Snowflake Inc. (SNOW) Earnings Call Transcript & Summary

So look, we're just delighted to have Snowflake and Michael Scarpelli, the CFO joining us today, which -- how many years -- when was the IPO?

2020, September '20.

Yes. And you did it before the IPO, I remember because you were like, this is great. I'm private I can whatever I want.

I did say that.

And maybe this is the last time. So for anybody doesn't know.

Hopefully. Hopefully.

Michael announced his pending retirement and transition out, but you got to find someone new first, right? So we'll get to that. All right. So Snowflake is a $60 billion market cap. They had a fabulous quarter. They grew 28% on the top line nearly $1 billion in revenue just underneath it. 43% free cash flow margin in the quarter. We shouldn't expect that. Q4 is really good...

Q4 is always going to be the biggest quarter.

Q4 is really good. And it marked a year of the new CEO, and it's been pretty remarkable, actually, I would say, the amount of sort of innovation and change that happened in a year. So let's start with House business and then we'll talk about how Sridhar pulled this off, the business.

Sure. No, business is good. Listen, we grew 28% quarter -- year-over-year quarter. We grew 30% last year, which was very good. We improved our operating margins in Q4. We went through a -- Sridhar really drove this as well as a number of others with what we called our Project Get Fit, where we really rationalized a number of teams, cut-out management layers. We did a lot of performance management while still hiring people, and we're still hiring quite a few people this year. And we just guided for the quarter, and I feel good about the guide we have, it has an implied second half acceleration in our business. And you have to remember, Q1 is our toughest comp just given Q1 last year was a good quarter, but it also had the extra day. And I would say business is good. I would say it's definitely not euphoric, but it's a stable environment within our existing customers, which are the ones that drive revenue. They continue to consume, and we've identified a number of new workloads with many of our customers to go live on. There is a lot of uncertainty with new projects, so in terms of landing -- and what I say that I haven't seen it yet, but as everyone knows, there's a lot of uncertainty with what's happening with tariffs and reciprocal tariffs and whatnot, and you hear it on the news all the time. I haven't heard a single customer mentioned that yet and the ones that I've talked to, but clearly, that's got to be on the minds of many people.

Yes, you just can't get away from it.

Yes.

it's actually been nice the last 2 days not watching CNBC.

Yes.

It's nice to be in here with you guys. Tariffs don't affect you economically in any way, right?

No.

And just so -- because it's kind of interesting. So it does -- why is it not -- why do tariffs not impact software?

Well, first of all, it's -- that's a good question. I don't know but it could. But we do what we are seeing more so and there's a lot of countries that are being really aggressive with transfer pricing right now, which, in essence, is a form of tax we're paying like India is being very aggressive on transfer pricing right now. And that's probably the country where we pay the most cash taxes. And as we do more in India, and for instance, they're taking the stance that your transfer pricing should be cost plus 26%, and then you pay a high tax rate on that 26%. That is a form of taxation that is going against U.S. businesses. So we do pay -- our customers do pay VAT on the software that they buy. So that is -- so to be determined what -- if they do anything on software.

What is transfer pricing?

So when you are incurring costs in a country, you have 2 models you can do. You can either do a -- and if you're a selling entity or a back office or R&D and a selling entity, you have 2 choices. You can do cost plus reimbursement for those people and that is transfer pricing. And you can do a distributor model, where you -- the distributor makes a profit, in essence they buy the software from you and resell it. Most companies do a cost-plus reimbursement on the sales side. Some countries, you have to do a distributor model because it's more tax advantage to you. And on the cost plus where you have an R&D function or a back office G&A, internal IT, you typically -- in those countries, you have to do a markup on that and pay -- you have to guarantee a profit and pay tax in those countries on that.

All right. That's something I got to look into. Let's talk a little bit about why you're leaving us?

I am old. I've been doing this for a long time. This is my fifth public company. When I was at ServiceNow, I had indicated I wanted to leave and was going to retire. And to be honest, if Frank never approached me back in April of 2019 to join him, and remember, he asked me to come and have a coffee with him and he said, "Hey, if you'll do this, I'll come out of retirement and do it with you." And so the more I learned...

