With all the interest over the Snowflake IPO, I wanted to share a few thoughts. Now I am not the right person to tell you if these valuations will pay off and for whom. You’ll have to look elsewhere for investment advice. But I can talk about what is happening with the product, which is driving a movement.
Snowflake is a leader in a new class of super scalable analytics platforms, and these platforms are becoming central to the explosion of analytics all around us.
Now, there are at least four serious contenders in this space (Amazon Redshift, Microsoft Synapse, Google BigQuery, and IBM Db2), and they are consistently leapfrogging each other with capabilities. However, Snowflake is well-positioned as arguably the first to market, but there is plenty of space for all of them.
While data warehouses have been central to analytics for at least a generation, it is improvements in architecture and a change in practices that have allowed these platforms to take off.
There have been several changes in the architecture of these systems. But what it really comes down to is leveraging the elastic computing model which allows customers to start as big or small as they need and then scale up or down on demand. This is made possible through operating in the cloud, major advances in database architecture, and the separation of compute from storage.
Combined with a pay as you go usage-based model, companies can now avoid big platform investments up-front locking them into a fixed capacity.
The new architecture has also facilitated an equally important but sometimes overlooked change in practices.
In earlier generations of data warehousing, much of the business logic (or processing) was built into the loading the data. This created the dual problem of having to define business logic upfront and a flexibility problem that would require the data to be reloaded or reprocessed when the business logic needed to be changed.
This approach made for fast analytics for a narrow well-defined set of problems but took a long time to implement and was incredibly rigid.
Alternatively, current practices allow the data to be loaded in relatively raw form into the storage which best suits the data without predefined business logic. Business logic can be built outside the analytics platform without changing the raw data, and processing can happen when needed.
This approach has put the definition of business logic closer to the problem and allows new business applications to be developed as needed without reloading the data.
Why does this matter?
Together the innovations in architecture and practices have made enormous improvements in the time it takes to implement an analytical application, provided a highly efficient cost model, and allowed flexibility to solve new problems incrementally. And these improvements are bringing big benefits to customers and attracting more and more of the analytics workload.
So, in a world where data is exploding and analytics is central to transforming so much of our business, these platforms are experiencing massive growth.
There is no doubt other analytics platforms are in the race, but Snowflake is one of the leaders set to ride this wave. And it is this recognition which is at the heart of the investment explosion.