More than 50% of data leaders recently surveyed by BCG said the complexity of their data architecture is a significant pain point in their enterprise. Companies hampered by legacy data architectures are often plagued by a high total cost of ownership (TCO), an inability to govern data, and a lack of scalability as their data volumes grow. “As a result,” says BCG, “many companies find themselves at a tipping point, at risk of drowning in a deluge of data, overburdened with complexity and costs.”
Your technology stack should accommodate growth—in data volumes as well as in your business. It should empower people with varying skill sets to effectively use data and harness the power of AI and machine learning (ML). It should foster collaboration across functions. And finally, it should support exceptional security and compliance practices without limiting value and hindering innovation.
The Snowflake Data Cloud gives you the flexibility to build a modern architecture of choice to unlock value from your data. Snowflake was built from the ground up in the cloud. Unbound by the limitations of a legacy on-premises solution, its multi-cluster shared data architecture separates compute from storage, allowing data teams to easily scale up and down based on their needs. Companies and lines of business can securely collaborate on data regardless of their underlying data architectures, be it a data mesh, data warehouse, data lake, or a data lakehouse. Snowflake’s platform provides industry-leading features that ensure the highest standards of governance for your account, users, and data. All of this is delivered as a fully managed service, simplifying administration, upgrades, and helping customers do more with their data.
Here’s how one of our industry-leading customers modernized their data architecture with Snowflake.
Marriott simplifies its data platform and achieves lower TCO
It’s hard to believe that Marriott International started as a root beer stand in 1927. Today, the company comprises 32 global brands across 139 countries, with 8,300 hotels offering 15 million hotel rooms, and 100,000 home and villa properties. As Marriott’s business has grown over the past century, its data infrastructure has become more complex. In 2019, the company embarked on a mission to modernize and simplify its data platform.
Prior to 2019, Marriott was an early adopter of Netezza and Hadoop, leveraging the IBM BigInsights platform. Its technology team then moved Marriott Data Platform to AWS, which integrated batch and streaming data into cloud object storage, and processed data with a variety of technologies including Spark, Terraform, and Snowflake as a data warehouse for users to consume data through Tableau. Many of those technologies made the stack complex, costly due to expensive upgrades, and difficult to operate. Data engineers spent 20% of their time on infrastructure issues such as tuning Spark jobs.
Now, the team is on an ongoing mission to use Snowflake’s data platform to simplify the complexity of its tech stack. New functionality arrives automatically, eliminating the need for testing or infrastructure fixes. Snowflake simplifies data ingestion by consolidating batch and streaming, increasing Marriott’s speed to market—as soon as a customer transaction occurs, the data is available for consumption. With Snowflake’s Kafka connector, the technology team can ingest tokenized data as JSON into tables as VARIANT. Then, they parse and flatten that data using Snowpark.
Simplifying with Snowflake also allows the team to better support different people in the company at a lower cost. Business users have better access to data with less IT oversight needed. Analysts and data scientists can query data directly with their preferred language and tools. And third-party data is easily accessible from Snowflake Marketplace.
Today, simplifying its data platform on Snowflake has enabled Marriott to achieve transparency and control of its data, faster speed to market, improved collaboration and data sharing, a better user experience, and lower TCO. With Snowflake, Marriott has seen a dramatic improvement in performance and cost savings in comparison to Spark and Hive-based workloads. Many users from Marriott have commented on their improved experience with Snowflake, mentioning queries that used to take 5 hours or time out on Netezza now take 1 hour on Snowflake. Data that previously took 48 hours to one week in Hadoop is now available near-instantly in Snowflake.
Resources for modernizing on Snowflake
Thinking about a move to Snowflake? Here are a few materials to help you get started:
- Start planning a migration to Snowflake with one of our Migration Kits.
- Find out if you could benefit from migrating Spark workloads to Snowflake in our migration guide. See the results customers are getting by moving to Snowpark.
- Take Snowflake for a spin by joining an instructor-led, virtual hands-on lab.
And if you’d like to hear more from Summit 2023, explore sessions on-demand to watch the most popular presentations from Snowflake customers and dive into all of the new product announcements.