Where did you guys had the coffee?

We had the coffee at Ruby Hill Golf Club, where he was living at the time. And I just happened to be golfing one Saturday morning there and said, we can meet there.

That's funny. Okay. And so...

And I had always told Frank when I joined Snowflake is, listen, I will guarantee you when you leave, I will stay minimum 1 year. And it's been a year now. But I am in no rush to leave.

Yes, yes and we appreciate that you are in no rush.

And I just thought it was going to be better that we do a public search for a new CFO. I think you get a lot better candidates when it's a public search where they know that I am really leaving. I've tried over the years to hire a #2 with the hope that it will -- like the promise to them that they will become the CFO. But to be honest, there's such a shortage of CFOs it seems out there. And it's really hard to hire someone and make them wait 1 to 2 years to be the CFO. So I'm ready to hand the reins over when we find the right person. I've said I'll do -- if that person wants me to do a transition for whatever period of time, I'll give them a title, I'll stick around. But if they want me to leave because I think it'll be disruptive for them and harder for them to do their job, I'll leave the day they start, but we have to find that person first.

Yes, all right. So you have someone doing it. You have some high end...

We're kicking off a search meeting with a recruitment firm, myself, Sridhar, and our Chairman of our Audit Committee, who is getting involved in the search.

And what are the characteristics you're looking for?

First and foremost, Sridhar would like someone who is very good with investors and credibility with investors. He'd like someone with a very strong FP&A background that is very operational, someone that can deal with customers. A number of the functions that I have will get moved and report directly into Sridhar. And when I leave -- like, for instance, our CIO reports in to me, including all of our security and everything. That will move. Our general counsel and the 100-something lawyers we have reporting to me that will move into Sridhar. But they have to be very -- they have to be someone that is respected by the field. That's a really important piece and is someone that can also work across the company with all the various leaders in the company.

That's a lot. All right.

We're going to take our time and find the right person

Yes, that's a lot.

You want to jump?

No. I don't know how you guys do it.

And you got to understand taxes.

Yes, yes, no, definitely not. But keep my eyes out for who the right person is. Yes, that's a lot. All right. Cool. So over the last year -- so you guys acquired Sridhar's company, right?

We did in May -- Neeva in May of '23.

Yes. And then -- so I guess he was just -- I forgot, he has a word for the period before he was CEO. I forget what he calls it?

So when we were acquiring Neeva, I was negotiating a deal with Sridhar. Frank really wasn't involved in negotiating that deal. And I remember when I talked to Sridhar literally before we even closed on the deal, I'm like Frank, Sridhar could be your successor. You need to spend some time...

Really?

Yes, no, it's true story. You need to spend some time with him and the more Frank spend time with him and the Board got to know him as well too. We thought he was the right person.

Yes. And so...

And I still believe he's the right person.

It seems like it. It seems like it. So what -- just stepping back for people over the last year, what are the 2 or 3 biggest things you'd point to that he has done?

I would say the #1 thing that the immediate result was much, much better accountability within the engineering organization and product organization for ownership of the features and products that we're building. Really drove a lot more alignment between engineering and product management and sales. I would say the speed at which we release thing, I think we've released over 400 features last year, which you don't hear about all of them, but Sridhar is a big believer in -- our old engineering leader and product management, would always want to work on all these new features, get them all bundled together and then take their time to release them and Sridhar is like, no, I want to release with -- at velocity when we have a new feature done, let's just roll it out and continue to do that, so...

It's real time, when it's ready you let it roll.

Obviously, we do a lot of testing on everything, too. And in many of the cases, my IT organization is customer #1. We're huge users of Snowflake and we test everything in production ourselves before it gets rolled out to customers many times.

All right. So a big one was product innovation a lot fast. That's the right word is, velocity of product innovation.

