저자: Kevin · 외국인
요약 보기
2020년 상장 예정인 Snowflake IPO를 분석. 매출 급성장(분기 2,900만→1억3,300만달러 수준)과 순고객유지율 158%를 강조하나, 마케팅비가 매출보다 큼. 단순 모델로 2년간 150% 성장·P/E 100 가정 시 주당 약 61~169달러 산출. IPO 직후 변동성·락업 해제(물량 증가) 리스크를 언급하며, 공모가 대비 하락 시 매수 고려 가능성을 밝힘.
다듬은 글
Hello everyone, this is Kevin. On Tuesday, September 15, 2020, a company with a lot of hype is going public, and, my friends, it will certainly attract a lot of media attention. The company is called Snowflake. It is a cloud-based software platform based on Google, Microsoft, and Amazon services that allows companies to conduct near- real-time analytics to better understand their customers.
I know it sounds complicated, which is why I'm making a video about it, because at first I barely understood what they were doing. But don't worry, I've spent the last six or seven hours trying to figure out Snowflake as much as I could. And I'll tell you, I've only studied the subject superficially.
So these are more like my first impressions of this company. So take it as it is. Also, take the free $20 that M1 Finance will give you for signing up and starting investing with them. You deposit $100 and get $20 as a gift. And you can even withdraw your money afterwards if you want. Okay, let's get to what I found out.
So, first of all , Snowflake's IPO, or the day it hits the stock market, is Tuesday. This means you could potentially buy Snowflake shares on Tuesday. The opening price is expected to be in the range of $75 to $85. Although often these IPOs intentionally try to set a slightly lower price, and then the shares immediately jump to $100 a share.
So I wouldn't be surprised if they say, "Oh yeah, we'll list them at $75-$85," and they actually sell in the $90-$100 range. We'll see. But it is expected that there will be a lot of hype around it. Mainly because Warren Buffett's Berkshire Hathaway invested $250 million in the company, along with Salesforce's Marc Benioff, who also invested $250 million—and one of those facts is actually quite surprising.
You see, Warren Buffett is an investor who made it a rule not to invest in unprofitable companies. And friends, Snowflake is a loss-making company. And this really, really interested me. Why the hell would Warren Buffett invest in a company that loses money and breaks his own rule? Well, partly because they are growing at a breakneck pace.
In fact, here are some of their earnings reports showing revenue growth in recent quarters alone. Friends, these are neighborhoods. These are not years. So, look down here. In the third quarter of 2019, their revenue was $29 million. Now, in the second quarter of 2021, and they have a kind of quarterly calendar.
Okay, that's about July 31, 2020. But anyway, they have about $133 million, which is crazy growth. And they're not just growing wildly, they also boast a great thing—positive net customer retention. This essentially means that you are not losing customers. If the rate were below 100%, it would mean you are losing customers.
Let's say you had 10 customers and a retention rate of 90%, that means you have nine customers left. Since they have a positive rate, it means they are making more money on each customer, having a net retention of over 100%. And it is 158%. And they have 56 clients that generate over $1 million in revenue per year.
And they have about 317 clients in total , a number they expect to grow significantly. And they are spending huge amounts of money to ensure this growth, which, believe me, we will talk about later. So, first of all, what the hell does this company even do? I made a little introduction at the beginning, but let's be honest.
What does this company do? I spent a lot of time trying to figure out what exactly she does. And the best I can explain it is this. You know what a spreadsheet looks like, right? Like an Excel spreadsheet , something like Google Sheets, or even, you know, here's a spreadsheet that looks something like this.
Here, this is a spreadsheet. And I can enter data here, like 100, 105, 200, whatever. And I can do interesting little things, like, "Hey, iPad, why don't you tell me the average of all this?" Oh, 135. Cool. It was great and fast. I like it. Well, imagine if you could perform such calculations as quickly, but not just for three values, but for a much larger number.
In fact, have you ever seen large spreadsheets, perhaps at school or for your business, where you have a mailing list with, say, a thousand or ten thousand email addresses? Hell, I've worked with spreadsheets like that, and let me tell you, computers start to slow down . You start to see the "loading wheel " quite quickly, even with just 10,000 rows in the table.
