Snowflake: Some Challenges And More Solutions

Summary:

• Snowflake Inc.’s recent guidance for 22% revenue growth has disappointed investors and caused a decline in share price.
• The company’s new CEO, Sridar Ramaswamy, is seen as the right leader to address challenges and drive growth.
• Snowflake’s AI initiative, Cortex, and other new products have the potential to significantly impact revenue and profit projections.
• Snowflake has developed competitive solutions to address competition with Databricks and other alternatives in the market.
• Snowflake’s current guidance is likely to be exceeded, providing a much more reasonable valuation framework than might be in the minds of many.

Snowflake: Does a new leader and some new technologies signify a return to hyper growth?

It has been 8 months or so since I last wrote about Snowflake Inc. (NYSE:SNOW) for Seeking Alpha. I rated the shares a strong buy then – it wasn’t the greatest call. The shares have essentially been flat since then, while iShares Expanded Tech-Software Sector ETF (IGV), probably the most representative software ETF, has appreciated a bit less than 20% over the same time span. It hasn’t been the most pleasant journey with an expected growth inflection disappearing over the course of a couple of quarters embodying disappointing forecasts for revenue growth.

It can be easy to give up on the shares of this erstwhile hyper-growth company and conclude that it has lost its competitive mojo. And there is no doubt that Snowflake has challenges. It also has solutions for its challenges and increasing opportunities.

Notionally, the shares aren’t cheap, and that is even after considering the company’s rapidly growing free cash flow margin. And the company’s percentage growth has been declining for the last two years without much relief for weary investors, including this writer. I have lots of patience with investments…some might say too much, but at this point, I think it does make sense to do a deep dive into what actually has been happening

The most recent Snowflake earnings were released on February 28th. The earnings release itself wasn’t a tragedy. The forecast for the new fiscal year certainly was. Snowflake was built to be and has been valued as a hypergrowth company. The forecast of 22% revenue growth for the coming year is nothing like what most investors – and again that includes this writer – were expecting. In addition, the replacement of Frank Slootman as company leader by Sridar Ramaswamy was unsettling to some.

I must confess that I was a bit surprised to see Mr. Slootman, after enjoying huge successes at his last two companies, Data Domain and ServiceNow, go out on a bit of a sour note, but as I explore the balance of this article, it was probably the appropriate time for a CEO transition. And I believe that Ramaswamy is the right leader at the right time for Snowflake.

Snowflake has presented an unusual forecast in that all but one of the major growth drivers of the next several years were not included in the projection. In addition, the company has forecast impacts from such initiatives as tiered storage pricing, the introduction of what is called Iceberg tables for storage, and a compensation plan focused on consumption expansion. These changes have costs that can be calculated mechanically and benefit from price elasticity that can be difficult to determine in advance.

Overall, in an effort to achieve consistent beats beyond those forecast throughout the year, the company simply didn’t forecast any revenue driver for which there is no historical trend – anything different would be more of a guess than a forecast, and the CFO has been very clear about his strategy for guidance. The company has a significant (private) competitor, Databricks, whose management has been vocal about its competitive advantages vis-à-vis Snowflake. And there are some who essentially buy that contention. Again, as I will explore later in this article, the new Snowflake CEO has acknowledged the issues, developed and implemented appropriate strategies, and has fired more than a few shots across the bow of Databricks and will continue to do so.

Snowflake these days is a fairly large ship. And like other large ships turning it is a process of some duration. So why buy the shares now? Last week the CEO made an open market purchase of about 31k Snowflake shares valued at $5 million. Usually, when I see insider sales or purchases, I don’t consider them as great guides regarding the current outlook of a company. That said, I used to think an insider purchase was more significant – that was until the example of Dustin Moskovitz and Asana. But I am inclined to think this purchase was of a different nature.

