Autodesk builds a culture of self-service data for 13,000+ employees

Autodesk is a Fortune 1000 company that creates software products for a variety of global industries. With over 100 products that span architecture, engineering and construction, product design and manufacturing, and media and entertainment, you’ve likely come across a place or product with Autodesk software at its core, from city skyscrapers to movie imagery. 

With a large portfolio of software comes a lot of data. In fact, it’s petabytes of data. The business turns to Mark Kidwell, Chief Data Architect of Data Platforms and Services to manage its Analytics Data Platform. 

Mark’s key responsibilities are to:

• Modernize and optimize the data platforms with best-of-breed tools to improve performance while managing costs
• Accelerate the delivery of data to the teams who need it for analytics, including marketing, sales, revenue and product

Delivering data-as-a-service with reliable pipelines 

The Analytics Data Platform at Autodesk is the go-to center for teams to understand critical product usage – everything from subscriptions to customer behavior to the health of the business. But with a data lake architecture, accessing data for analytics was complicated and fragile. 

“We couldn’t even come close to enabling self-service with any of our previous tools and processes. We had a central ingest team that solely took ingest requests – they were the only team that could do that work.”
- Mark Kidwell, Chief Data Architect of Data Platforms and Services

To enable self-service data to more than 13,000 employees across the enterprise, Mark knew the analytic data sets would need to be housed in a data warehouse to be easily queried. Mark looked to modern data tools, including Fivetran, Snowflake and dbt, to revamp the ingestion, transformation and orchestration. 

“It’s as simple as choosing between an unmanaged, low-level pipeline that breaks, or a pipeline [like Fivetran] that handles schema evolution, back-fill and other cases that we know we can’t support,” Kidwell adds. 

This move towards a modern data stack was a large business effort involving multiple business units, including two of the recent four construction product acquisitions: BuildingConnected and BIM 360. The Analytics Data Platform team moved all of its business-critical data to run on Snowflake. 

The initial phase involved replacing just a handful of pipelines. After these were successfully migrated and the value of features like schema drift handling and incremental updates was proven out, the team moved onto its most critical data sources.

“We migrated our Salesforce pipeline, our most critical ingest product, to Fivetran. We’ve unlocked more users and more use cases along the way, so our usage has grown quite a bit over the years.”
- Mark Kidwell, Chief Data Architect of Data Platforms and Services

The result is a reliable data movement and warehousing solution that ultimately saves the business time and money. “It just works,” says Mark. 

[CTA_MODULE]

Enterprise-grade capabilities that address Autodesk's complex, highly-secure data requirements

The Analytics Data Platform team quickly and securely onboarded teams and users to Fivetran to add their own connections. Fivetran provides users with access to the pipelines and destinations they need to do their jobs. Enterprise data governance features like team and connector permissions, custom roles and team access, and single sign-on enabled the Autodesk team to scalably enforce platform security with its large user base.

“We’ve successfully migrated nearly 60 data teams that are ingesting, consuming and sometimes publishing their own data products onto the Analytics Data Platform. Over half of those teams are working within Fivetran, assigned to different destinations and different roles. This self-service motion means that we set the platform rules, and the teams can access their data without coming to us. ”
- Mark Kidwell, Chief Data Architect of Data Platforms and Services

Modernizing the data lake

With the foundation for self-service in place, Mark and his team are focusing on making the platform even easier to use for internal customers and exploring additional use cases like:

• Metadata API: Autodesk recently selected Atlan for its data catalog and plans to integrate that with Fivetran this quarter to enable data lineage.
• REST API: To become an organization that can build services and products that are API first, and Fivetran’s API will be key to their efforts.

[CTA_MODULE]