Snowflake — the company everyone should be talking about + its ridiculously overvalued IPO

I was recently reading Scott Galloway’s (professor of marketing at NYU, Stern) newsletter. While the entire newsletter is about Airbnb’s valuation, Scott also mentions the company Snowflake, a cloud-based data warehousing company, in passing.

Snowflake, a company that has solid business fundamentals (more on that below) was founded in 2012 and had its IPO in September 2020. The IPO holds the record for the biggest software IPO EVER. I hadn’t even heard of Snowflake until I read about it in Scott’s newsletter (which I highly recommend, by the way. Scott is awesome).

Before I knew it, I was deep down a rabbit hole on the internet, about all things Snowflake. This company is something. And their massive IPO made them (Snowflake) an even bigger talk of the town.

Through this post, I want to talk about three things-

1. What Snowflake does as a business and how it is different from its big-tech competitors (Amazon, Microsoft, Google)

2. Snowflake’s recent IPO along with all the exciting events that were in the run-up to it. (Warren Buffet’s Berkshire Hathway invested $750m in Snowflake, and Salesforce has $250M of their funds on Snowflake’s IPO)

3. Why Snowflake has not made any profits till date

1. First things first, what does Snowflake do?

Whenever I try to understand anything (literally anything), I go to Wikipedia. And I did not do anything different, here.

Wikipedia says this, about Snowflake-

“Snowflake Inc. is a cloud-based data-warehousing company that was founded in 2012 by 3 data warehousing experts, in San Mateo, California. Snowflake offers a cloud-based data storage and analytics service, generally termed “data warehouse-as-a-service”. It allows corporate users to store and analyze data using cloud-based hardware and software.”

-Wikipedia

If that didn’t make enough sense, the next bit will.

But before I get there, I’d like to convey my personal feelings towards the term ‘data warehousing’. ‘Warehousing’ sounds like a boring activity that happens with your data in a dark old basement when that is actually not the case at all. Lots of cool stuff happens in data warehousing I wish we had a better name for it.

Anyway, I digress.

Snowflake is a SaaS (Software-as-a-service) based company. So no hardware. They are a pure software company. They are a company that was born in the cloud.

(‘Born in the cloud’ means that the company was founded at a time where all of its IT assets have always been and are currently in the cloud. It has never owned physical servers or understands what a data center is).

And therefore, Snowflake is not a data storage company (since it primarily a software/SaaS tool). While they still offer storage services, that is not what companies primarily hire them for. Snowflake’s core value prop is its analytics and data optimization. It is more likely that clients have their data sitting on Amazon’s AWS/ Microsoft’s Azure/Google’s/IBM’s Cloud etc and hire Snowflake to do something with that data. Something like organizing the data into a readable form and also gathering business insights from the data (among other things).

Snowflake runs on all popular cloud provide platforms. It started with running on AWS (back in 2014) but is now compatible with Azure (since 2018) and GCP (Google’s cloud since 2019). What this means is that a company that has its data sitting on AWS/Azure/GCP (or even a combination since many companies use multi-cloud today) can use Snowflake to run analytics on that data. Previously, Snowflake was compatible only with AWS. So your data had to be sitting in AWS for you to be able to use Snowflake. But now, Snowflake works with all clouds.

It’s obviously in the best interest of Snowflake to be available as a SaaS offering in not ONE but ALL clouds because more clouds equals more clients equals more revenue.

Additionally, if you have data sitting across multiple clouds (AWS, Azure, GCP — a lot of companies today can use a combination of Azure and AWS to store their data), Snowflake is also able to gather all of the data into one place and analyze this comprehensive set.

As a Snowflake customer, you can simply load your data and start querying. Without having to worry about everything that goes on in between (Because Snowflake is doing that for you)

“Food delivery service DoorDash says moving its data to Snowflake in early 2019 has doubled the average speed of a data request at half the cost of its old system.”

-Fortune, Sep 2020

Here is a 7-minute graphic filled video that summarizes what they do, pretty neatly.

So is Snowflake competing with Amazon, Microsoft and Google?

The short answer is YES.

They are competition to the big-tech cloud leaders. But it’s a little more nuanced than that.

Snowflake is also a partner to these big techs.

Because a client storing data on AWS may want to use Snowflake exclusively for their data warehousing service while using AWS for everything else.

Hence, to this extent, Amazon partners with Snowflake to ensure that its platform is able to host Snowflake’s solution.

Why does AWS have to offer Snowflake as a solution on its platform?

Amazon has no choice but to offer Snowflake’s solution on its platform. It has to ensure that its clients have the luxury of a variety of cloud solutions to choose from. And if Snowflake is the service of their choice, they should have that choice.

If AWS does not have this variety, they run the risk of a client leaving to a competitor (say Azure) which has Snowflake on its platform. This is how and why most clouds, though competing with Snowflake, partner with them.

Snowflake and Amazon are competitors too.

Because AWS has a database warehousing service of its own. It is called ‘Amazon Redshift’. Amazon would like clients to use their Redshift service for warehousing instead of Snowflake’s.

