AUTODESK PUTS SNOWFLAKE AT THE CORE OF ITS CUSTOMER 360

Software company Autodesk is changing how the world is designed and made. More than 100 million people use Autodesk software to solve challenges in architecture, engineering, construction, product design, manufacturing, media and entertainment. Autodesk’s customer 360 Analytics Data Platform (ADP) supports a variety of BI, data science and customer-facing use cases. 

Autodesk’s data lake architecture was operationally burdensome to support and cost-prohibitive to scale. Data ingestion workloads relied on large amounts of homegrown code that led to frequent troubleshooting and unreliable data. Data-access-control limitations presented data governance challenges. 

Performance issues with its Spark and Athena ecosystem inhibited Autodesk’s product teams and business users from accessing timely insights. “People would go to use a BI dashboard and half the time it wouldn’t load,” says Mark Kidwell, Autodesk’s Chief Data Architect of Data Platforms and Services. Lack of trust in ADP caused teams to consider building their own data environments. 

Seeking to elevate self-service analytics and reduce system complexity, the data platform team developed a plan to rearchitect a large portion of ADP.

After comparing multiple solutions, Autodesk chose Snowflake. According to Kidwell, “We wanted to make the most people the most productive in the shortest amount of time, and Snowflake was the clear winner.” 

Snowflake’s multi-cluster shared data architecture and elastic scalability solved Autodesk’s performance issues, which led to an enhanced experience for BI users. Per-second pricing provided a cost-savings opportunity for Autodesk’s analytics workloads.

Snowflake’s native SQL support and extensive network of connectors, drivers and programming languages simplified data ingestion and transformation. “Our goal was to enable that end-to-end pipeline entirely in Snowflake,” Kidwell says. Snowflake Marketplace offered a convenient way to access ready-to-query data sets without the cost and effort of traditional ETL processes. Configuring a bidirectional sync between Snowflake and Autodesk’s cloud object storage accelerated adoption. 

Snowflake’s near-zero maintenance reduced administrative work and freed up technical staff to focus on increasing analytics.

Autodesk’s reimagined data architecture allows the data platform team to support even more self-service analytics use cases. For example, Autodesk’s CX team relies on help center interaction data in Snowflake to understand and improve the customer experience. Analyst teams use Snowflake to build data products that predict churn, identify business opportunities and anticipate the impact of process-related changes. According to Kidwell, “Analyst teams are asking—and answering—big questions with Snowflake.” 

Empowering a growing number of Autodesk’s users and teams with self-service analytics has led to increased awareness for Snowflake.

Migrating data, users and workloads to Snowflake helps Autodesk achieve a highly efficient and cost-effective data environment. ETL processes now run in minutes, not hours. Accessing web event data via Snowflake Marketplace expedites user behavior analysis at Autodesk. Alleviating data reliability and performance issues shortens development cycles for new data products. 

Autodesk’s data platform team needs 3x fewer people supporting Snowflake compared with the previous data lake architecture with Spark and Athena, which frees additional capacity for innovative work. 

“We’re doing things that would have been difficult to impossible on the previous data lake architecture,” says Kidwell.

Snowflake also supports a growing number of data products that are embedded into Autodesk’s software. “With the data in ADP and the ecosystem for ingesting, processing and analyzing data, we’ve built products that display insights to end users,” Kidwell says. Data-driven reports and feature recommendations help Autodesk’s customers maximize utilization, collaboration and productivity.

Increasing self-service access to data and insights is a top priority for Autodesk’s data platform team. Continuously optimizing ADP to make data even easier to find, use and share will advance innovation and reduce unnecessary bottlenecks. Leveraging Snowflake as Autodesk’s machine learning (ML) query engine and source of truth is also on the roadmap. According to Kidwell, “We’re building the ML strategy to enable the most users with self-service access while following the principle of Snowflake first.”