Now, instead of thinking about tens of thousands, as I just showed, let's not even think about hundreds of thousands, millions, or hundreds of millions of data points in a table . Instead, imagine a table that contains a billion data points. A billion data points. This is an insane amount of data. Now imagine that you can calculate the average for a billion data points and the program will give you the answer pretty quickly.
This is the goal of Snowflake. And it's not just about calculating averages. This is necessary so that companies can access their own data and then seamlessly conduct tests, comparisons, and establish connections between different data sets. This greatly simplifies business decisions and conclusions without damaging the underlying data.
One of their main advantages is centralized data that everyone can work with and draw conclusions from without changing the original “table ,” so to speak. So hundreds of remote workers at, say, Nike can simultaneously run various Snowflake analyses on the underlying data collected by the company, draw their own conclusions, and advise management, marketing departments, or other departments.
The best example I could find was the one Snowflake used in their marketing. And, to be honest , I don't think that was the best example they could have chosen, but that's what they did. They said, "Okay, here's a sample of data from City Bike." You know, those bikes you can rent in New York or somewhere else.
They discovered that they had a ton of data about each trip: length, duration, time of day, date, everything. And suddenly, around February 13, 14, and 15 of a certain year, they noticed that the number of trips dropped sharply, and then immediately increased again. So, in a few clicks, they were able to pull together a bunch of different spreadsheets that they had -- I call them spreadsheets, they called them other fancy words, but I'm simplifying -- and they were able to say, "Oh, okay, it looks like our average passenger is between 28 and 35 years old, maybe someone who works at one of the companies in New York." So, you know, this kind of millennial employee, and , oh, interesting.
We had snow on the 13th, 14th, and 15th . OK. So maybe that's why passenger traffic has decreased. And now I think to myself: well, that's enough for you. I could just pick up the New York Times that day and tell you the same thing. I didn't need all of this to come to that conclusion. But on the other hand, I am a one- person business.
This is a completely different matter than, say, Apple. You see, Fortune 500 companies, not individuals like me, like this . Forget about bicycles for a moment . Forget about me for a moment. Think about Apple. What data does Apple have? Well, they track the whereabouts of almost everything you own.
They know how often you use their devices. What's up, here's an iPad. Here's the Mac Pro right here. There's an iMac over there. They have, they have many access points to me. Good. They even know my pulse. Thank you, Mr. Apple Watch. I think I have power saving mode turned on . And now they even have your credit card details.
Well, of course, if you signed up for Apple Card. Not that I 'm an Apple fan or anything. I don't know what could have made you think that. But they're even in our cars as CarPlay. Apple is literally everywhere. And each of these interactions creates more and more data. And if we can give companies like Apple all this data, then perhaps we should think about privacy first.
But Apple, as a business user, and I, putting on my investor hat rather than my privacy advocate hat, say, "Hmm , Apple could do something powerful with this data." Imagine being able to correlate a bunch of things, like Apple Mobility. For example: "Um, it's February 28th, and suddenly fewer people are leaving the house." Um, they spend more time online and don't go to the office like they usually do.
In fact, they don't even log into their office computers. They stay home. Ahem. Oh, people are spending less on credit cards, saving more, and paying off debt. February 28. Oh, that was right before the pandemic. No guarantees. I just made this example up, but I imagine this is exactly the kind of data that analysts collect.
I'm just a YouTube guy, but there's a reason Apple has openings specifically for Snowflake engineers. Actually, here she is. And I must say, I was impressed myself. This vacancy was published on July 10. So, they're hiring, well, either they haven't closed the vacancy yet, or they've been looking for a long time.
But according to Glassdoor, Apple and other companies pay such employees somewhere between $130,000 and $170,000 just to work with Snowflake's infrastructure, presumably to draw some conclusions from the information received . The conclusion for me is this: wow, I can't believe that big companies actually hire people to work with this particular platform.
But believe me , it happens. And when you search for Snowflake vacancies, you see more and more of them. Okay, I'm impressed. What about the company's financial performance? This is where it gets interesting. And as an investor, this is the most important thing for me. You see, this company spends a lot of money on marketing.
