Snowflake held a Snowday event on November 7th in San Francisco, as part of its Data Cloud World Tour. Similar to prior Snowdays, the Snowflake team used the occasion to showcase new product innovations and to interact with customers. During the event, they made a number of product announcements and program updates. The product team emphasized that these were not just a rehash of announcements made during the Summit conference in June, reinforcing the fact that Snowflake has been continuing to innovate on its platform.
Audio Version
View all Podcast Episodes and Subscribe
These Snowflake customer events are worth monitoring by investors, as they provide signals around product strategy, traction and the opportunity ahead for Snowflake. The keynote presentation and product specific segments are available to view online.
The first major announcement from Snowday was the general availability of Snowpark for Python. At Snowflake Summit in June, the team had promoted Snowpark for Python to public preview. Snowpark is Snowflake’s developer framework with DataFrame and custom function support, used by data scientists, data engineers and application developers to build efficient data pipelines, machine learning workflows and data applications. These capabilities are supported in multiple languages, including Java, Scala, SQL and now Python. As part of Snowday, the Snowflake team highlighted a number of improved capabilities, including support for streaming, Iceberg tables and external tables for on-premises storage.
The addition of Python was important to broaden adoption as Python is the most popular language for data science. Included in the addition of Python is support for the Anaconda libraries, which add a variety of common functions to data science use cases.
Since its public preview announcement in June, Snowflake leadership shared that demand for Snowpark for Python has exploded, with customer adoption increasing by 6x. They now have hundreds of customers making use of the capability, including Charter Communications, EDF, NerdWallet, Northern Trust and Sophos. The benefit these companies cite is the ability to build applications with access to their data directly on Snowflake (versus porting it to another data source).
Snowpark for Python has created new opportunities and use cases for our team to build and deploy secure and compliant data pipelines on Snowflake, so we can more efficiently provide our customers with the tools needed to handle every aspect of their finance journey. Snowflake’s continued investments in Python allow us the flexibility to code in our programming language of choice, and accelerate the speed of innovation for our end users.
Sathish Balakrishnan, Director of Data Engineering, NerdWallet
Included in this was the announcement that Snowflake has been moving forward on the integration of Streamlit into the core platform. This will allow developers to build data visualization applications on top of their data within Snowflake, maintaining security and governance over the data by obviating the movement off to another platform. They expect to have private preview Streamlit integration available at the beginning of 2023.
Streamlit serves as the interaction engine for the vast majority of our Data Science & Machine Learning models today, actively transforming how our teams build, deploy, and collaborate on powerful applications with other stakeholders across the business. With Snowflake’s Streamlit integration, we can go from data to ML-insights all within the Snowflake ecosystem, where our data is already present, making it easier and more secure for us to create impactful applications to further mitigate the negative impact of flight disruptions, provide more predictability to our operational planning teams, and more customer personalization to give our customers the best possible experience.
Sai Ravuru, GM Data Science & Analytics, JetBlue
Other new capabilities announced included Dynamic Tables and Observability features. Dynamic Tables were previously introduced as Materialized Tables at Summit. Dynamic Tables supports streaming of data by pulling data from multiple sources, automating incremental processing of it and forwarding it to a streaming pipeline. This allows data to be processed on Snowflake in near real-time.
To improve the developer experience, Snowflake is building native observability features into pipeline monitoring so that developers can build, test and maintain them. They are adding support for alerting, logging, event tracing and history.
Snowflake’s SVP of Product also shared an update on Iceberg Tables, which was introduced at Summit and is in private preview now. More than 50 customers have participated in the private preview, including Dropbox, Indeed and DriveTime. Announced at Snowflake Summit, Iceberg Tables combine unique Snowflake capabilities with the Apache Iceberg and Apache Parquet open source projects, to provide customers with another way to interact with their data inside of Snowflake. As an open source format, it addresses customer concerns around lock-in to a proprietary system design. The most common question from customers is when Iceberg Tables will be generally available.
Snowflake introduced their new transactional workload, Unistore, at Summit as well. Applications can interact with Unistore as a transactional database with support for fast reads and writes, but the data is also immediately available for analytics and large query processing. This prevents the need to move data from a transactional database to a separate analytics database. The SVP of Product mentioned that “demand is through the roof” for this. Customers include Novartis, Tecton, Iqvia, UI Path, Adobe and Tapestry.
Sponsored by Cestrian Capital Research
Cestrian Capital Research provides extensive investor education content, including a free stocks board focused on helping people become better investors, webinars covering market direction and deep dives on individual stocks in order to teach financial and technical analysis.
The Cestrian Tech Select newsletter delivers professional investment research on the technology sector, presented in an easy-to-use, down-to-earth style. Sign-up for the basic newsletter is free, with an option to subscribe for deeper coverage.
Software Stack Investing members can subscribe to the premium version of the newsletter with a 33% discount.
Cestrian Capital Research’s services are a great complement to Software Stack Investing, as they offer investor education and financial analysis that go beyond the scope of this blog. The Tech Select newsletter covers a broad range of technology companies with a deep focus on financial and chart analysis.
