How JetBlue’s integrated use of Snowflake and Fivetran is a model for the modern data stack

Commercial airline company, JetBlue, is using the integrated Snowflake data stack to power insights into its customer services and operational activities.

Ashley Van Name, general manager of data engineering at JetBlue, recently explained how a tight grip on the vast amounts of data it holds is crucial for the airline. Her team turns information into insight by using a combination of Snowflake’s Data Cloud and Fivetran’s data integration service.

This combination represents the modern data stack. Data management is a complex challenge, and information comes from multiple directions and myriad sources. So, any attempt to turn both structured and unstructured data into useful insight requires a broad collection of techniques and technologies.

For the past few years, Van Name’s work has centered on a massive data consolidation effort. By using an integrated data stack that draws on a combination of systems and services, JetBlue decision-makers benefit from powerful insights that create a competitive advantage.

“We believe that data should be looked at as a product,” said Van Name. “If you lead a product team, and you build a software tool that is consumed by an end user, then you need to ensure you are meeting their needs and building something that can make lives easier.”

Generating valuable insights

JetBlue’s three-year journey to a modern data stack has involved careful integration work. Snowflake sits at the center of the stack and stores enterprise information as it would in a data warehouse. However, this cloud-based technology also integrates with a range of other services to build the reports that power decision-making processes.

JetBlue uses a variety of software-as-a-service (SaaS) applications, with data streaming into the data warehouse from ServiceNow, Qualtrics, JIRA and Salesforce. The company also uses third-party SQL servers and Oracle databases.

Van Name notes that the key to creating a single view of this information is to ensure raw data enters the Snowflake Data Cloud. That’s where Fivetran then plays a crucial facilitating role, helping the data engineers to ensure JetBlue’s data is centralized rather than siloed.

Fivetran provides a pipeline-as-a-service, which combines data from multiple sources for analysis. The service provides more than 200 source connectors that replicate information sources directly to the Snowflake Data Cloud.

From this consolidated location, JetBlue’s data engineers get to work. Van Name explained that the combination of Fivetran and Snowflake ensures employees are focused on generating insight from information rather than tinkering with the technology.

“We believe that data engineers should spend time making their data useful rather than just moving it around,” she said. “Fivetran creates this world where gone are the days of sitting and laboring over building a custom data pipeline. Instead, you can reap the benefits of the output of that pipeline with just a few button-clicks.”

Building the modern data stack

JetBlue’s tactical use of Fivetran is illustrative of a new phase in the data stack. While Snowflake aims to provide a one-stop shop for executives that want to manage the large volumes of information their businesses collect, the enterprise software giant also recognizes it can’t meet this objective in isolation.

Snowflake CEO Frank Slootman explained in a talk earlier this summer how his company has goals that reach far beyond running databases in the cloud. Either through developing its own tools or — even more crucially — by working with trusted partners like Fivetran — Snowflake wants to create an integrated data stack.

“From the standpoint of the business model, it doesn’t really matter to us whether we own the tool or somebody else does,” Slootman said. “We have a very different strategy than public cloud providers who want to check every box and be able to say to customers, ‘ere’s the whole kit and caboodle.’ We’re a best-of-breed company.”

Take security, which is one of the key areas where Snowflake is investing in. Here, the aim — by working with partners such as Hunters, Panther Labs, and Securonix — is to ensure customers can pick best-of-breed applications that meet their business requirements and which run securely on the Snowflake platform.

In short, Snowflake’s vision of a modern, integrated data stack is one that uses tools to help executives at companies like JetBlue make the most of their information, regardless of who creates the tool.

As Snowflake cofounder Benoît Dageville told to VentureBeat, his company believes the key to delivering an integrated data stack in the longer term will be allowing external developers to build and run their own applications on the Snowflake platform.

“Snowflake will be mostly not be built by our employees – that’s the vision,” Dageville said. “We’re all about removing friction and democratizing things so that everyone can be an application builder.”

Reaping the rewards

Back at JetBlue, Van Name said the ability to work with a specialist pipeline provider to help push data quickly to Snowflake is already paying dividends.

She noted that  it is not uncommon for data engineers at other firms to spend around 40% of their time building and testing the ingestion part of the data pipeline. Since implementing Fivetran, Van Name estimates just 10% of a JetBlue data engineer’s time is spent testing, validating and making sure the data is being pushed into the warehouse.

That shift in effort means engineers can now spend 90% of their time focused on how data is consumed in useful products.

“You get the most bang for your buck from your data engineers because you can allocate them to harder, more complex data problems,” she said.

One area where JetBlue is using data to create a competitive advantage is in customer experiences. Van Name pointed to the analysis of a post-flight email survey that is sent out by its partner, Qualtrics. Once customers have entered their survey answers, the responses are logged back in Qualtrics before Fivetran’s pipeline pushes the data into Snowflake.

“This process allows us to replicate the responses very easily and quickly from our customers directly into our Snowflake Data Cloud in their raw form, so that analysts can access the information and derive insights from it,” said Van Name, who also noted the data has sponsored improvements in areas such as in-flight entertainment experiences and digital interactions.

JetBlue is also using data to boost operations maintenance. The airline runs a computerized maintenance management system via an on-premises Oracle database. It uses Fivetran’s HVR technology to replicate the mission-critical operations database in real time. Employees can then use insights from this data to make timely interventions.

“That approach decreases the amount of time it takes analysts to build reports because they have one place to go and pull the information they need,” Van Name said. “It also empowers more consistent reporting; you’re getting closer and closer to that source of truth of a certain set of information that your company cares about.”