Building an analytics ecosystem with dbt and dbt Cloud for a healthy data mesh at Rivian from Coalesce 2023

“[Data mesh] is this amazing modern recommendation around this age-old problem of centralization versus decentralization in development in the data space…but there's also this question of 'How do we get there? How do we build this?'"

Will Bishop, Manager of Data Science Analytics at Rivian, discusses the journey of Rivan, a new American car company, in implementing data mesh in their organization. Will shares the story of their data analytics journey over the last two years, starting from scratch and gradually evolving towards adopting a data mesh structure.

Rivian's journey towards implementing a data mesh

Rivian's journey towards implementing a data mesh involved a two-year evolution that began from scratch. Their efforts are a work in progress but significant progress has been made, particularly in leveraging data analytics for the company's operations.

"Are we there yet? No, not yet at all, but we've made a ton of progress in that time, and we are evolving in the right direction," says Will. He explains how Rivian's approach mirrored the data mesh concept, growing from serving consumers locally to expanding into a larger global data economy.

"We want to build a healthy free market. We want to have effective production and trade. But we also want to have standards, and regulation, and great visibility as we build that out," he adds. Will also emphasizes the importance of establishing a simple language and strategy for handling data to facilitate better communication with other teams.

The importance of iteration and constant value delivery in data management

Emphasizing the importance of iteration in data management, Will underscores the value of delivering constant and achieving incremental success at each step of the process. He also highlights the importance of making good practices easy to fall into and focusing on solving people's pain points before moving towards delighting them.

"We want to iterate, delivering value at each step along the way," states Will. He also points out the necessity of making good practices easy for people to adopt, saying, "We want to make those good practices easy to fall into and use."

Empowering people in data transformation and solving domain-specific problems

Will highlights the importance of empowering individuals in the data transformation process. He notes that those who use data are often best equipped to help transform it and advocates for equipping individuals with the tools they need.

"People who use data are often the best equipped to help transform the data," says Will. "We want to bring power to the people in transforming data."

He also stresses the importance of focusing on one domain, emphasizing the importance of a targeted approach: "We want to solve for one domain really well and expand from there.”

Insights surfaced

• It's important to iterate and deliver value at each step of the journey
• Establishing a simple language and strategy for handling data is crucial
• Modeling good practices and sharing them with others can help in learning and growth
• Making good practices easy to follow can lead to better adoption
• Solving people's immediate pain points can lead to greater acceptance and delight
• Solving for one domain thoroughly before expanding to others can lead to better results