And then the next big thing that he spent a lot of time this is where investors were like, oh, he really does have the sales side and stuff. Sridhar is an operator. He is really good at process. He's really good at understanding, and he is a student, he learns very, very, very fast. And he really revamped our sales organization to hold them more accountable, the management side of actually managing reps and getting leading indicators for who's going to be successful and not just looking at how many meetings a week are people taking, how quickly are they following up on things. He's great at being able to pull a lot of data just out of your Google calendar to understand who are the reps who are doing one meeting...

He'll do it himself, right? He doesn't need someone to come do it for him.

I've learned how to use the Google Suite so much better through him, products. But then he also really has upped the level of our SE organization. We have a new leader running our SE organization, who really built it out at Azure. He's been with us for a while doing different roles, but he's really reinstituted that every SE needs to pass a test. And if you don't pass the test, you get 2 chances, you're going to find another job. And really holding the bar higher. He took -- he insist on taking 30 of our best SEs and went through an extensive boot camp on AI, so they could get in front of customers and really preach about what Snowflake is doing in AI and machine learning.

And these are the sales engineers? They support your sales?

Yes, these are sales -- Yes. But you really realize that the sales engineers are the really ones that drive a lot of new workload adoption within our customers that leads to revenue.

Yes. Are you increasingly looking to hire salespeople who actually have the skill set to be a sales engineer. I mean, we're like they can build the demo and they can do the sale?

You know what I'd say, I see a day where that is going to be the place where a rep and an SE is interchangeable. We're not there yet. But what I will say is our best reps tend to be people who are SEs at a stage in their career, and you have to understand technology.

So that's a little bit of a trend. I've heard from people who are interviewing at Salesforce that if you can't build the agents and everything yourself, they're not interested. It makes sense, right? Like the whole point is you're trying to say that these things are easy to adopt, you better be able to do it yourself. Okay. So the other thing that I've heard is, people are like -- what do we need for a new CFO, and they're like, "Well, it's got to be somebody works 20 hours a day, 7 days a week." Does he have a really high like work ethic?

Sridhar has no hobbies. He works all my e-mails -- Well, he does go to the gym every day for an hour.

And that's supposedly post his power lifting stats on the slide.

Yes, that I don't know about it. I don't see that, but he does go to the gym with the trainer every day in the office. But my e-mails with him usually start at 5 in the morning. We typically do a lot of early morning calls, but he also has no problem doing calls at 8, 9 o'clock at night. And I get e-mails from him and many times we talk on Saturday, Sunday mornings, and he works a lot. He's super high energy.

I heard -- I might have told you the story last year, but I had an executive of another company who was really excited that he got an introduction to Sridhar's for dinner. And then when the details came in, the dinner was in the office, bring your technical team. We'll order in.

I actually -- I get invited for lunches all the time with people. I'm like, I don't have time for lunch. I work. If you want to come meet me, I'm happy to grab a sandwich in our cafeteria and I can bring it up to my office and we can sit in there and talk, and he does that for dinners. We don't do Board dinners outside of the -- our Board dinners are in the office, in the boardroom.

Okay. Why don't we open it up to our audience for questions. I have a bunch of other stuff, but let's see what people want to ask you, Mike.

You are stuck asking me...

I'm counting in the room They'll be playing over. Yes, in the back, go ahead. And we have to remember to repeat the question.

What are the primary competitive pressures Snowflake can see from others in the market like Databricks and others?

Databricks, we coexist in many of our large accounts. They really came at things from a data science persona and a lot of Spark. We really came at it from a data warehousing and we've converged somewhat. But if you talk to our customers, they view us very differently in how they use us. I would say the biggest competitor to Snowflake, and we've said this since the IPO are really the hyperscalers. And quite frankly, if the hyperscalers could do what we do, we wouldn't exist today. And I would say Google is probably the most competitive from a technology standpoint in the core data warehousing with BigQuery. Microsoft is probably #2. And if they can get Fabric right, they're still not there, that they will be more of a competitive threat. With that said, our partnership with Microsoft is probably the best it's ever been and improving. And AWS is the least competitive. Yes, they have Redshift, but we still replace it left and right. And the one thing that AWS and Microsoft Azure figured out and they get it is, if Snowflake is running in AWS or Azure, it gives them an opportunity to sell other software services around Snowflake. Google still hasn't got that yet, but they don't have a lot of other software services to sell. Yes, they have Gemini now and how can we help them drive revenues associated with Gemini, and that's something we're working on. In terms of Palantir, we really don't -- Palantir is really not a -- I don't see them in deals at all. I'm not hearing about them, where they tend to exist more, where they're very big and I'd love to get some of their work is in the federal space. But so much of that stuff today, what they're doing is different is still so much customization. There's a lot of people involved in the work they do. So on the private side, it would definitely be Databricks. But listen, this is a massive market and growing and there's going to be many successful people in the market with this because of how big the market opportunity is. We're still growing very well. I think there's not many companies at scale growing at the pace we're doing. Why? Because we win more than our fair share of deals. And every deal is competitive. They always have been.

