Snowflake - What's going on?

Their chief competitors are AWS and Google Cloud. AWS offers a data analytics service called Redshift, and Google offers BigQuery. The competitive advantage that SNOW offers is apparently ease of use, cost, and being cloud-provider agnostic. Hope that helps!

That does help. It is amazing how many experts are willing to contribute on this board. No matter what the subject someone is willing to step up and give their knowledge to all of us. Thanks to everyone that gives to this board.

Andy

9 Likes

The difference between Report Generation and Analytics can be hard to understand at first. Basically, Report Generation is a straight forward process. A report might contain how many widgets your factory built each day, or how many widgets were sold, or how many problem tickets were filed. Analytics involves a deeper dive, often pulling data from different sources and finding non-obvious correlations that provide useful insight or actionable data.

smorgasbord

This is a request for clarification.

As a complete novice in this area I derived some generic conclusions from Snowflakes description of its capabilities.

First that they had developed something called a ‘datalake’ . The benefit of this was to permit seamless amalgamation of all forms of stored data from all sources and any warehouse and to make the data accessible for analysis and to developers. This as I understood it constituted an innovative and somewhat unique data storage structure offering a huge competitive advantage and a strong moat.

Second Snowflake talked about creating tools to make it simpler for developers to access and utilize the data.

How close a description is this to what is actually going on and does it explain the current level of exuberance.?

draj

2 Likes

First that they had developed something called a ‘datalake’ . The benefit of this was to permit seamless amalgamation of all forms of stored data from all sources and any warehouse and to make the data accessible for analysis and to developers. This as I understood it constituted an innovative and somewhat unique data storage structure offering a huge competitive advantage and a strong moat.

Data Lakes are common - they’re basically just a collection of data, often from a number of different sources, put somewhere so it doesn’t get lost. When a Data Scientist is tasked with producing insights they’ll pull data from the lake and put it into something (like Snowflake) from which they can run analytics. That can be part of something called ETL: Etract, Transform, Load. Think of it as prepping data for further analysis.

Amazon’s S3 is an example of a data lake. Snowflake does propose that its product can be used as a Data Lake, but it’s more often used as a Data Warehouse, which is more structured and more able to run analytics better. Usually with Data Lakes you want the cheapest reliable storage and don’t care about compute or performance since you access the data in it infrequently.

Read more here: https://www.snowflake.com/guides/what-data-lake

Second Snowflake talked about creating tools to make it simpler for developers to access and utilize the data.

That’s one of Snowflake’s advantages - that’s it’s easy to access data stored within it since it’s based on standard SQL queries, and because it’s integrated with lots of different BI (Business Intelligence) tools.

Read more here: https://www.snowflake.com/trending/right-analytics-tools-rig…

8 Likes

It should be noted that the plus of supporting SQL is that lots of people know SQL and there are lots of tools that communicate via SQL … not the SQL is a fabulous query language. One can, in fact, query a wide range of data sources with the DataDirect connectors ( https://www.progress.com/datadirect-connectors ). SQL is ancient, not new. In particular, SQL was created as a query language for relational databases, which makes it easy to understand, but may not be optimal when accessing a non-relational data source.

1 Like

What about SNOW’s business model makes them deserve a premium far beyond even its richest peer?

There is no question that SNOW is a great business. But we invest in other great businesses here at a fraction of the valuation of SNOW.

What is SO SPECIAL about SNOW that it deserves this kind of premium?

Thanks,
Rob

3 Likes

I work with Snowflake. Here are some points to consider. I don’t own Snow at this time but on the prowl.

Positives

Snowflake has a “data sharing” business model on the cloud. This has “network effects”, similar to Apps on iPhone.
Anyone in the world can share data seamlessly and monetize it. This sets up interdependencies.
Other products are not architected for data sharing and security.

Snowflake’s competition is weak

  • Appliances like Netezza, Exadata are history
  • Redshift, BigQuery - Snowflake is better, faster and cheaper
  • Hadoop/Spark ecosystems - too complex and slow
  • Mongodb, Cassandra are document stores - not relational/dimensional

Big Data and AI is hot, growing more rapidly than anything else.
Snowflake has latched onto the most lucrative item on cloud and this will continue to increase in the foreseeable future.
The nature of data is it always increases and rate of increase also increases. More easy Cha Ching.

Their roadmap BI/Reporting tool, PubSub tool, a data pipeline tool, connectors etc. is very well thought out.
They provide the enterprise ability to be multi cloud (aws, azure), so easier decision for management.

They are consistently the top 5 company to work for and attract talent.

Negatives:

I don’t like the fact that the company is NOT run by the founder but instead run by a hired hand.

It seems expensive for $109B but all these high growth companies do. I missed out on Google early due to this logic.

Weak or no support for unstructured (pdf, docs), audio, images and video data / processing.