For one thing, Ramaswamy is not a founder of Snowflake. For another matter, despite the payday he must have had when his company Neeva was acquired by Snowflake, his net worth is far, far less than that of most founders of larger IT companies. A $5 million commitment to the shares might reasonably be assumed to indicate something of substance – some sign that current guidance can readily be exceeded. My belief is that this insider purchase is likely to mean something substantive about the company’s business outlook. Of course, I wouldn’t predicate a newly updated recommendation simply based on an insider purchase.

The company has appeared at a number of investor conferences subsequent to its earnings release. Such conferences usually are fairly anodyne and generic affairs. In my view, these recent investor presentations suggested that a forecast of 22% growth is not really a forecast that growth for Snowflake this year is most likely to be 22%. How much greater than 22%? Enough to matter for the share valuation, in my opinion. The company CFO, in my understanding of his explanations of his forecast, set the table for a series of beat and raise quarters. I will let the CFO’s commentary speak for itself later in this article.

Snowflake shares have depreciated by almost 35% since the earnings release on 2/29. The valuation contraction was reasonable just by looking at headline guidance. It will not prove to be reasonable if results for the current quarter that will end on 4/30/24 exceed estimates materially, and if guidance is raised. Of course, I don’t know if that will happen, but as I will explain, the setup for what’s to come seems quite favorable to me.

I have been a Snowflake shareholder for some time, although not in a huge position. I have avoided making an outsize commitment to the shares because of their record of volatility. If, indeed, Snowflake does beat its most current projections (product revenue growth of 26% and non-GAAP margins of 3%), and forecasts a somewhat greater percentage revenue growth than is the current consensus (22%) for the full year, the likely share price appreciation has the potential to be significant.

In other words, I think buying Snowflake shares now, before the company demonstrates a growth reacceleration, is a reasonable strategy given the way the current forecast was engineered.

I recommend the purchase of Snowflake shares at this time and at this price with the understanding that timing a growth inflection is a fraught undertaking and with some expectation that buying the shares in advance of a return to accelerating growth visibility could be catching a falling knife.

I am willing to endure some knife cuts if it means I have caught a multi-bagger at a reasonable entry point.

Snowflake shares have always been a “risk-on” investment. They still are. Much of their short-term valuation will be a function of how markets trade. As I write this, hi-growth IT shares have been consolidating for the last few weeks as traders speculate about the timing and cadence of Fed interest rate actions. This is not an article about interest rates, but a review of Snowflake’s prospects over the next year and beyond. Some interest rate projections have been seemingly upended by the release of the ISM manufacturing survey covering February that was released on April 1st.

Interest rate angst is a major factor in short-term moves in high growth IT stocks and interest rates have backed up noticeably in the past few weeks. So, I caution readers that Snowflake shares will be subject to lots of volatility and will be pressured in periods in which interest rate expectation reflect concerns about “sticky” inflation.

Reviewing Snowflake’s earnings release and its guidance

The latest quarterly earnings report from Snowflake did not provide investors with lots of positive data points. And the guidance was just plain awful in terms of its divergence from prior expectations. Just to review: product revenues for the quarter that was reported (ended 1/31/24) were $760 million, up by 33% from the year-earlier quarter and a beat of 2%. Non-GAAP EPS for the quarter was $0.18, compared to prior expectations of $0.09. Free cash flow was $324 million last quarter, up by 3X from the prior quarter-much of that is the seasonal end-of -year influence of renewals. That said, the 12 month free cash flow margin was 29%, compared to a 25% margin the prior year. None of that was disappointing.

What was disappointing was the revenue growth guidance. Product revenues for Q1 are forecast to be about $750 million, or growth of 27%. Prior estimates for product revenue had been for product revenues of around $765 million. For the full year, product revenue growth has been projected to be 22%; the prior consensus had been for revenues to grow by more than 30%.

Why was the guide so disappointing? Part of the explanations has been a change in guidance methodology. In the short term, Snowflake’s revenue growth is all about consumption. While Q4 was indeed a revenue beat, consumption trends did not return to their pre-2024 patterns, and the company uses the latest trends in consumption as the basis for its forecast. There are some other factors in the guide as well. So, despite commentary from the CFO regarding stronger consumption trends the past couple of months, the forecast is based on disappointing consumptions trends noted through most of FY ’24.

