Snowflake's Performance Optimizations Help ESO Reduce Costs by 60%

ESO is the largest software and data solutions provider to emergency medical services (EMS) agencies and fire departments in the U.S. With a mission to improve community health and public safety through the power of data, ESO makes software that helps save lives. If you call 911 and a fire or medical team responds, it’s likely they’re using ESO software to make sure you get the right help fast.

ESO’s customers—first responders, ambulatory transporters, hospitals, fire departments, and regulatory bodies—use ESO’s software to document and share a caller’s journey from the time they call 911 through their discharge at the end of the emergency. Emergency response teams can look at historical data to understand how they can improve their performance for future events. This data is also used for legal and regulatory purposes.

Currently, ESO has more than 10,000 SaaS customers that include more than 500,000 users. Its fire and EMS record data sets alone comprise over 220 million records and counting. These records have a high churn rate with hundreds of transactional changes daily. For example, a change in blood pressure can affect yearly aggregates. Each record has to be processed as a snapshot, and many of the records are from third parties, so they need to be integrated accurately. 

And because the system supports emergency responses, it has to be available 24/7, with the ability to support high concurrency and scalability in cases of public emergencies or natural disasters. But there are also periods of time when no reports are being run, causing low usage and concurrency.

As a result, ESO needed a data platform that could flex to support high availability and scalability, and that also helped the company to manage and optimize its spend. 

To minimize TCO by saving both time and manual effort, Data Platform Technology Lead Andy Brown realized he would have to replace both of ESO’s legacy data platforms with Snowflake.

ESO’s data analytics platform was previously based on Cloudera running Scala and Spark. Although it was performant, running a big IaaS data cluster in Microsoft Azure was costly and time consuming. It required resources for not just a large number of servers, but also for a team of people with specialized technology and management skills. The team spent a great deal of time making sure the cluster was running and data was loading correctly, leaving little time for rolling out new features. 

Brown’s team replaced its Cloudera cluster running the analytics application with Snowflake in January 2022. Now, the Snowflake infrastructure can support the high concurrency needed when many customers run reports simultaneously. And because of Snowflake’s consumption-based pricing, ESO has reduced infrastructure and operational costs by paying only for actual usage, eliminating software license fees and recovering storage and server costs. Snowflake also enabled ESO to federate its data governance and gain visibility into the load versus query workloads. 

Since November 2022, ESO has experienced a 64% reduction in monthly infrastructure costs to run the analytics data platform. The median execution time for analytical reporting decreased by 60%. Reports that used to take 10 to 12 seconds now take only 3 to 5 seconds.

“It very quickly became apparent that Snowflake would save us money,” said Brown. “Originally, we were supposed to run the Cloudera and Snowflake data clusters side by side for two months to make sure we had a backup, but it was seamless. We ended up only running them together for one month, which resulted in additional cost savings.” 

ESO also saved on TCO by moving its larger ad hoc reporting platform to Snowflake. The median report execution time went from 30 seconds to under 10 seconds as a result of Snowflake’s segmented workloads. Most significantly, query usage for exploratory reporting increased 178% between August 2022 and February 2023, and expanded data storage increased ESO’s data footprint by 224%. But Brown was able to use Snowflake’s native cost optimization capabilities and learning resources to manage—and actually decrease—costs. 

“As the data usage and storage numbers increased, you would expect the cost to increase as well. But thanks to Snowflake’s features that enable continuous cost and performance optimizations, we’ve reduced our costs every month to date, and that’s resulted in a 60% reduction in monthly spend,” said Brown. 

Brown’s team has also been able to rely on Snowflake’s support team as it has transitioned and maintained its new data infrastructure. “We’ve been with Snowflake now for almost three years, and I can’t even quantify how much Snowflake’s engineers have helped us with our questions,” he said.

Now, ESO’s customers can use one dashboard that combines data from all entities—fire departments, emergency medical services, hospitals, and personnel management—to create a 360-degree picture of performance. Do medics perform better when they are staffed on 8-, 12-, or 24-hour shifts? Is there a degradation in care as their shift progresses? When is it appropriate for medics to refresh their training on emergency procedures? By combining information from all of ESO’s domains, Snowflake supports ESO in its mission to empower customers with data to improve the health and public safety of the community. 

In the future, Brown is looking forward to unlocking more capabilities enabled by the Snowflake Data Cloud. This includes self-service and automation features for employee user accounts, internal data cataloging that will support ESO’s data mesh implementation, and using Snowflake Marketplace to both publish ESO’s data and to enrich its own data with third-party data sets. Brown’s team also plans to use the time and money saved with Snowflake to analyze how ESO’s customers are querying data to create more personalized, account-based products. 

“It’s unique that the vendor that you’re paying based on your consumption actually wants to help you reduce how much you’re consuming, but that’s what Snowflake does,” said Brown. “And spending less on doing the things we’re already doing enables us to do more.”