SageMaker & Machine Learning

The Challenge

Leveraging past customer behavioral data and analytics to predict future customer behavior, engagement, and growth is valuable to all companies. Holland America’s goal was to leverage machine learning and artificial intelligence to gain additional insights into onboard offerings and to find ways to increase their customer experience and satisfaction.

Holland America’s additional challenge was to accelerate the efficiency of the company’s process for revenue prediction as well as identify key attributes for customers who are likely to use various services while on a cruise. For example, a key attribute would be to identify individuals who might use the spa with an incentive to do so. Additionally, adding segmentation for passenger profiles will allow Holland America to categorize and predict future behavior.

Why Amazon Web Services 

Holland America was already benefiting from the use of Amazon Web Services (AWS) for various data services. However, they were interested in determining a better solution for machine learning models that can increase its accuracy and efficiency of their forecasting. To address this, Holland America engaged with 1Strategy, an AWS Premier Consulting Partner to evaluate the capabilities of machine learning solutions on AWS.

“1Strategy helped Holland America Line build and deploy a machine learning solution that identifies passenger behavior, giving us an advantage, while enabling us to improve how we operate to deliver the most value to our customers,” said Lem Hebb, Director IT Strategic Planning at Holland America Line.

The Benefits

1Strategy conducted a Machine Learning Discovery to identify potential areas where new AWS technologies and methods could be of benefit to Holland America Line. Following the discovery, 1Strategy outlined and explained several opportunities for improvement, including the benefits of utilizing Amazon SageMaker and machine learning technologies. In addition, 1Strategy aided Holland America Line in optimizing the company’s existing machine learning algorithms and the data science behind their passenger revenue prediction solution and provided insights into forecast modeling improvements, which included the following:

• Assist with the migration of data from RDS Oracle to Amazon Redshift/Amazon S3 data lake
• Develop a machine learning solution on Amazon SageMaker to reduce runtime and deployment complexity
• Provide reports that analyze customer behavior that can be used to more efficiently predict passenger behavior in order to increase cruise revenue

Holland America Line had experimented with machine learning outside of AWS, but their method for prediction was not as accurate as they desired and took more time than existing forecasting methods.

In contrast, 1Strategy compared the company’s results from their experimental forecasting method utilizing Amazon SageMaker. This test resulted in the following:

• Increased Holland America Line’s existing machine learning algorithms accuracy to +80%
• Decreased the solution time for previous revenue prediction algorithm from hours to less than 4 seconds

Amazon SageMaker allows Holland America the ability to leverage a fully-managed platform on AWS that will allow the company to quickly and easily build and deploy machine learning models at any scale.

The capabilities and potential benefits of Amazon SageMaker for Holland America Line also include:

• Fully-managed service for building, training, and deploying machine learning models
• Host Jupyter notebooks
• Automatically scale to meet performance needs
• Support any model built into a Docker image

The work described in this engagement was originally completed by 1Strategy, a TEKsystems Global Services company acquired in 2019. As of June 2023, 1Strategy has fully integrated with TEKsystems Global Services to continue to deliver AWS expertise to customers. Learn more about our AWS solutions.