The company introduced tiered pricing recently; its guidance reflects a revenue loss from tiered pricing that was generated more or less mechanically. In addition, it is expected that some customers will leverage what are called Iceberg Tables for their storage. The advent of tiered pricing and Iceberg tables are really not negatives for the company’s revenue growth. They are actually expected to increase revenue growth because they improve the relative price/performance of Snowflake-they are anticipated to tap demand elasticity. But the forecast doesn’t reflect that.

The company’s AI initiative, Cortex, is just now in public preview. It is expected to be generally available in June. The company is projecting a significant level of incremental costs related to Cortex. It is not forecasting any incremental revenues. The company had earlier announced a new commission structure that incents sales people to focus on generating new Snowflake workloads. Workload acquisition and deployment are basically how consumption grows. The company is forecasting greater commission expense but no benefit at all from the new go-to-market strategy.

The following is a quote from Mike Scarpelli, the CFO of Snowflake who has a long history with both Data Domain, and ServiceNow as a colleague of Frank Slootman:

Mike Scarpelli

Well, I’ll start with Q4. In Q4, we had a good bookings quarter. But to be honest, we still fell short of expectations on the revenue side from what we thought internally. And going into this year with the CEO transition, last year was a rough year, we were hoping, from a consumption standpoint, we’d see customers return to consumption patterns pre-2024 — as a reminder, we’re in fiscal 2025 right now — and we didn’t see that.

Yes, consumption trends are good right now. But with the new CEO coming onboard, we have so many new products coming to GA that we just don’t — we’re not going to forecast those until we start seeing a history of consumption. So, I do think there’s, call it, conservatism, I call it prudent guidance, and we’ll take it one quarter at a time. And we would really like to see Sridhar succeed as our new CEO.

In the short term, consumption trends will drive the company’s revenue growth. Consumption, as I have written before is not something that can easily be forecasted. It is not something that can be managed or controlled by a vendor except through the tiered pricing and through sales compensation plans. But the impact of those initiative probably won’t drive quarterly consumption trends in the next couple of quarters.

Longer term, of course, bookings are a key factor that correlates with future revenue growth, although with a company such as Snowflake, the impact is lagged. The irony of the growth angst is that bookings have been at record levels:

Actually, the spending environment is pretty good and you see that based upon the commitment customers are making with their bookings,

The spending environment is actually pretty good with our customers. I just think a lot of the customers we have now that are ramping on Snowflake are a lot more disciplined, more mature companies than a lot of the digitally-native companies where they had these massive valuations and money being thrown at them that they just spent euphorically.

The customers we’re bringing on today and ramping up, big telcos and banks and stuff, have always been very cost-conscious, and they’re going to do things at their pace.

So, in the near-term, there’s a more prudent approach to ensuring that you’re going to be in a position to beat and raise on a go-forward basis.

The management transition: Does the new CEO have the experience and drive to take Snowflake to the next level

Frank Slootman, the retiring CEO of Snowflake has been a legend in his own time. He has a reputation for being one of the savviest and hardest-driving operators/sales executives in the enterprise software business. His success with both Data Domain and then with ServiceNow, Inc. (NOW) is well known. One of the factors that led me to make a commitment to the shares initially had been Slootman’s track record. Now he is retiring.

The other day, the CEO of Databricks asserted that Slootman was retiring because he missed the bus with regards to AI. It is a provocative thesis but has less than nothing in terms of facts to recommend it. I confess that the new CEO, Sridar Ramaswamy, is far less known – both to me and to most investors. And yet he, too, has a track record of success. After listening to his presentation at the JP Morgan conference and looking at his resume, my conclusion is that he is the right man at the right time to lead Snowflake.

