The upcoming Snowflake (SNOW) IPO is set to be one of the hottest tech IPOs to come out of the gates this year, with Warren Buffet’s Berkshire Hathaway as well as Salesforce Ventures essentially being cornerstone investors in the deal; they will subscribe via concurrent private placements at the IPO price (indicated range of US$75-85). But does a Buffet deal make it a good deal? Should we ride on the coat-tails of the legendary investor?
The focus of this article will be on the Snowflake’s financials and my calculation of a fair valuation range, rather than the pros and cons of Snowflake as a company, as the latter has been well covered by other Seeking Alpha contributors. I should also qualify that I am also not a tech expert, and so I probably do not fully appreciate the technical advantages of the Snowflake platform compared to its competitors. Nonetheless, quoting Mr. Buffett: “You don’t have to know a man’s exact weight to know that he’s fat
Source: Snowflake S-1 Filing
Snowflake was founded relatively recently in 2012, and made its first product sale in 2014. As such, it is a young company that enjoys a correspondingly high revenue growth rate. Based on the quarter ending July-2020, Snowflake’s y-o-y revenue growth rate was a whopping 120.7%. That is the highest growth rate amongst the data warehousing/analytics comparables in my list. On a trailing 12-month basis, Snowflake’s top line growth looks even more impressive at 238% given the exponential ramp-up in sales over the last 4 quarters vs. the year ago; this is again much higher than the peer average of 40%, shown in the table below.
Comparable Company Analysis
Source: Snowflake S-1 Filing, company filings. Snowflake T12M EBITDA Margins are estimated by the author.
Meanwhile, Snowflake has been showing an improving margin profile as economies of scale come through, with operating margins having declined steadily from -136% in the Oct-18 quarter to -58% in the July-20 quarter.
I would take the improvement in operating margins over the last two quarters with a pinch of salt though, as there have been some one-off declines in costs due to COVID-19. Snowflake’s S-1 filing states the following:
After the outbreak of COVID-19, we have seen slower growth in certain operating expenses due to reduced business travel, deferred hiring for some positions, and the virtualization or cancellation of customer and employee events. We intend to continue to make significant investments in research and development as we add features and enhance our platform. We also intend to invest in our sales and marketing organization to drive future revenue growth.
Nonetheless, the fact remains that Snowflake’s margins have been improving consistently every quarter, displaying considerable operating leverage.
In terms of the balance sheet, Snowflake has zero Debt, and has Cash, Cash Equivalents, and Short-Term Investments of c.US$591m as of 31 July 2020. That is a comfortable level of liquidity, given that operating cash outflows amounted to US$45.2m in the half-year ending 31 July 2020.
From a financials point of view therefore, Snowflake is checking all the boxes.
Strong revenue growth? Check.
Improving margins? Check.
Healthy balance sheet? Check.
Snowflake is offering 28m shares of its Class A common stock, at a price to be determined on 15th September but within the range of US$75-85 per share. Each of Salesforce Ventures and Berkshire Hathaway will also purchase 3.125m share in a concurrent private placement at the IPO price. Post the IPO and private placements, there will be roughly 279m shares outstanding.
This means: i) the free float of the company will be roughly 10%; and ii) based on the mid-point of the pricing range (i.e US$80 per share), the market cap of Snowflake would be around US$22.3bn.
Extract from the S-1 relating to the offering
Source: Snowflake S-1 Filing
Given the early stage of growth that Snowflake is in, and the astronomical sales growth that follows, I prefer not to engage in a peer multiples comparison at this point. For example, the P/S multiple will be distorted by the high sales growth numbers that should drop rapidly over the next few years on a higher base effect. Based on my estimates (i’ll explain further below how I come to these estimates), Snowflake’s annual revenue growth will decline from 174% as of FY20 (ending Jan-2020) to 47% in FY25. That’s a big yet expected decline given that not many more mature companies can sustain revenue growth rates >100%. Thinking about using P/S multiples to value a company like Snowflake, the question is then – which growth rate do you use? There is no right answer to this.
Author’s estimates of Snowflake’s revenues and revenue growth
Instead, I prefer to do a Discounted Cash Flow model for Snowflake as this better reflects the growth in the near-to-medium term. I explicitly model out the next 5 years on a quarterly basis, and the subsequent 10 years as an interim period on an annual basis.
My Discounted Cashflow Model is based on a few key assumptions:
Assumption #1: total customers grow from 3,117 as of July 2020 to 9,217 as of July 2025. Snowflake had been adding about 400 customers per quarter pre-COVID and I extrapolated this at a declining rate going forward. I also think this fits well with the statistics that in the US (where the bulk of Snowflake’s sales are booked), there are about 26,500 firms with annual sales over US$100m which is a good starting point for thinking about what the addressable market is for Snowflake in the US. 9k out of 26.5k firms using Snowflake sounds like a bit of a reach, but I assume there is some global expansion built into that as well.
Assumption #2: net revenue retention rates are 200%/175%/150%/125%/115% for customers who have used Snowflake for 1/2/3/4/5+ years respectively, and a 1st year customer spends US$21,750 on average. How I calculated this was an iterative process using the number of customers and net new customers quarterly, and triangulating that back to the total revenue number for the July-2020 quarter. Given that Snowflake’s reported net revenue retention rate was 158% as of July-2020 and the average customer should have been using Snowflake for ~2 years give or take, these assumptions seem to tie in.
Assumption #3: I assume steady-state operating margins of 10%, with Snowflake breaking even in 4QFY2022 (ending Jan 2022). This is probably the most challenging assumption to make, and the least accurate. Why I say that is because while recent margin trends seem to indicate substantial operating leverage, when I look at competitors that have been around for 5-10 years longer than Snowflake, many of them have not managed to turn profitable (see chart below). So I’m taking a wild guess here, with 10% as a middle ground of sorts between the two pieces of conflicting information. It’s important to note that:
Source: Company Filings, author’s calculations
Assumption #4: I am using a weighted average cost of capital of 10%. I find this is a good rule of thumb to use as a discount rate for riskier companies from my experience. Oil & Gas companies use a 10% discount rate on their future production and that level of risk seems somewhat commensurate with the level of risk we are looking at with regard to the limited visibility on Snowflake’s margin outlook. In any case, better to be conservative than not.
With that in mind, here is the Discounted Cash Flow model:
Source: author’s calculations
As you can see, my discounted cash flow model yields a target price of only US$36 per share at a market cap of about US$10m for Snowflake. Why the big gap vs. the asking price, I ask myself.
I don’t think my sales numbers are conservative; I’m expecting sales to grow 40x in the next 10 years. The key is therefore margins. To arrive at a US$80 valuation, I would need to assume operating margins increase to about 23-24% for Snowflake over the next 5 years and stay at that level in perpetuity. While this is certainly possible, given the lessons learnt from competitors who have been around for longer, my view at this point given the limited information and guidance available from Snowflake is that I wouldn’t bank on the company hitting the 23-24% level, which a US$80 IPO price seems to imply.
What would I pay?
While the model I shared above says US$36, that fails to account for much growth outside the US, and the possibility of margins being >10%. I think the Berkshire investment also lends Snowflake some credibility and it’s possible that Berkshire knows something that we don’t. I am also aware of the premium being afforded to secular growth companies in a returns-scarce world, and that isn’t going to change in the near-term. Getting to a target price is more of an art than a science, but based on the above, I would be happy paying US$50 per share for Snowflake.
Disclosure: I/we have no positions in any stocks mentioned, but may initiate a long position in SNOW over the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.