And nothing shows a higher return on marketing costs than going public. Last fiscal year, the company spent more on marketing than it earned from sales. The same thing happened a year earlier. That is, the company is literally, if you recall the "input- output" scheme from the third grade. Imagine a teacher drawing on the board: "Hey, everyone, here's a company that puts a dollar in this box and gets 50 cents out." Oh, I don't like the "divide by two" box.
I like the "multiply by two or ten" box. Well, that's their hope. They hope that the dollar they invest now, getting only 50 cents, will turn into $10 in the future. It's like a machine with a deferred result. After all, remember that increase in income retention they show? And 26% of Snowflake's revenue comes from Fortune 500 companies.
When Fortune 500 companies start loving Snowflake, they can continue to hire people and use the product more and more. And the more people use the product, the more money they make. It's all very simple. So, in terms of cash and debt, this company also has enough cash and accounts receivable to actually pay off all of its debt.
So debt is not such a big problem. However , Snowflake has competitors. Amazon has a service called Redshift, and Google has BigQuery. And it seems that all three services are strong in performing their individual tasks. So it wouldn't be surprising if companies like Apple or other large corporations would start using all three services, building on the strengths of each .
This makes sense, because despite the existing competition, especially from giants like Amazon and Google , Snowflake's growth has been impressive. Some say it's partly because they believe Snowflake does everything better, faster, and easier than other companies. Maybe that's true, but as a guy who records videos for YouTube, I have no idea.
All I can tell you is that it is very difficult to value a company that is making losses. So I'm going to try this incredibly difficult task of valuing a loss-making company, which is just crazy. But we'll try and draw some quick conclusions. Okay, I've moved some of their data for fiscal year 2019, which is actually 2018, and fiscal year 2020, which is actually 2019 plus January.
It's some ridiculous story they made up over the years. This is ridiculous. For example, according to their logic, we are now in 2021, maybe they are just trying to get out of the pandemic faster. Okay, I guess I can't blame them for that. But, be that as it may , if you look at this growth, it is impressive.
That is, they grew from 96 million in revenue to 264 million. This is crazy growth. So that's two and a half times, right? This is simply incredible. Well done, they are young. Maybe they will even be able to maintain this momentum going forward. We'll see. And that's always a big question for growing companies, isn't it?
You know, we have a cost of revenue here that is actually becoming less costly. They spent only about 44% of their revenue on cost of revenue. They also have marketing, research and development costs, and general and administrative expenses, and that's the norm. The good news is that their R&D costs are falling and general administrative expenses are currently holding steady, but it would n't surprise anyone if both of these figures decline once their revenue starts to grow exponentially.
That's usually the case . What is worrying is that they are forced to constantly increase marketing costs. Now they spend, well, at least last year they spent 110% of their revenue on marketing. That is, their income was $264 million, and they spent $293 million . So far, this doesn't sound like a good business model.
The year before, they earned 96 million, but spent 125 million on marketing alone, and that's despite having other expenses. So the company is losing money. OK. But then again, that's normal for these growing companies: they come up with an idea and say, "Let's spend every spare penny on marketing, and you know what, let's borrow some money too." Although they don't actually have that much debt.
So that's pretty interesting. They get money, obviously, from investors who invest in the company. But in any case, their net losses also increased sharply. But let's move on to looking at their most recent data. Here's their latest data, and I want to show it separately so you can start to spot some trends.
Look at this. Their latest data for the period ending July 31, 2020 shows that the cost of their products has decreased. Look at this. Previously it was 53%, then 44%. Now, in half a year, the figure has dropped to 38%. This is a really good reduction. That's good. We want this indicator to fall. We want to see income grow and we want other expenses to decrease .
The fact that this indicator is falling is good. Marketing as a percentage of revenue. Look at this. It's less than 100%, and it's still growing, so remember, and it can be a little confusing. So I want to explain. This represents 6 months . And this represents 12 months. So, in 6 months they earned almost as much money as in the entire previous year.
So all the numbers you see here are for the last six months. Of course? So they spend less, but actually earn more. I mean, look at this. They spent 190 million on marketing, which is almost 100 million less than the previous year . However, they spent 190 in six months. Remember, it's 6 months, but they've almost matched the income.
