What inspired you to go with Starburst?

Before Starburst and Trino, we were struggling to connect our data sources. We wasted too much time copying and transferring our data and were looking for a solution that would allow us to access all our data together and spend more time creating models and performing analytics. We were already using Presto, now Trino, at the time, and it was a natural progression to transition to Starburst in order to get the most out of the enterprise only functionalities.

Where does Starburst fit in your overall data strategy? Which critical technologies are you integrating with Starburst?

Our entire Analytics and Model Training stack is integrated with Starburst. It is used to power all analysis at the company, and to support feature development, training, retraining and monitoring of our hundreds of ML models in production.

What are your three favorite Starburst features?

My favorite Starburst features are:

• JSON Parsing: Our data consumers are able to have fast access to different data sources, including ones stored in JSON
• Cost Based Optimizer: We’ve been able to accelerate queries by up to 20-30%, which significantly cut our computing costs
• Security: We have access to Personally Identifiable Information (PII) and Personal Health Information (PHI), so maintaining a high level of security is of the utmost importance

Which business outcomes are you most proud of?

We are really proud of the ability for Data Scientists to really explore and access data from across all our data stores quickly. In order to get the most out of our data, our data scientists have to have the full picture. By being able to access all our data from one access point, we enable them to be curious, explore, and innovate with our data.

In addition, we have also been able to enable robust modeling. With these models, we can optimize agent licensure and ensure sufficient insurance coverage to all our customers. This, in turn, allows us to provide more insurance products to more customers.

What success have you been able to measure?

Our time-to-insight has been reduced from months to weeks and from weeks to hours. This has led to multiple business optimizations across the entire organization Our conversion rates have also improved significantly. We can generate better insights resulting in an increase on conversions by over 10%.

Do you think decentralized data access and Data Mesh are the future of analytics at scale?

Yes! Each team should definitely have access to data and access to how it relates to others to be able to scale.