However, Snowflakes’ unique value proposition (which I discuss soon, below) may lead a client to pick Snowflake over Red Shift.

Similar to Amazon’s ‘Redshift’, Snowflake competes with Microsoft’s ‘Azure SQL Datawarehouse’ and Google’s ‘BigQuery’.

Snowflake is a classic example of ‘competitor and partner’ for these big techs.

In fact, there have been multiple cases of clients that used Redshift for years but have recently started switching to Snowflake, leading to a permanent loss in revenue/client for AWS. Similar thing with Azure and GCP. The rivalry got to such a point that Microsoft started explicitly enticing its customers away from Snowflake to use its own product.

“Microsoft has been selling Snowflake’s database product on its Azure cloud platform since 2018. But recently Microsoft upgraded its own data analytics product and has been trying to persuade its customers to skip Snowflake in favor of Microsoft products, the people said.”

- The Information, May 2020

What makes Snowflake unique from its competitors?

The primary difference between the data warehouse services of Snowflake and those of Microsoft, Amazon and Google’s is that Snowflake allows clients to pay for storage and compute separately. While the big techs often bundle storage and compute together, Snowflake ensures that the client has the freedom to individually opt for both according to their need and consequently, pay proportionately.

This independence of compute and storage functions also allow for quicker analysis and reorganization of data in the back end, leading to a better and seamless client experience.

“Snowflake’s architecture allows similar flexibility with big data. Snowflake decouples the storage and compute functions, which means organizations that have high storage demands but less need for CPU cycles, or vice versa, don’t have to pay for an integrated bundle that requires them to pay for both. Users can scale up or down as needed and pay for only the resources they use. Storage is billed by terabytes stored per month, and computation is billed on a per-second basis.”

- Stitchdata

Snowflake has gradually been stealing clients from its competitors IBM and Teradata (This Forbes article speaks to that)-

In fact, it is rumored that the bank, CapitalOne which, for the longest time worked with Teradata (Snowflake’s competitor), recently switched over to Snowflake. CapitalOne believed that Snowflake was a better candidate than Teradata, in leading and managing the bank’s move to the cloud.

IBM customers that have moved to Snowflake — OnDeck, Linden Lab, and eXelate/Nielsen

Teradata customers that have moved to Snowflake — Capital One, Office Depot, and ARC.

“CapitalOne picked Snowflake for its journey to the cloud. CapitalOne hadn’t taken full advantage of the cloud and were looking for a shift forward. They concluded that Teradata could not take them into the cloud. And they picked Snowflake because it could run 250 concurrent data analysis queries — compared to 60 for Teradata — at a much lower price (25% to 30% of what Teradata charges),”

- Bob Muglia, former CEO of Snowflake

Here is a 2-minute video — Capital One’s own testimony of their work with Snowflake.

Oracle is another big competitor of Snowflake. (In fact, 2 of 3 founders of Snowflake were former Oracle employees). Other hot startups that Snowflake competes with are Panoply, Incorta, and Yellowbrick.

And, in case you are interested in reading further on competition, Here is a comprehensive analysis (down to the last detail) of how Snowflake differs from its competitors in offerings and architecture.

2. The biggest software IPO, EVER

And now to the more spicy part — the IPO

The company listed itself on the New York Stock Exchange (NYSE: SNOW) on Sep 16, 2020.

It’s share price before going into the IPO was rumored to be at $75-$85.

The official price that Snowflake IPO’d at was 120$.

The price of Snowflake closed at 254$ on the first day. So if you invested in the IPO, you’d have made more than double your money.

At a closing price of $254 (on IPO Day), Snowflake’s valuation of ~$70B makes it the biggest software IPO EVER. This valuation of over $70B makes Snowflake bigger than 2 Twitters, VM ware, GM, Ford and many big companies that have existed for decades.

How did this happen? How did they get such a high valuation so quickly?

Let’s rewind the clock just a little.

In Feb 2020, Snowflake raised $479M (led by Dragoneer Investment Group). At this point, it was valued at $12.4B (which was already triple the company’s previous valuation at $3.9B. So basically from being valued at 3B to 12B to 70B and today, 75B — all of this in the span of one year. It’s ridiculous. And definitely overvalued.

Let’s look at a simple revenue multiple analysis of Snowflake. I say simple because revenue multiple is very simple to understand and is often used to understand whether a company is overvalued/undervalued/rightly valued.

Revenue multiple is also referred to as ‘price to earnings multiple’ or ‘sales multiple’.

If a company earns 100$ a year and is valued at $1000, it’s revenue multiple is 10. That is, for every dollar that the company is earning, the market values it at 10$. (Value/Revenue = 1000/100)

Snowflake made revenues of $402M in the latest year. They are valued at $70B (give or take)

That gives it a revenue multiple of 187. (70B/402M =187 also written as 187x)

Hence, Snowflake is valued at 187$ for every dollar that it makes. This is very high for a revenue multiple in any industry.

Obviously, this valuation reflects more the future earnings potential of Snowflake than its past performance. Nevertheless, a valuation of 187 times its revenues is bound to raises eyebrows.

