Snowflake S-1 Analysis
Data warehousing company Snowflake filed its S-1 last week on 24th August, 2020.
Snowflake has till date raised $1.4bn in total funding according to Crunchbase and is backed by investors including Dragoneer Investment Group and Sequoia Capital. Goldman Sachs is leading the IPO and the San-Mateo based company plans to trade under the ticker “SNOW”. Snowflake offers a Continuous Intelligence Platform giving insights across a wide range of use cases. The architect of the platform is cloud-native, multi-tenant and build on modern, micro services based applications.
The USP of Snowflake among other data warehouse solutions available is the instantaneous, real-time, always-on experiences. The insight lies in the very fact that the build of Snowflake’s architecture is derived from Snowflake Schema (a popular Schema type used for data warehousing) which is further refined and normalized form of Star Schema. This is different from Hadoop which started trending in early 2010’s, and the vision was to be a bespoke data warehousing platform for Big Data. Hadoop is great for storing and processing large data sets but becomes complex and slow with small data sets that sometimes can be quickly scaled and replicated across nodes for blazing results. Snowflake has restored the academic frameworks and taken the industry back to first principles — its also complete ACID compliant! Old (Although a bit refined) is gold :)
The founders had an intimate familiarity and hence leveraged (As Tom Gonser calls it) the “unfair advantage” to engineer a fast-growing company aiming to break down data silos and derive value from rapidly growing data sets in secure, governed, and compliant ways. Snowflake has a global footprint with customers spread across sizes and industries (Advertising, Media, Financial Services, Technology, Online services, Healthcare, Manufacturing and Retail). They had 3,100+ customers as of July 2020, a 100% increase and and did $133mn in quarter ending July 2020, up at a staggering 121% YoY. Snowflake was founded in July 2012, first offered the platform for sale in 2014, has 2,037 full time employees across 19 countries.
Platform and Use cases
Snowflake supports a growing variety of use cases that can be cross utilized for garnering deeper insights, faster data transformations and improved data sharing.
Business Model and Key Metrics
Snowflake is driven by a consumption-based business model and not a vanilla subscription based model. The contracted revenue can have a overage or can be rolled to future periods. “Customers have the flexibility to consume more than their contracted capacity during the contract term and may have the ability to roll over unused capacity to future periods, generally on the purchase of additional capacity at renewal.” Consumption-based pricing is becoming a key to customer satisfaction, especially where the contract sizes are large and value to be derived from the platform becomes critical. While this method has been revered in literature, it is quickly getting pulsed within multiple software companies.
But for Snowflake, deferred revenue is still not significant. Most of the customers that demand for a contract change end up exceeding the contracted limits and opt-in for purchasing additional capacity and for customers not utilizing 100% of the contracted capacity, have an option to roll over the usage to future periods often at an option of purchasing additional capacity.
- Snowflake recorded a revenue of $264.7mn and grew 174% in fiscal 2020 and 133% for the first half of fiscal 2021(ending July 2020). Product Revenue as a % of total revenue has been decreasing from 99% at the end of fiscal year 2019 to 94% as of 6 months ended for fiscal 2021. The other stream of revenue majorly comprises of deployment fees/ Revenue on-demand/ overage fee. Snowflake’s Implied ARR (most recent quarter’s product revenue*4) stands at $499mn. Though the losses have reduced a bit they are still draining a lot of money and had a (72)% operating margin for the first half of fiscal 2021
- Net Revenue Retention for Snowflake was 158% for 6 months ended July 2020 which significantly beats the normal between the SaaS companies that went public in 2019 (~126%). This figure was 169% for first half of fiscal 2020
- Their implied ACV (Annual contract value) is up 7% comparing with the same period last year and stands at $160K (Implied ARR/ total number of customers)
- Revenue Concentration: umber of large customers with trailing 12-month Product Revenue greater than $1mn stands at 56 as of July 2020, more than 2.5x increase YoY! Snowflake also counts 146 of Fortune 500 companies as its customers and this cluster contributes to 26% of their revenues. One point of concern on the revenue concentration is that 11% of Snowflakes revenue is being contributed by a single customer group (Capital One!) and they claim to bring this figure down to under 10% in the coming quarters
- A 100% increase in number of average daily queries processed (>500mn in July 2020 from 254mn in July 19)
Snowflake’s gross margin has improved from 50% to 62% for first half of fiscal 2020 but is way below industry normal of 70–75% for fast growing software companies. But appreciative fact is that the margins have reached this level by developing and selling differentiated solutions despite the platform being built atop of public cloud vendors.GM gives confidence for further correction as the company grows because it was a result of “higher volume-based discounts for purchases of third-party cloud infrastructure, and increased scale across our cloud infrastructure regions”
- Sales & Marketing as a % of Revenues stood at 79% (down from 132%) for the first half of fiscal 2021. The implied rationalization was largely because of reduced spend on travel, hiring and related heads due to COVID-19. But it still is quite high considering the stage Snowflake is at (As a reference, Public market comparables are are at 33%)
- Below are the key financial and operational metrics and ratios:
Go-to-market
GTM for Snowflake is largely driven by 1) Direct sales team for acquiring new customers and 2) initial self-service trial through web traffic redirection. The breadth of Snowflake’s platform helps them engage at every level of the organization including data analysts and data engineers through self-service model and C-suite executives through direct sales team on large cloud transformations. To me if the product is capable to being sold in a down-up model where data team and developers can influence the organization to make a buy, this lever can definitely play a key role in enhancing execution, reducing sale and implementation cycles and ultimately driving up efficiency.