Sridar had a lengthy career at Alphabet Inc. (GOOG), (GOOGL), initially as an engineer, and ultimately as a leader of the company’s advertising business. He was also responsible for launching Google Pay. Sridar eventually co-founded Neeva, a search engine utilizing AI technology which was bought by Snowflake. Sridar has been involved in using AI technology for almost 20 years. At Google, he helped to build prediction engines for that company’s ad click-through product.

Frank Slootman has been an operator, and a legendary one, but he has never purported to be a technologist or a thought leader in emerging technologies. Sridar is self-described as a technologist with significant experience in leading large cross-functional organizations. Obviously the list of prospective candidates to lead Snowflake was large. And equally, the list of prospective employers looking for an executive with domain expertise in AI was equally long; Sridar had many different choices-he ultimately chose Snowflake.

Transitions happen in life. Some are better and more productive than others. It is my belief that Frank Slootman, recognizing what Snowflake needed to revamp the company’s growth engine, timed his retirement as the company’s AI initiative, Cortex is nearing general availability. Slootman is 65, he is several times a billionaire and he has been in high stress roles for well more than 20 years. Occam’s razor essentially suggests that the simplest explanation is most often correct. Retirement made sense now for Slootman and his departure shouldn’t be read as having some dark, existential meaning.

Of course I have had no personal contact with Sridar any more than I have had personal contact with Frank Slootman. But his resume ought to suggest that his leadership of the company will resolve any product issues that may have been a factor in decelerating growth.

Snowflake’s Cortex: Just what impact will it have on Snowflake’s growth-and when?

There has been much written about which company is or is not a beneficiary of the AI revolution. Much of what has been written from a 300k ft. level. Almost all enterprise IT companies have announced some kind of AI capability. Few of them have been generating much revenue from their AI offerings – certainly at this point I haven’t read of any large IT vendor – with the possible exception of C3.ai, Inc. (AI) – with more than 5% of their revenues derived from generative AI. And even in cases such as C3.ai, Inc. and Palantir Technologies Inc. (PLTR), the proportion of revenues coming from generative AI as opposed to traditional AI is still small. I’m just echoing Gartner and other 3rd party commentators in suggesting that investors who are looking for an inflection measured in a month or two simply will not find it from generative AI offerings.

There has also been much written about AI being something of an investment bubble. I think that is a substantial misreading of what is happening. AI is really a revolution and will ultimately impact most phases of work and leisure activities. I am not a futurist or a technologist, per se, but this is something on the order of the changes wrought by telephones, electric light bulbs and automobiles. No doubt there is exuberance, but it is perfectly rational.

For example, I mention a company called Mistral below. Mistral started in April 2023. The company raised more than $400 million in October last year, and by the end of the year it was valued at more than $2 billion. It now has a partnership with Microsoft and its data models seemingly are considered to be the most advanced and functional. When investors see that kind of cadence, they look for ways to invest, and while the valuations may not seem rational when looking at traditional financial analysis, it is the very human desire to find a place on a rocket ship that is leading to what some describe as bubble valuations.

The fact is that AI is very real, it will bring change to most forms of work and leisure, and it will create a lot of wealth…along with, of course, some spectacular flameouts – probably gaudier than what is now being seen in the EV space. It is my belief that Snowflake is likely to be a significant beneficiary from AI over the coming years, and that its Cortex platform will have a major impact on the company’s revenue and profit projections.

Notionally, Snowflake’s technology ought to be a good fit for the creation of generative AI applications. Snowflake’s core technology is that of providing users with an extensive data base that is cloud resident and offers strong performance. That is part of what is required to create generative AI applications. Cortex is designed to take the creation of generative AI apps to another level. It is almost impossible at this stage in the evolution of AI deployment to rank different solutions. When I wrote about GitLab Inc. (GTLB) the other day, I was able to cite some early adopters, and a third party analysis to bolster my conviction about that company’s solution.