This is incredible growth . And this is where we want to see the beginning of a decrease in this ratio. And look at general and administrative expenses as a percentage of revenue—they've dropped significantly. R&D costs have decreased significantly. Good. This is exactly what we want. We want these percentages to drop closer to 10-15%, that would be good.
OK. Uh, then we have a little bit of tax expense, net performance, etc. Okay, cool. So, we move on to the final result. Now, they will still have some tax benefits because of the losses they had over a long period of time, at least I assume they have loss carryforwards. So I'll keep it simple and not focus on taxes, interest, or other expenses right now.
All I want to do is see how much the company has to grow in the next two years for this to make sense. That's what I'm doing here. I took the latest numbers as of July 31st and multiplied them by 150% growth rate over 2 years. This literally means that if you have $10 , you convert it to 25, which is 2.5, and 10*2.5 is 25, and again by 2.5 is 62.50.
So, we should see $10 turn into $62.50. This corresponds to 150% of the growth rate accumulated over 2 years. Yes, this is a huge growth. This is more than doubling. That's a 2.5x increase, right? Uh, and you can see it here. To quickly check with me, you can take that 241 and multiply it by 2.5 , and then by 2.5 again, and you get $1.5 billion in revenue.
If they can get $1.5 billion in revenue, and we can reduce the cost by 30%. Since we are seeing a downward trend, gross profit will be around a billion dollars. Good. Okay, that's simple. I understand. OK. But here's the main nuance. I cut their marketing costs by more than half. And there's a reason why I had to do it .
You'll see why I had to do this, because it sounds crazy right now. Like, Kevin, why do you assume they should grow 2.5 times next year and the years after that? Well , you'll find out soon. OK. Uh , then I significantly reduced their marketing costs — up to 35%. And yet, somehow they still have to demonstrate such insane growth.
Good. Next, I took their research and development costs, as well as general and administrative expenses. And I set them at 15% of income. Okay, that's probably where they should be in the long run. Can drop to 10%. I set them at 15%. It doesn't matter. We'll see. We'll work with these numbers a little later.
You know, then I just left some general items here, like other expenses, interest, stuff like that, and taxes. Well, it's not really important. OK. So now we move on to net profit. This company would actually generate some net profit. and it would be its first net profit of about $85 million for the first time in a six-month period .
Remember, this is 6 months of data. Now we will soon calculate this on an annual basis. There are 278 million shares outstanding. And if I divide those two numbers and then double them to get the annualized figure, I get earnings per share of about 61 cents. So, with all the growth we demand, we will only get 61 cents per share.
And if this company with all this growth has a price-to- earnings ratio of 100, then the company is worth $61 per share. Yes. Wait, but I said they were going to IPO at $75 to $85. Yes, that's because, in my opinion, this company's projected share price has a crazy level of growth built into it. I'm certainly not an expert, but I'm trying to understand this stock and understand the potential valuation.
Let me do this. Let me take you to the stock page itself, or rather, not the stock page, but the table itself. Let's change some of the things we just worked on. Let's say we leave the P/E ratio at 100, but try to get this company to show us real growth in its share price. For example, what should we see?
Well, let's say they spend less money on research and development and on general and administrative expenses. Let's reduce each of these indicators to 10%. Well, if I reduce each of them to 10% of revenue, then suddenly this stock or this company's value per share is at 100 times earnings, which is a very high level.
You know, Salesforce has that number around 95. That's why I use that, because this company is probably going to grow over the next decade before we start valuing it like Microsoft or Apple, somewhere around 30 to 40 times. It is then quite acceptable that the stock price could be around $169 per share with 278 million shares outstanding.
So, if you buy Snowflake or are thinking about buying Snowflake, know that you are investing in huge growth potential and huge expectations for growth. These are two different things, right? Because if the potential is not realized, the share price could plummet. Now here's what else you need to know about it and when it might be worth buying these stocks.