Yes. I do feel like the volume of sort of the competitive discussion between the two of you has diminished somewhat over the last year?

What I would say on that is that's more just because it's like Databricks has ratcheted down their tone on stuff and is not as public as they were. I don't think if you talk to our field, nothing has changed in that.

SAP -- sorry.

You can ask.

Now we'll get to it.

Speaking of federal, given the change in the last 6 weeks, what are you guys seeing so far...

So the question was, given the change in the federal administration with the government, how does that impact our outlook for the year? Well, the federal vertical is our smallest vertical. It's like less than 1% of our revenue. And so as I've always said, it's upside, and if anything, the -- with what's happening with DOGE and efficiency, I think we can add a lot there. With that said, we spent the last 2 years investing very heavily in our federal segment getting the certificate IL-5 FedRAMP high. There's some protect or BCAP and there's some other things as well as onboarding some new partners. There's some interesting RFPs we're competing on. But from a revenue perspective, it's really not until next year that some of these -- if they turn out will pay off in terms of revenue, but I would say there's zero downside and only upside.

Yes, in the back.

[indiscernible].

So I'm not going to comment on rumors in the news, but we partner with RedPanda, we partner with Confluent, and we have a team of engineers that are working on real-time streaming as well, too. It's a pretty small team. We will partner and we do think we'd like to own the whole data life cycle, and that is something that we are spending time on. So stay tuned.

Can you just explain why the streaming is so important?

First of all, I'm not a technical person but i will.

I know but keep it high level, just...

The reality is the more you can control the writing of that data and the movement of data, it's owning that data life cycle, getting it into Snowflake faster for our customers. What we always hear from customers as well too is people don't want to procure from various vendors. They just want to procure from one place. We hear that all the time. And so to the extent we can own it or partner and if we partner, we're looking at doing things like letting it go through our marketplace, so it can all be on our paper to -- we don't get the revenue. It's just the transaction fee we get, but making it easier for customers. And we do that right now with a number of our partners like RelationalAI and we're working even with Fivetran and Matillion, can they go through marketplace so our customers don't have to procure separately those services.

[ Three ] contracts instead of one of what they're paying...

Yes.

It's not really specific to Snowflake but I thought I'd ask you anyways. Is all this stuff about data center orders maybe being scaled back because of -- maybe any thoughts on that?

Yes. It doesn't really impact us and why I say that is we leverage all 3 clouds and pretty much every region around the world where they are today. And I don't have any access getting GPUs, and I don't have any access getting normal CPUs or storage. So I don't see that...

No challenges getting access..

It's really gotten better. Price still needs to come down.

I think that's going to happen. All right. 30 seconds. Anyone have one last one.

You were going to ask about SAP.

Oh, yes, you saw that. Well, the SAP Databricks, was that something that you guys were interested in getting?

We're working with Databricks and I mean in SAP and similar to do bidirectional connectivity with the movement of data. But the reality is, is we do -- it's the same thing we have today with ServiceNow and Salesforce, and the vast majority of customers just use that to move data into Snowflake. Very few customers, we see take their Snowflake data and push it back into those other platforms. Why? Because they weren't built be the data warehouse platform that we are.

Yes. All right. Mike, thanks so much. It's a pleasure to have you here. We always appreciate it.

Thank you.