In this case, Cortex has not yet made it to general availability-it is just now in public preview. I have linked here to a product description. As mentioned earlier, the fact is that the company’s CEO has domain expertise with regards to AI and generative AI – this is what he specifically knows about, and that is why the fit between him and Snowflake makes lots of sense.

A few weeks ago, the company announced a partnership with Mistral AI. Mistral is considered one of the leaders in the creation of models. The company was founded barely a year ago and its models are said to have the best performance currently available. I have linked here to a description of Mistral. I think adding Mistral to Cortex at such an early stage is a function of a new sense of urgency regarding AI at Snowflake and bears the imprimatur of the new CEO. Cortex can also utilize Large Language Models (LLM) from Meta Platforms, Inc. (META), Google, and through the Python serverless function.

Cortex is said to be easy to use – one of its attributes is its appeal to so-called “citizen developers.” And it is advertised as cost effective. At this point, there isn’t an SKU for Cortex; that will come, presumably, when the product is fully released in 2 months. Initially, much of the revenue impact from Cortex is likely to be based on drag. In other words, some users will choose to buy the full Snowflake stack because Cortex is a component of a holistic solution. At the very least, Cortex is designed to create apps and these apps in turn are likely to drive workloads and Snowflake consumption. How much that will be initially is pretty much a guess at this point.

A few easy to understand functions of Cortex are its ability to analyze text data and to generate predictions and insights on structured data with machine learning capabilities. Cortex and associated products are most likely to initially have an incremental impact on Snowflake’s revenue growth. Most revenue impacts will start to be visible in the 2nd half of the fiscal year, and will have a greater impact on FY’26. But I do believe that the first half of the current fiscal year will be the nadir for percentage revenue growth.

• I am going to quote below one comment made by the CEO at the MS conference. Two things of note: Frank Slootman would never provide this kind of answer and two, the CEO has deep product knowledge that is perhaps unmatched in terms of the understanding of what a broad range of IT users want to do with AI:

This very morning, a product manager that I was talking to said, Sridhar, I wanted to just take a look at the new use case data that we have sitting in Snowflake. This is our reps entering use cases for all of their customers, what drives consumption. He said, I was looking to see if by running sentiment detection on the text that they enter for their comments, whether I could predict the win rate or not.

It turns out when people have positive comments about use cases, our win rate is like close to 99%. On the other hand, if it comes out negative, then the win rate is far, far slower.

Now, this person knows nothing about language models, nothing about deployment, how they should use it. It was just like a query that they ran. 15 seconds later, they had that insight. So, that is power.

In addition, Cortex also has a semantic index. You can think of this as a vector index, so startups are all the rage these days, built natively into it. And our aspiration here is that you should be able to build a chatbot on any corpus [ph], a bunch of PDF documents, maybe your help desk, maybe some customer support cases. You should be able to do that in five minutes and stick it into a Streamlit app. (Streamlit is a development tool that is designed to take machine learning models and turn them into interactive apps.)

Taking a look at the Snowflake forecast: What is missing is perhaps more important than the specific data on which it was built

When Snowflake reported its quarter and provided guidance, I think the basic reaction of most shareholders – and that includes this writer – was one of nausea and disbelief. How could this happen. In my case, it was worse, as my thesis when I last wrote about the company 9 months ago, was that there was a revenue growth inflection coming – and 22% revenue growth is a growth inflection, but in the wrong direction.

Basically, the issue is that the CFO, Mike Scarpelli has made a conscious decision not to include some significant revenue growth drivers in this year’s forecast. Why? Part of the reason, actually articulated by the CFO, is to provide a low bar for the new CEO. This isn’t some conspiracy guess; Scarpelli actually maintained that part of the reason for the guide was to ensure that the new could be seen as successful.

But beyond that, Snowflake will have many new products, even beyond Cortex, that will be in general availability. But because they are new products, the staff at Snowflake is not really able to make reliable forecasts. The one newish product that is being forecast is Snowpark, which is forecast to be 3% of revenues this current year.