25 Likes

“Weak or no support for unstructured (pdf, docs), audio, images and video data / processing.”

dividends20, thanks for your inputs.

Snowflake announced support for unstructured data in Nov 2020 ( which was in private preview).

Unstructured Data – The portion of the world’s data that is unstructured is quickly increasing. In addition to structured and semi-structured data, Snowflake announced support for unstructured data such as audio, video, pdfs, imaging data and more – which will provide the ability to orchestrate pipeline executions of that data. Unstructured data management in Snowflake means customers will be able to avoid accessing and managing multiple systems, deploy fine-grained governance over unstructured files and metadata, and discover new revenue opportunities thanks to gaining more complete insights. This feature is currently in private preview

https://www.snowflake.com/news/snowflake-announces-new-featu…

They did mention about that in the earnings call as well.

Cheers!

ronjonb

4 Likes

Is the support for unstructured data mean simply that they have added a BLOB datatype, i.e., one can store the unstructured data and retrieve it based on the structured data, but one can’t query the unstructured data itself. If so, this would seem to be a significant disadvantage compared to document databases.

1 Like

Is the support for unstructured data mean simply that they have added a BLOB datatype, i.e., one can store the unstructured data and retrieve it based on the structured data, but one can’t query the unstructured data itself. If so, this would seem to be a significant disadvantage compared to document databases.

You can query unstructured data like a BLOB or CLOB, it’s just extremely inefficient. But for the same reason, you can’t query the unstructured data in document databases either. You still need to provide a structure to get fast results. What a document database might have is tools that can parse the unstructured data for keywords/identifiers etc to make the querying faster, by adding those identifiers to a list that are then associated with the document. But you can’t just dump a bunch of images into a document database without spending a lot of time rigorously defining and extracting relevant data identifiers.

2 Likes

Searching images is going to be inherently difficult, but searching JSON CLOBs coming from an IoT source is one of those things which document databases are very good at … and one of the kinds of things one is going to want to do a lot of going forward … hence the wondering about what Snowflake can do here. Yes, with the document database one may have to do something in advance to make the search fast … on the nature of notice whenever you see such and such a tag and make note of where it is … but given that one can do that, it is very powerful and covers a lot of use cases.

Smorgasbord

Thank you for correcting a few misconceptions.

draj

Hi Dividends20 and other knowledgeable members,

I am curious on how do you view Talend (TLND) products vs Snowflake… TLND talks about data fabric that does appear to be similar to some part of SNOW products, at-least to casual observers… yet TLND is not really growing on overall basis… (although its cloud revenue is growing at very impressive rate).

TLND has built integration with SNOW. That integration allows to read from Amazon S3 buckets or Azure Container Storage and write the date to Snowflake.

Talend was one of the first partners to support Snowflake Partner Connect with the availability of Stitch on the portal. Additionally, Talend is certified as “Snowflake Ready Technology” and is one of only nine Elite Technology partners in the Snowflake Partner Network.

Does that bold section speak something?

There’s also recent news here if you’re interested in reading.

https://seekingalpha.com/pr/18108490-talend-expands-snowflak…

https://seekingalpha.com/news/3641386-talend-expands-partner…

Cheers!

ronjon

1 Like

nilvest -

Talend plays in a different space. It is “data integration” (ETL/ELT) set of tools.
It is datastore agnostic, meaning it can work with Snowflake, Oracle, Redshift and others.
Informatica has been the Gorilla in this space for long time. It is a crowded space. Talend is based on opensource and good.

A crude analogy would be

a) set of trucks/pipes/processes* that bring the raw materials from various places to the “Factory” would be Talend.

b) The Factory itself is Snowflake that stores this raw material and processes it in stages.
This factory scales dynamically (the usual elastic compute / elastic storage points for SaaS).

*processes push down can be done with Snowflake easily and role of tools like Talend/Informatica is less value added.

4 Likes

Talend has a 10 page e-book (PDF), written in 2016, with the title “WHY YOUR NEXT DATA WAREHOUSE SHOULD BE IN THE CLOUD”

Available here: https://info.talend.com/rs/talend/images/WP_EN_IP_Next_DataW…

And it’s heavy on touting Snowflake. Note this was before Talend was public.

1 Like

And it’s good to hear this straight from AWS CEO Andy Jassy( if you’re wondering the threat to Snowflake from Amazon Redshift).