Just to tell you how ridiculously high Snowflake’s revenue multiple of 187 is, let’s look at revenue multiples for Snowflake’s big-tech competitors. (Most recent revenues and valuation (all figures in USD))

Alphabet’s revenues are $166 B. They are valued at $1.1 T. That is a revenue multiple of 6x

Microsoft’s revenues are $143 B. They are valued at $1.6 T. That is a revenue multiple of 11x

Amazon’s revenues $322 B. They are valued at$1.6 T. That is a revenue multiple of 5x

Oracle’s revenues $39 B. They are valued at $180B. That is a revenue multiple of 5x

So while competitors are valued at anywhere between 5–10 times their respective revenues, Snowflake is valued at 187 times its revenues.

Snowflake’s revenue multiple is way higher than it’s competitors, which makes it a very overvalued company according to most analysts

Now that we understand the possibility of an exaggerated valuation (though they are solving an extremely meaningful business problem), let’s talk about what made the IPO a big success. How did they manage to gather so much confidence from the market?

Two events that mad the IPO, a success, in my opinion–

First, Snowflake’s CEOs —

Right from the beginning, Snowflake has had very strong leadership. Two of their three founders were ex-Oracle (which understands data warehousing very well) and the third founder had co-founded another cloud database company before Snowflake (Vectorwise, now called Action Vector). After working in stealth mode, Snowflake finally started raising money and disclosing its business model. (In business, stealth mode is a company’s temporary state of secretiveness, usually undertaken to avoid alerting competitors to a pending product launch or another business initiative).

This is when they appointed former Microsoft veteran Bob Munglia as the CEO. Bob understood the database warehousing world very well and led the company to tighten its fundamentals, raise money, and increase the popularity of Snowflake. They were able to land big clients like Adobe, CapitalOne, and Nielsen. Bob took the company to the next level.

In May 2019, there was news around Snowflake’s IPO and that’s when they brought in Frank Slootman as CEO, replacing Bob Munglia. Frank Slootman was sailing for 2 years and was called to be offered the option to return from a holiday to lead Snowflake. He took it. Frank Slootman is popular for taking 2 companies public before this — Data Domain and Service Now, both of which were successful IPOs. Snowflake was going to be the third company for him, to take public. He is quite the IPO expert and his reputation preceded him (in a good way). Lots of investors wanted to be in after they heard about Frank being at the helm.

Here is a 10-minute interview of Frank Slootman with CNBC immediately after the IPO, where they chat about everything, Snowflake.

Second, the vote of Confidence (from Berkshire Hathaway and Salesforce) —

Apart from Frank Slootman’s leadership, Snowflake had other things going for itself in the run-up to the IPO, which made the IPO such a blockbuster.

One of these was Berkshire Hathaway’s (BH, hereon) investment in the IPO. The word is that Warren Buffett (principal at BH) has given Todd Combs, a rising investment manager at BH (rumored to succeed Buffett) the authority to lead tech investments. And therefore, it is highly likely that Todd led this $750M+ investment (part IPO stock and part purchase of shares from former CEO, Bob Munglia). BH more than doubled their money in a single day with that investment.

This is not all. THE cloud CRM company, Salesforce has also invested $250M in the IPO of Snowflake.

Having both these names along with credible VC investors like Sequoia Capital and Redpoint Ventures gave it the vote of confidence and stamp of approval. This definitely led to the IPO being a big success, for Snowflake, increasing investor confidence and driving the stock price to more than 2 times its initial listed price (of 120$).

Side note — The last time BH invested in an IPO was in 1956 when Ford was going public. Hence, it was headline news when BH was going in on Snowflake’s IPO with almost a billion dollars. (after 64 years!!)

Snowflake — Timeline of events- From founding in 2012 to IPO in 2020

3. Snowflake is yet to make profits

With most of their business value proposition and IPO outlined in the first two sections, it may be interesting to point out that Snowflake is yet to make profits. While Snowflake’s revenues are expected to increase at a high rate, they have been making losses so far. For 2019, Snowflake recorded revenues of $265M but a net loss of $350M. And the losses are expected to continue for the next 2–3 years. This is primarily due to their high spend on marketing and advertising.

Snowflake’s spend on marketing alone is more than its revenues. (For 2019, Snowflake spent approx. $300M on marketing and advertising while its revenues for the year were $265M).

The trend is expected to continue with losses reducing and Snowflake eventually expects to make significant profits. As it adds more blue-chip companies to its roster of clients, Snowflake will continue spending on marketing.

It is also likely that they invest in adjacent offerings of the cloud (apart from just data warehousing) for which it will need significant upfront capital. But if there’s one thing Snowflake does not have to worry about now, is access to capital. With credible VC backed investors and a resounding vote of confidence from the equity market (as evinced by the IPO’s success), Snowflake can have all the money it needs.

What they do have to worry about, however, is meeting the expectations of its investors and clients. There is a lot at stake for Snowflake and Frank Slootman.

Excited to see how it pans out for them.

All eyes on Snowflake! :)

Perpetually beta. More at www.ashwinchhabria.com