And the company has made a decision to use the latest consumption data as the basis for its revenue forecast. Consumption fell short of expectations last quarter. Consumption trends are significantly better now – they are at levels that the company had hoped to see the prior quarter. But those trends aren’t in the forecast, and they won’t be in the forecast until after the company actually has a quarter with decent trends in consumption for a full fiscal period.

What is not embedded in the forecast is Streamlit, the impact of tiered storage pricing, Cortex, the co-pilot offering, the availability of Iceberg Tables for Snowflake, and Document AI. In addition, the CEO talked about having a notebook product in private preview. It is supposed to be ramped with “much ferocity and velocity.”

What might all that mean for the revenue growth forecast for the current fiscal year and beyond? The way the forecast has been constructed shows growth percentages declining from about 27% in Q1 to less than 20% by the end of the year. Such a forecast is counterintuitive, to say the least. I expect that revenue growth in Q1-2 to be the slowest of the year, with a noticeable increase by Q4 as some of the new products start to generate specific workloads and to buoy consumption patterns. I expect that, by the end of the year, Snowflake percentage revenue growth can exceed 30%. For modelling purposes, I used a revenue estimate of $3.6 billion, or revenue growth of 28%.

The current forecast for free cash flow calls for a 29% margin, comparable to the free cash flow margin of the prior fiscal year. Most likely, if revenue is greater than forecast, much of the increment will be captured as free cash flow. And some of the new products are highly likely to drive large multi-year commitments that can have a substantial impact on bookings, and on deferred revenues. These estimates don’t make Snowflake shares cheap – they do change the valuation calculus notably

Competition: Is Snowflake’s growth being stifled by competitors

One of the key investment debates with regards to Snowflake has been that of whether its declining percentage revenue growth has been a function of competition. Much of the attention has centered on competition from a private company, Databricks. Databricks is a formidable competitor and it has enjoyed stunning revenue growth. Last year, it reported that revenue growth was greater than 50% with revenues reaching $1.6 billion.

As Databricks is a private company, no other financial details have been made public; it is likely that it is still loss-making, and burning cash. The company recently raised $500 million in a venture round with a valuation of no less than $43 billion. That is an EV/S ratio of …27X!, exceptionally elevated these days. If Databricks grows another 35%-40% this year, than the EV/S ratio would be 19.5X-still elevated. The company has no present plans to go public-probably because it remains loss making and wants the flexibility to invest in products and sales capacity.

Just looking at reported revenue growth, the conclusion has to be that Databricks is taking share in the market. Snowflake “only” grew revenue by 36% for the full year. In the prior year, Snowflake’s percentage growth actually had been greater than that of Databricks. But before leaving it at that, market share and revenue are two very different things. At the end of the day, Snowflake did have a record bookings quarter, and regardless of the guidance as it relates to reported revenues in the current fiscal year, Snowflake’s CFO has described the current sales environment as favorable.

Q4 was an exceptional booking quarter for us. Bookings are not leading indicator of revenue; they do signal an improving macroenvironment. Remaining performance obligations grew 41% year-over-year to $5.2 billion. Of the $5.2 billion in RPO, we expect approximately 50% to be recognized as revenue in the next 12 months. We signed our largest deal ever in Q4, a five-year, $250 million contract with an existing customer. Our international territories returned to meaningful growth, outperforming expectations for the first time in a year.

I am not going to quibble about bookings not being a leading indicator of revenue. Bookings do indicate the intention of users to choose workloads that run on Snowflake infrastructure. It has been typical for users to take 9-12 months before they start to deploy workloads they have contracted for. So in that sense, bookings are not exactly correlated with revenues-but they certainly mean something with regards to future reported revenue growth and market share.

Snowflake’s growth problem is a consumption problem. It is not a bookings problem because bookings are at record levels. Those bookings, in time will turn into consumption, and then into revenue. The real question is whether Databricks has stifled Snowflake growth by being selected for workloads. That might be the case in certain situations – certainly in those enterprises who use what are called notebooks. The Databricks notebook product is resonating with users, and sometime choosing Databricks for its notebook software does lead to selecting the company for its whole suite of cloud data warehouse solution. But that said, competition between the two vendors goes back and forth with some wins and some losses.