He kind of thinks Snowflake on AWS as Amazon’s own product ( since they’re driving immense consumption on AWS)

This is a must watch video if you’re thinking about investing in SNOW.

https://www.cnbc.com/video/2020/12/01/amazons-cloud-chief-on…

I’m glad to have bought that small position in SNOW just after the IPO and am hoping that the lock-up release on Dec-14 will provide some buying opportunity. Will wait and see…

Cheers!

ronjonb

7 Likes

The two major mistakes I made in the past about growth stocks are: #1 Mistake. I sold out too early and bought back later at a higher price couple years later. e.g. Paycom, Shopify. Although I no longer own them. #2 mistake: I bought too little for the best companies. e.g. I bought couple shares of ZM, NET, DDOG, CRWD, PTON early 2020 and forgot about them! After almost a year, they went up couple hundreds percents and recently I’ve finally decided to make a full position in CRWD,PTON. and I missed out on all the big gains. I wished I bought full position in some of them at the beginning. Back them, I hold too many stocks and now I aim to hold 5 to 10 stocks.

For SNOW, I don’t want to repeat the same mistakes. For best company with good product and hyper growth, valuation is not a huge issue. The share price will follow the revenue growth rate until it slows down.

For concerns about competitions between AWS, Google, MSFT etc… I think of it this way. SNOW to infrastructure providers are like search engine to websites. There are two kinds of traffic to websites: Direct traffic and indirect traffic from search engine. Direct traffic do not compete search engine traffic. Both types of traffic benefit the website. SNOW is not a search engine but SNOW bring customers to infrastructure providers. Infrastructure providers like both types of customers no matter how it’s introduced.

I built a full 20% position in SNOW as of Dec 04, 2020.
Here are my purchase of SNOW and price.

Date Price Weight
Sep 16,2020 $269.74 0.82%
Nov 30,2020 $325.30 9.62%
Nov 30,2020 $326.60 1.85%
Nov 30,2020 $323.10 0.98%
Dec 04,2020 $380.50 1.16%
Dec 04,2020 $382.70 4.83%

Total weight: 19.25%

3 Likes

I built up a small starter position in Snowflake soon after IPO. The main thing that keeps me from building up a larger position is simply a lack of understanding of the business model and the pricing. Hoping to gain some insights from this board.

Risks:

  1. Model is not a recurring revenue model. CEO Slootman clarified this a no. of times and

See his comments at the start of the Q3 Earnings:
Before I discuss our results and guidance, I would like to spend some time discussing our unique and powerful business model. We are not a SaaS model. We are a consumption model. Our business model is a key differentiator for us and is designed to drive customer success. Our customers purchase credits and when those credits are consumed, we recognize the revenue. Unlike a ratable model, we only recognize revenue, if the customer uses our platform.

For this reason, there is no shelfware in our revenue. For many customers, it takes several months up until they are up and running at full capacity and this model gives them the flexibility to purchase the amount they plan on using without wasting credits or exceeding their original contract if they consume more than planned.

For these reasons, we do not focus on the same metrics that the SaaS business would. We focus on product revenue and remaining performance obligation. Product revenue, which excludes professional services and other revenue is the most transparent disclosure we offer. It gives full insight into how our customers are actually using our product in the period reported. If a customer purchases credits and does not consume, their revenue will be zero dollars.

Not currently sure if this is better or worse than SaaS as there is no limit to how much data a customer may use, and the revenue is built on consumption, so you may have many customers but no consumption for a while. The fact that Slootman interupted the interview and began with this discussion seems to me a very important point how it plays out with future earnings remains to be seen.

  1. Current short interest on the stock is 28 % which seems pretty high to me might be due to the tremendous valuation and short term run up.

Q3 Earnings Transcript is here:
https://seekingalpha.com/article/4392690-snowflake-inc-snow-…

2 Likes

Three thoughts, two that I need help diving deeper on.

  1. SNOW is 3.5x higher P/S than ZM while sporting 1/3 the Rev Growth rate of ZM. Hence, my question on market and valuation below. I apologize that it is technical in nature.

  2. Float - Shares Available - is 10% (28 million shares), this is a reason I’ve researched that the price continues to drive further. Those that own, simply may not want to sell.

2a. Salesforce ($250M) bought shares of IPO. Is this ~2.1MM shares a piece at the $120 / share IPO?

2b. Berkshire bought a reported 6.1MM shares

2c. Salesforce and Berkshire own 8.2MM shares of SNOW, or 25% of the outstanding shares

If this is the case, then I think the fact that most shares are owned by longterm investors who are unwilling to part with the shares (Berkshire infamously), further impacts price action.

Anyhow - does this mean the lockup date matters or not? Typically, I’ve never seen ANYTHING significant from a lockup date. However, perhaps this is similar to APPN?

Going back to #1, I’ll add another flavor, CRWD at 85% growth with superior margins is valued 5x less than 109% growing SNOW.

I don’t think valuation by itself is a reason to avoid a stock, but I do think relative valuation should make us ask, why?

Just a Fool

10 Likes

ron, smorgas, dividend.

thanks for clarifying TLND w.r.t to SNOW. That makes sense.
I will wait for SNOW lock up period before starting to build a position.