Performance Enhancements
In addition to adding new features and capabilities, Snowflake continued its background work to make the core engine more efficient and lower cost. Customers receive this benefit automatically, gaining more performance for each Snowflake credit without having to make changes. Snowflake leadership introduced two capabilities in this domain. First, processing of sporadic large queries can streamlined, as Snowflake can apply a temporary burst of additional resources. This creates cost efficiency by preventing the need to maintain more capacity for short bursts in data processing.
Second, Snowflake added cost optimization features. Customers will be able to run a cost-benefit analysis to determine the magnitude and impact of data loads or modifications on tables through new history views (private preview), and lower costs with Search Optimization now supporting column selection (public preview) for data optimized for searches. Additionally, customers can now analyze long-running and time consuming queries more easily, and identify and resolve performance problems before they impact the end user through programmatic access to query metrics (in private preview).
The Snowflake team is also making enhancements to their cross-cloud Snowgrid capability. Snowgrid provides an interconnect layer between Snowflake instances across hyperscalers, and is responsible for much of Snowflake’s collaboration and data governance capabilities. For collaboration, new enhancements include listings for private sharing, streamlined access to data shared across clouds and richer user analytics. For data governance, Snowflake has added features to help organizations comply with international regulations and gain unified visibility across their data. Improvements include automated data protections and accelerated searches on Masked Data.
To help maintain continuity and resiliency, Snowflake added pipeline failover to its other data replication capabilities. Set up and management of failover can be controlled through an intuitive UI, that includes controls over configurations for sources, destinations, replication objects and replication intervals. This failover capability is pretty slick, as it can work across different cloud providers (AWS to Azure for example).
Data Cloud Ecosystem and Powered By Program
Finally, Snowflake announced some exciting progress around the data cloud ecosystem with their industry solutions and Powered by Snowflake program. Powered By has been experiencing extraordinary popularity, increasing in participant size by 6x year/year as of July 2022. In fact, during the Keynote, Snowflake’s SVP of Product shared that the 4 fastest growing companies from $1M to $100M in ARR are built on Snowflake. Powered By provides participants with tools and resources to build, market and operate applications in the Data Cloud. Some example participants include:
- Blue Yonder: Supply chain solution provider
- Panther: Cybersecurity
- Piano: Media and advertising
- Simon Data: Marketing technology
They also highlighted the launch of four separate industry-focused Data Clouds to date, with over 120 partners supporting them. The four industry-specific Data Clouds include the Financial Services Data Cloud, Media Data Cloud, Healthcare and Life Sciences and Retail Data Cloud. Within each Data Cloud, Snowflake collaborated with partners to provide interoperability, secure data sharing and best practices per industry. These industry-focused Data Clouds serve to attract companies within the same industry to share data in a controlled environment.
Snowflake has an earnings report coming up on November 30th. I expect continued growth with large customers, particularly as Snowflake continues to roll out these new capabilities. We could see some softening of demand due to the macro environment. I think the interesting counterbalance will be if we start to see meaningful revenue generation from the Powered by program. As these companies are building their business on Snowflake, they can quickly ascend past the $1M customer threshold. During the Investor Session at Summit, leadership shared that 9% of $1M+ Snowflake customers are part of the Powered By program. I suspect this will increase as more developers choose to build and monetize applications in the Snowflake Marketplace or run stand-alone businesses that use Snowflake as the data processing layer.
NOTE: This article does not represent investment advice and is solely the author’s opinion for managing his own investment portfolio. Readers are expected to perform their own due diligence before making investment decisions. Please see the Disclaimer for more detail.
Additional Reading
- For more coverage of Snowflake’s Snowday, please check out the coverage over at Hhhypergrowth. Muji goes into a lot of detail on the event and adds additional perspective on the product direction. This is a premium post, but well worth the subscription cost.
Thanks for the take-aways.
Really interested to have your view on recent AWS re:Invent announcements and their impact on Snowflake. Products/enhancements like AWS Clean Rooms, Zero ETL integration between Aurora and Redshift, Redshift integration with Spark seem moving towards SNOW’s path. Is competition firming up?
I would also be interested. I don’t have a good understanding of it, but I think AWS might be at least trying for better integration of their many products. On Gestalt IT, they praised Amazon’s DPU, and said the latest Graviton processor is OK (but not brilliant) at floating point arithmetic. Apparently that’s been a weak area for ARM processors, while X86 CPUs have accelerators for it. One point was, AWS Nitro could reduce cost, but also lock in customers, and I don’t know how customers will split between avoiding cost and avoiding lock-in. I also don’t know if all the benefit’s of those things are available via Snowflake.
Hi Nikos – it’s hard to interpret how these moves will impact Snowflake. On one hand, AWS is clearly creating products that target similar use cases as Snowflake. On the other hand, AWS’s success with these copycat products has been mixed, if not poor (DocumentDB, Elasticsearch, Contact Center, etc.). I think it is a positive for Snowflake that AWS sees enough customer interest to justify the investment of resources into data sharing and a single data store.