Now, when these shares come to market, during the IPO, there will only be about 28-32 million shares available, which sounds like, " Seriously, Kevin? 28-32 million shares?" Sounds like a lot of action. Yes. Yes, that's a lot of stock. And that is very, very true . However, between about November 26th and December 6th, they're going to issue potentially up to 11.2 million more shares, which means you literally go from...Oh, I think I broke my board, but we'll use it anyway.
You literally go from 32 million shares outstanding to adding about 11 more, like we said, two? So that would be about 43.2 million shares potentially. I mean, if all these people who have these stocks sell them. OK. So, there will be some dilution of shares in November-December. But wait 15 days after that, if the stock price is trading 33% higher than at the time of exit.
In other words, if the IPO price is between $75 and $85, then if it trades somewhere between $100 and $113 for at least 10 of the 15 days after November 26th and December 6th. Yes, you can rewind 10 seconds to listen to it again if you want. Sorry. If that happens, okay, that's the main point. It is possible that the people who control another 38 million shares will be able to sell them.
This does not mean that they will necessarily do it. But what if all these people decide, "Dude, I just got a big raise and I'm going to Tahiti." I don't care about the pandemic. I am retiring. Wherever you want to retire, okay, it doesn't matter. Good. And then they sell. Well, let's see how many shares we have now.
We go from 32 million on IPO day to a potential 11.2, uh, well , let's do the math . 11.2 + 32 + 38 equals 81.2 million shares. 1.2 million shares, probably sometime by mid-December. So, as a Christmas present, and then sometime in January, if there are any other restrictions on other people's shares, any other types of lock-ups, will they expire two days or the second day after their second public earnings report?
Yes, it's complicated, I know, but the point is, there might be some logic to it, and I'm not urging you to do it because, look, stock IPOs can be crazy. This stock can literally enter the market at $75 and end up at $100 the same day. It could end up at $150 that same day, because that's how the market works.
Sometimes it seems quite irrational. There's a lot of growth built into this, but people are so excited about big data right now that it could happen. And if that happens, great. You will make a big profit. Unless you're someone who buys at $100 or $150 a share, uh, you're better off looking at the long term, is what I would say about those prices.
And that doesn't mean it can't grow, but let's assume they continue to beat earnings forecasts over the next two quarters. We'll see a 2.5x revenue growth, which would be cool, but I guess we'll have to wait and see. And they will cut their marketing budget . Okay, great. Let's assume that all of this happens.
I still can't shake the idea that there might be a better opportunity to buy this company between November 26th and maybe sometime closer to Christmas, December 25th, somewhere in that range, because there were complex time frames there. In my opinion, this is the point where the company won't have all this "oh my god, Warren Buffett invested in this IPO" hype.
This hype , in my opinion, is potentially very, very risky. Will I be looking at this stock on the day of its IPO? Probably. Do I think any of us should buy it? I think you should make your own decision whether to buy it. Personally, I'll probably abstain. And that could be a big mistake, because I really think there's a lot of long-term real value in this company.
I think big data is a big game. Eh, but that kind of volatility is too much for me. I don't like stress. So I'm not saying " no" because I think this deal is bad or that it needs crazy growth. I'm just saying it would be too stressful for me. That being said , I think I'll be able to drink eggnog and buy these stocks when the snowflakes start to appear in the air, because right now there's only ash in the air , because I live in California, and it sucks.
A quick note: if these stocks suddenly drop, let's say the IPO is at $75, and then the price drops significantly, I might be more willing to buy them. I think they are very overrated. Eh, but I like the long-term growth prospects. But, uh, then again, I might prefer to wait until the stock dilution passes.
And, uh , let me see if they can consistently lower marketing costs while growing. This is a trend we want to see continue. It's a little early for me. This may mean I'm missing out on an opportunity, but that's okay. I won't have FOMO over this. Thank you. Well, that's all , friends. If you enjoyed this breakdown, please share it with those interested in Snowflake .
And remember, this is just a general overview. I only touched on Snowflake superficially. I don't know everything there is to know about Snowflake. I still have a lot to do in analyzing this company, but honestly, with these lock-in periods, I have time. I'm going to Let's be. Oh, wait. This is water. OK. See you next time. Please subscribe and , uh, bye.
Comments 0