The reality is that this is a large space, and the success of Databricks is not necessarily at the expense of Snowflake. Databricks has been very successful in its sales motion with data science groups within an enterprise. Snowflake’s success has been at the large enterprise level. Snowflake has been focusing some of its product strategy at filling holes that may have given competitors, including Databricks opportunities. Snowpark is one such offering and has seen some success. Streamlit, which was introduced just a couple of months ago, is another such offering which focuses on improving building and sharing data apps.

I have linked here to a comparison between Databricks and Snowflake. One thing that is striking, at least to me, is that all of the components in which the analysis found advantages for Databricks have been recently addressed by Snowflake.

With all of the attention being paid to the rivalry between Databricks and Snowflake, as the linked analysis shows, the leading competitors in this space are the hyper-scalers and in that group, in addition to Amazon.com, Inc. (AMZN), Google and Microsoft Corporation (MSFT) are Oracle Corporation (ORCL) and International Business Machines Corporation (IBM). These are the competitors that compete most often in the space simply because they are incumbents as cloud vendors. They do have their own cloud data platform alternatives, but these tend to be both expensive and hard to customize.

Snowflake has been slow in releasing some add-on products, and it probably has been slow to bring Cortex to the market. But at the least, it has remediated these perceived deficiencies. It is not surprising to see the CEO of Databricks make rather ambitious claims about the ability of his company. He can do so because Databricks isn’t public. But I would take much of what he has been saying with several grains of salt.

Snowflake has some competitive challenges. And it is resolving them quite rapidly. It has a CEO who probably is more aware of what market requirements are for Snowflake to excel competitively compared to most other senior executives in this space. That is his challenge and he seems up to the task.

Snowflake’s business model

Snowflake’s business model has been improving steadily and at a significant cadence. Most investors in Snowflake are looking for growth, but in looking at valuation, I incorporate a free cashflow margin along with expected percentage revenue growth. Last quarter non-GAAP operating margins were significantly above expectations at 9% compared to the 4% the company had forecast. Part of that was the beat on revenue of about 2.5%. Indeed, the company’s free cash flow margin was 42%. To an extent, the strong bookings discussed earlier led to increased collections and that in turn caused free cash flow margins to over-attain noticeably.

But the company has also continued to constrain expense growth and that has lead to consistent improvements in non-GAAP expense ratio. Non-GAAP gross margins have reached 75% up by 400 bps year on year, and flat sequentially. Non-GAAP sales and marketing expense was 36% last quarter compared to 38% in the prior year quarter and compared to 37% in the prior sequential quarter. Non-GAAP research and development was 23% last quarter, up from 19% in the year earlier period, and .up from 21% in the prior sequential quarter. Finally, the non-GAAP G&A expense ratio was 7% last quarter, down from 8% in the year earlier period and at the same level as the prior quarter.

Overall, the company’s non-GAAP operating income margin last quarter was 9%, up from 6% in the year earlier period, but down 100 bps sequentially. That brought full year, non-GAAP operating margin to 8%, up from 5% in the year earlier period. The company has forecast that its operating margin this year will also be 8%. Given the way the revenue forecast was constructed, and some of the assumptions that I pointed out that were more than a bit conservative, I think it is likely that the company will continue to grow its operating margins, both this current year and on into the future.

As mentioned, the company’s free cash flow margin was 42% last quarter. Free cash flow rose by 50% year-on-year. Part of the increase in bookings shows up in the strong increase in the deferred revenue increase which rose 27% year-on-year. The full year free cash flow margin was 29%, and the projection for the current year also call for a 29% free cash flow margin.

Not terribly surprisingly, Snowflake does use stock-based compensation aka SBC. Last quarter, SBC was 39% of revenues, a ratio that some will feel presents a valuation issue for the shares. That said, SBC did decrease from 42% of revenues in the year earlier quarter. Given the constrained level of hiring, which usually is a significant factor in SBC levels, I expect the ratio will be declining throughout the current year.

