and ML and AI.

Download ebook »

Foundations of Revenue Management

This comprehensive guide provides in-depth detail about all the stages of revenue management from data collection to marketing.

Download ebook »

The AWS IT Buyer’s Guide for Revenue Management

Discover the AWS services, architecture references, and AWS Travel and Hospitality Competency Partner solutions available to help you maximize revenue.

Download ebook »