Market and Competition
According to IDC, the market fabric for Analytics Data Management and Integration Platforms and Business Intelligence and Analytics Tools has a combined value of $56bn as of 2020 and will grow to $84bn by 2023.
Snowflake claims to have covered all of these offerings and additionally call out that they offer data sharing capabilities that haven’t yet been quantified in the above numbers and pegs the total addressable market opportunity to be at $81bn as of January 2020.
Apart form private players providing similar services to Snowflake’s, the major competition is faced from the giant cloud infrastructure companies — Amazon, Microsoft and Google. Snowflake has reportedly taken over 2,500 customers from Amazon’s Redshift warehousing services including Adobe, Instacart, Deliveroo and Strava (As per one of the public comments made by Frank Slootman, Snowflake’s CEO).
Though Snowflake is competing with the biggies, they themselves have benefited from Snowflake’s customers as the offering is indeed built on top of these very vendors (Eyebrows up -Virtuous??).
Network Effects
Snowflake believes that the observed growth in business is accelerated due to the network effects as the data cloud keeps getting rich as customers move their siloed data to clouds. And because of the open data marketplace “the more customers adopt the platform, the more they can share data with, or receive data from, other Snowflake customers, partners, and data providers, enhancing the value of the insights delivered by the platform for all users”.
Other Industry Approved SaaS Metrics (Rule of 40, Sales Efficiency and Dollar Burn to IPO)
Rule of 40: In order to ensure that high growth SaaS companies burn within guardrails, an adaptation of Rule of 40 states that the Revenue growth rate added to profit margin should be greater that 40. Snowflake fares awesomely well at this metric as the latest quarter’s product revenue YoY growth rate is 117% with a (58)% profit margin which gives a score of 59 to Snowflake.
Sales Efficiency: This is to measure how effectively sales and marketing dollars are converted to revenue and a corresponding measure is payback period to determine how much time is taken to recover the cost for sales (Calculated as inverse of CAC Ratio, where CAC Ratio = New ARR added*gross margin/ Sales and Marketing spend of the previous quarter). And Magic number is sans margin and is used to see how fast is the cost recovery in terms of revenue generation (simply seen as New ARR added/ Sales and Marketing spend of the previous quarter). In the below analysis the New ARR added has been calculated as Implied Net New ARR (Difference of ARR between two consecutive quarters)
Snowflake’s implied payback period is 21 months and lies at par with SaaS companies’ average that went public in 2019.
On Magic Number, Snowflake is just nearing breakeven of 1x which is decent but not super for a company of this growth and margin profile.
Dollar Burn to IPO/ ARR: Snowflake has ~$887mn cash and cash equivalents on balance sheet and has till date raised close to $1.4bn. This concludes that the company has burnt $513mn to reach the current scale of $500mn implied ARR, ~1x ratio which is a touching a required target in the enterprise software space (While many companies struggle to get to 1x, the best of companies are 10x+).
Illustrative Valuation
Since Snowflake is still loss making, it’d be valued on a multiple of forward revenue, benchmarking with rest of the high growth SaaS companies. The guidance we have from Snowflake’s was fundraise in Feb-20, that the company was valued at $12.4bn. This gives Snowflake a 47x EV/ LTM Revenue multiple (as of Jan 2020) and a 20x-22x EV/NTM multiple by assuming Snowflake will register a 120%-140% YoY growth rate for fiscal 2021 (Its YoY growth rate for fiscal 2020 was 174% and S1 doesn’t give out any guidance on forecasts for the upcoming quarters). This is way higher than the publicly traded high growth SaaS companies’ average (~14x including outliers like Zoom, Crowdstrike, Datadog) but slightly lower than EV/NTM Revenue multiples regressed against growth (30x+ for growth profile like Snowflake). [This data is as of Aug-20]
On a deeper thought, the implied 20x-22x EV/ NTM Revenue multiple for last round could be lower given at the time of fundraise, COVID talks didn’t hit the markets as of January 2020.
Considering these facts and back of the envelope analysis, the below table outlines the scenarios for illustrative valuation ranges for Snowflake.
Snowflake has cracked the data warehousing market by making data silos operate in tandem for generating valuable insights, in real-time and faster than most of the competition and is one of the fastest growing SaaS companies at the IPO filing. They have best-in-class retention metrics with net revenue retention at 158%! which essentially means even if they stop selling to new customers, they can still keep growing at 58%. Though the company is doing moderate on implied payback period, that is more to do with relatively lower gross margins rather than only sales efficiency (It fares decent on Magic number). Given their growth, public markets may offset the market scenario and value it at a further premium than calculated estimates.
