Join our upcoming online event - "CDO TechVent: Data Observability- Thursday, August 25, 11:30 a.m. Eastern Time.

Snowflake Summit 2022: Growth, Vision, and Strategies

ABSTRACT: At Summit 2022, Snowflake paints a picture of a utopian data future.

For Snowflake, in-person conferences are back. The 2022 Summit, which took place in Las Vegas in June, was its first in-person conference since 2019. In spite of Las Vegas’ triple-digit temperatures, thousands of attendees and hundreds of exhibitors celebrated the return of business travel and the growth of a vibrant community of data gearheads.

The Snowflake community has grown significantly. In 2019, more than 2000 people attended Snowflake’s conference, with about 50 partners exhibiting. This year there were over 10,000 attendees and 200+ partners. Since 2019, Snowflake has also grown (see Table 1). Its staff has more than quadrupled and the company has increased its customer base six-fold. In his keynote address CEO Frank Slootman said Snowflake’s revenue grew from “basically a popsicle stand” to reach the $1b mark “faster than any company in the history of enterprise software”. 

Table 1. Snowflake Growth 2019 to June 2022

Snowflake Growth

2019

June 2022

Conference Attendees

2000+

10,000+

Partners

~50

200+

Employees

938

3992

Customers

948

5944

Revenue

$96m

$1.2b

Product announcements

These growth numbers formed the backdrop for a dizzying array of product announcements and strategies. They’ve made performance improvements, added financial governance features, improved data governance, added support for Apache Iceberg tables, added native support for Python, improved support for streaming data, a new workload and table types for transactional and analytical processing, a new focus on data application development, and monetization of data and applications. Like I said, dizzying. 

Many others have covered the details of Snowflake’s product announcements, the biggest of which was about the Unistore workload for hybrid transactional and analytical processing. Matt Sulkis of Monte Carlo wrote an excellent summary of Snowflake's new features in his keynote recap. I’ll focus on Snowflake’s vision and strategy, which they discussed at length and deserves attention. I’m most interested in the strategy behind so many varied product announcements because it provides a unifying context. Without that, it might seem as if Snowflake is charging helter skelter in a lot of directions at once.

The Seven Pillars of Snowflake’s Strategy 

Snowflake encapsulates their strategy in seven pillars of innovation that, according to co-founder Benoit Dageville, make the platform “truly unique”. Each of the pillars describes an area of focus that contributes to Snowflake’s vision of a unified data platform that handles almost all transactional and analytical needs.

  • All data. The first pillar, All Data, expresses Snowflake’s objective to process and provide access to any kind of data. This includes structured data such as database tables, semi-structured data such as application logs or from IoT devices, and unstructured data such as documents, images, or videos. All Data also means any volume and any origin—and not just internal data but also data shared from partners and suppliers, and external repositories.

  • All workloads. Snowflake strives to optimize its engine to deliver data for all use cases, such as highly concurrent interactive dashboards that need fast response time, training and running machine learning models, running data pipelines, and now even transaction processing for data applications. With storage and compute completely separated, each workload can be instantiated quickly and right-sized for the type of work it has to do. It also means that workloads don’t compete for compute resources, which gives the platform nearly infinite scalability. 

  • Global architecture. Snowflake creates a single distributed cloud data platform that can span multiple geographic regions and multiple cloud services such as AWS and Azure. Its Snowgrid abstraction layer provides a seamless experience for data consumers, meaning that they don’t need to know where data is to use it. 

  • Self-managed. Snowflake is designed to take care of itself. It provides a common data engine, security protocols, and APIs regardless of the clouds it’s hosted on. It eliminates administration tasks such as infrastructure management, workload management, and query tuning. 

  • Programmable. Snowflake enables developers to program in the language that they’re most comfortable with. Snowpark is a developer framework that currently supports SQL, Python, Java, and Scala through optimized APIs. A common engine processes data with common data access policies regardless of the language a developer uses. 

  • Marketplace. The Snowflake Marketplace enables organizations to share and monetize their data, their insights, or their applications with any other Snowflake account. With the Marketplace, Snowflake encourages collaboration among its customers and provides the incentive of new revenue streams to engage them in their vision of a global data cloud. 

  • Governed. The final pillar, Governed, means all data is always secured. Data is encrypted in motion and at rest. Role-based access policies are applied across regions and clouds. Attribute level access masks, hashes, obfuscates, or nulls data at the cell level and does so dynamically based on role and regional policies or regulations.  

These seven pillars guide all of Snowflake’s product development toward their overarching vision of the global data cloud.

The Global Data Cloud: Centralized and Distributed

In his keynote speech, Frank Slootman began his discussion of the global data cloud by saying, ”You are the epicenter of your [data] environment”. Snowflake has been talking about data sharing for years. But this is a powerful statement because it contrasts with where most organizations still are today. They are not the epicenter of their data environment. They are on data islands struggling to build bridges that connect with other data islands to gather information, place orders, deliver services, and other interactions necessary to conduct business.

Snowflake’s data sharing vision is that organizations can seamlessly share data regardless of organizational boundaries. This is possible when everyone is on one big multitenant database running across multiple geographically dispersed data centers and on multiple cloud platforms. 

Snowflake calls a data sharing connection between customer accounts an “edge”. Some are active edges, related to projects or trial relationships, that come and go. Others are persistent data sharing relationships that are integrated into the way two or more Snowflake accounts do business with each other. These are stable edges. Snowflake now has 1550 accounts sharing data in stable edge relationships, up from about 370 in 2019.

The notion of stable edges brings another contrast into sharp focus. Data architecture evolved from centralized, monolithic systems toward distributed modular designs over the last decade. Snowflake’s global data cloud is at once centralized and distributed: centralized because everyone is on the same platform; distributed because each account is a separate implementation. Their approach uses two opposing forces to produce an effect that neither one could achieve alone, like the centripetal and centrifugal forces that keep our planet orbiting the sun.

Conclusion: Snowflake’s Data Utopia

Snowflake paints a picture of a utopian data future in which all data is treated equally regardless of where it comes from or what language is used to process it, where processing resources are limitless, and all data is secured and governed. All accounts on the platform can seamlessly share data with each other if they so choose. And they can generate revenue by selling their data and applications in the Marketplace. 

It’s a compelling vision, but it only applies to those on the platform. So how does Snowflake plan to extend this vision to more organizations? One way is through their sales strategy that emphasizes targeting larger companies, like those in the Forbes Global 2000. Large companies have a gravitational force that pulls other companies along with them wherever they go. Data sharing through stable edges can be an incentive for companies to encourage suppliers and partners to join them on the Snowflake platform. It will be interesting to see how the many product features announced at Summit 2022 enable Snowflake to fulfill its vision.

Jay Piscioneri

Jay has over 25 years of experience in data technologies including data warehousing, business intelligence, data quality, and data governance. He's worked with organizations in a wide variety of industries...

More About Jay Piscioneri