As I typically mention, I look at dilution as the real cost of SBC, particularly given how that metric is calculated. Dilution has been running at around 2.8% on an annualized basis, somewhat below average for most other high growth IT companies. That level of dilution brings my estimated outstanding share count for this year to 342,000,000 and in turn, that yields an estimate of a bit more than 13X for my projected EV/S ratio. Snowflake shares aren’t notionally cheap even with a projected free cash flow margin of 30%. The question, for investors is basically will the company’s more aggressive product and go-to-market strategies yield growth significantly above the consensus/company forecast. I believe that will indeed happen.

Wrapping up: Reviewing the case to buy Snowflake shares

Snowflake’s guidance laid an extra size egg-perhaps the size of an ostrich egg. The company provided a forecast for revenue growth this year of 22%-implying that by the end of the year, its growth would slip below 20%. In addition, the company has announced a CEO transition, replacing a legendary leader with a less well-known executive, generally regarded as a technologist. The net result has been to bring Snowflake shares down by almost 35% in the few weeks since it released its earnings report.

Snowflake as a company certainly faces challenges. It is dealing with an aggressive competitor, Databricks. Its AI offering, Cortex, will only reach general availability in June. It has made some changes in its comp plan, and has introduced tiered storage pricing, and will be introducing Iceberg Tables for Cloud Storage.

But my view is that these are problems that are already being successfully addressed. The company, for example, has targeted the successful offering of Databricks notebooks (that is a software product that is a tool for creating data science and machine learning workflow-not the kind of notebook that many readers have and use) with its own notebook that will be released in general availability before the end of the year.

Snowpark, the company’s solution that is focused on helping programmers create applications using Python to develop machine learning models, is already having some impact on Databricks as the link provided above suggests, and was called out by the CFO in his presentation at the Morgan Stanley conference. The comp plan change is designed to solve one business problem for Snowflake by incentivizing sales people to focus on workloads and not just bookings. I believe that price elasticity is alive and well and that tiered pricing and the introduction of Iceberg Tables will be a net tailwind for revenues.

The company’s guidance has explicitly not included estimates for Cortex, for Streamlit, for Document AI, and several other significant new product releases. In addition, the company’s forecasts are based on usage trends seen in what was a disappointing year, rather than on the most recent usage trends. And the company’s guidance doesn’t relate in any way to the exceptionally strong bookings quarter that included the largest deal in Snowflake’s history.

Further, my view is that it was an appropriate time for Frank Slootman to retire after a storied career. Hey, even Tom Brady is no longer a quarterback in the NFL. The new CEO, Sridar Ramaswamy, has a pretty decent resume of his own with lots of success at Google, founding an AI-based search company, Neeva, that has been acquired by Snowflake. He actually has been leading Snowflake’s AI efforts and has been instrumental in getting Cortex into the market and in striking a partnership with Mistral, these days considered to be one of the world’s top model makers.

I don’t want to suggest that Snowflake growth is going to leap from the levels forecast to some level greater than 30% in a single quarter. Things simply don’t happen at that cadence. But I do think Snowflake’s forecast for both revenue growth and non-GAAP operating margins has been set up for consistent beats, and raises throughout the year.

As I have commented in a recent article on GitLab, I think the time to buy Snowflake shares is not when its growth is obvious to everyone, but when growth estimates are controversial and when investors are unconvinced that the new CEO can restore the company to rapid growth and market share gains. If all that were obvious, then inevitably the share price would be noticeably greater than where it is today.

Overall, I believe that Snowflake will leverage its technology successfully and that growth for the next few years will significantly exceed published consensus levels. I own Snowflake shares now, and plan to add to my position. I believe that Snowflake will create substantial positive alpha over the coming year.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of SNOW either through stock ownership, options, or other derivatives. 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.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---