Following up on my last reply…and happy weekend!
Allow me to call this “Snowflake Demystified”
As tech companies are solving unique problems at the right time, the definition of the cloud is also evolving. So, you can think about three layers of cloud like…
Cloud Infrastructure: ( Azure, AWS, GCP)
Cloud Database/ Data Cloud: (??)
Cloud Applications/Services:( e.g. all SaaS companies like CRWD, CRM, DDOG, DOCU etc)
Well, Snowflake fits in that Cloud Database layer. Snowflake is build for the cloud and is built for performance, sclability and security of data. BUT, there’s a BIG difference. You won’t find many names in that Data Cloud bucket.
So the picture today is more like this…
Cloud Infrastructure: ( Azure, AWS, GCP)
Cloud Database/ Data Cloud: ( Only SNOW)
Cloud Applications/Services: ( e.g. all SaaS companies like CRWD, CRM, DDOG etc)
Just a few years back, I worked extensively with big data and was quite proficient in extracting, transforming and loading data ( also known as ETL). I used my skills to create metrics and data reports that were being used to drive new feature development and A/B testing at my company and also measuring the success of the business ( like calculations of DAU, MAU, ARPU, LTV etc). It was super-easy for me as I was a domain expert and worked in my own silo.
However, if you asked me to work with datasets from disparate sources, the complexity of my work would multiply many times. And if that data happened to reside in different public cloud providers, I would almost give up!
So, if someone came with a magic wand and told me that I don’t need to worry about where the data originates from and can just use that data and analyze it as if it resided in just one place; it would indeed be magical! Well, that’s exactly what Snowflake is doing. It helps Data Engineers and Data Analysts to do their job without the need of worrying about actually where the data originates from.
For the benefit of my non-techie friends, I’ll try my best to explain this without any tech jargon in it’s simplest form.
…Say your application is generating raw data like who’s using your service, how much time they are on it, what platform they’re on (iOS, Android, Windows), how much they are spending etc ( all stored as events in a AWS S3 bucket). You simply load that data into a Snowflake instance ( think of this as your Snowflake Cloud database and some virtual warehouses to process that data. Once the data is processed you can now use a BI tool to build reports and analyze that data.
The power and beauty of Snowflake is that you can simply do the above operations for data residing anywhere ( AWS, Azure, Google Cloud). Snowflake works with both structured and semi-structured data ( like JSON) very easily.
So, that’s basically what Snowflake does. What’s unique is that, no one else does that stuff better than them as of today and their pricing is pay only when you use. ( If I were on the Snowflakes sales team, that would be my sales pitch to Mr.Buffett :))
And Snowflake uses standard ANSI SQL, which most Data Engineers/ Analysts are familiar with and it makes their work easy. Snowflake also handles all indexing and partitioning automatically. Their multi-clustered shared data architecture allows data to be analyzed as it’s loading which is really cool.
If you haven’t heard about this, Snowflake has a Data Marketplace for sharing data. A good example is the Starchema dataset and Weathersource which helps relate COVID-19 data with the impact of Weather. AFAIK, This data has been requested by more than 2000 Snowflake customers. And you could also use datasets like Factset and numerous other data providers that make their data available on Snowflake Data Marketplace along with your own business data to build insights.
I won’t go into more tech details but would like to mention that Bob Muglia, the former CEO of Snowflake was a member of my leadership team at MSFT and I have great respect for him.
As mentioned in the previous post of this thread, I bought my first position of SNOW just after the IPO at around $240 and it’s up 61% as of today ( not too bad :)).
I won’t be selling my SNOW shares unless I see a competitor come up with a better solution. Right now I don’t see any on the horizon. This also somewhat explains why their net revenue retention rate is 162%.
As someone rightly said “data is the new oil” and I say “you need refineries to process that crude into more useful products”. That’s what SNOW does with all that data. I may keep adding on dips as it’s currently sitting at the bottom of my portfolio at around 5% ( after the 61% rise).
Hope you find this helpful!
Long <CRWD, DDOG, ZM, DOCU, FSLY, NET, SHOP, PTON, SNOW > and some SPACs.
P.S. Please understand that it’s very hard for me to express my thoughts without bringing in technical jargon but for the benefit of all users in this board. I’ve tried to explain what Snowflake does in very simple language. Email me if you’ve any technical queries and I’ll try to find time to answer them if my daily job permits
I created my twitter account @algocodetest today out of curiosity after seeing many in the twitter community caught off-guard with Snowflake’s ascent. I won’t be posting much there ( as I’m very much time constrained).