Snowflake Inc. (SNOW) Earnings Call Transcript & Summary

My name is Tyler Radke. I co-head the software sector here at Citi. Thank you for joining us for Citi's 2021 Virtual Tech Conference. Hard to believe it's the end of day 3 here, but we're very happy to have Snowflake. The CFO, Mike Scarpelli, joining us for the closing keynote. Mike, thanks so much for being here. It's great to see you.

Thanks for having me.

So Mike, I thought since this is new and Snowflake is a relatively new IPO, coming up on the first anniversary here, maybe we could just start off with a brief overview of Snowflake. What are the differentiating factors that you think help make this one of the largest software IPOs and the fastest-growing companies in software history?

Yes. I think, well, it all starts with, a, the product; but b, I think we're in the right place at the right time. There really is a huge momentum behind companies adopting the public cloud, which our timing couldn't have been better for coming up with our product. There's all the technical reasons why we are superior to others. But the biggest thing what's driving our growth and why we're being so successful is data is exploding within all companies around the world. The traditional on-prem model, you really didn't have access or the access to that data was not possible to all people within the company because of the constraints around the physical hardware. The public cloud really enables unlimited scalability. And being in the public cloud and with our technology, you really eliminate those data silos that were occurring in an on-premise world. We make data available to all in a secure governed way. Our customers know exactly who is accessing the data, they control that, but they can really see, from a governance standpoint, who is doing what with the data as well, too. So it's really the right place at the right time with a great product is what's really been the key to our growth. And I would say all of the 3 big hyperscalers that we compete with: AWS, Microsoft and Google, their architecture was actually never built to be in the cloud initially, and they're still trying to catch up to where we are. And we think they're quite far behind, but that's also why they tend to partner with us as well, too. A lot of that partnering though tends to happen when they failed on their -- on those hyperscaler solutions, and then they partner with us. I will say, over the last 1 to 2 years, those partnerships have become much more forthcoming from -- especially AWS and Azure, I would say, Google probably the least amount.

Got you. Okay. So Snowflake has talked about this concept of the data cloud. I think Frank Slootman has written a book or soon to come out a book called The Rise of the Data Cloud. What is the data cloud? What does this mean for Snowflake? And how is it different from existing data markets that investors might know such as data warehousing or business intelligence?

Yes. So we view -- so we really started out as a data warehouse, but we've seen what companies are using us for. And we're really a platform, and our platform offers workload execution across the entire data life cycle. Well, the Data Cloud bridges these performance capabilities with data access and content. It's super important in that piece. And we don't even talk about when we go and to talk to customers anymore, we're not talking about the technology or what we do. We're really talking about our customers' business and data operations and that whole Data Cloud vision is what's resonating with our customers that enables us to engage with the customers at that level and getting into the line of business more so than just IT. And the nice thing about the Data Cloud, and we just had -- we had our big financial services Data Cloud Summit we've done our media and others. And the Data Cloud really offers access to industry-specific data content through data sharing, which enables differentiated use cases. That's really important. Data sharing is what the Data Cloud enables.

Right, right. Yes. And I think that the Financial Services Summit, that was just earlier this week, right?

Correct. Yesterday.

Yes. So maybe to touch on the industry piece for a little bit. Help us understand, is this a kind of -- are you making these industry solutions on a product basis? Like are you kind of architecting the Data Cloud differently for those solutions? Or is it more just around the data sharing aspect in the community of customers and users that can now kind of share data between it and you're that intermediary?

The latter, it's really one product. And the whole purpose of our industry verticals, talking about the financial services, we'll do health care and others. It's really more having specific use cases that drive business value for those industries that we are delivering. It's the same product. We're a very horizontal product across all industries. But as these use cases in the network effect of the data sharing, which brings together similar people in, whether it's financial services or health care or whatever. And that's what it's all about. It's not different products.

Right. Right. And to take that financial services example, I mean, is this kind of a greenfield opportunity? What were those customers doing before to share that data or maybe they couldn't show that data for a variety of reasons?

Well, we replaced a lot of what was being done before in FTP downloads, which is a very -- it's an old -- it's a 50-year-old technology today or something like that. And that -- or maybe it's not 50 years, but it's very old. But we enable data being shared without copying that data. There's actually no data transfer. And you can -- through the data sharing, you can also hide PPII information so that you can't see from privacy and some of the stuff, especially in Europe, what's going on, you can protect yourself from that. And every one of our customers has an aspiration to do data sharing when we talk to them, no matter what industry you're in. But the first thing in a customer's journey is to first get their data into the cloud. And so first, there's that on-premise migration of their data warehouse. Or if they were in a first-gen product getting that and then the whole Data Cloud opens up to be able to start doing data sharing and other things that happen. What we're seeing is there's a lot of fintech companies that are actually born in Snowflake or very early on, adopted Snowflake. And you can just think of who those companies are as well as new lines of business within financial services companies that are starting out on as they're trying to become more online banking and stuff for doing stuff from Day 1 in Snowflake. And we see that with Goldman Sachs, with Marcus, that was talked about in our IPO.

Right, right. Yes, I think there's some on the asset management side as well.

Well, you have like BlackRock with the [ lag ] at the [indiscernible] Summit Users Conference. And that really creates a network effect as well, too. The more people we can get developing applications on Snowflake, it opens up their customer base to us as well to become Snowflake customers. Because if you want to consume that data through this new offering, you have to be a Snowflake customer to consume that data. And a lot of the -- think of a lot of the financial services, whether you're on the investment side or whatever you're getting your S&P data, your FactSet data. Whatever data was coming through, these were coming through the FTP downloads. Now these things are all just done through data sharing. You have the capability to do it.

Yes. Yes. Interesting. So maybe help frame for investors the typical customer journey. I know that might be hard because customers are all shapes and sizes. But are you landing customers for a new use case? Do you start with replacing a Teradata or an Altezza or Hadoop? Just help us kind of understand that progression for acquiring and scaling a customer.

Yes. So it varies greatly. We historically started out early on. We were doing a lot of first-generation cloud product replacement, think Redshift and others. Then we started getting into the Teradata and the [indiscernible] and the big on-prem data warehouses. But there's also a lot of Hadoop, SQL, Oracle that we replaced as well, too. And when we get into these discussions with customers, they may be a big on-prem data warehouse in Teradata or whatever, but they generally have a lot of other stuff, too. And it really depends upon the company, customers. Some start with doing the Teradata. Others will start with the Hadoop, SQL and other stuff, and then it's a journey. It could take a number of years. I'll give you an example. We have a large retailer in Europe that we signed up 2 years ago, and they are progressing along very nicely. They're not going to shut Teradata down until the end of next fiscal year is when they will have that shut down. But that's a customer that has grown from -- we did an initial $5 million deal in calendar '19. In the last -- and last year, they did barely $1 million in revenue. This year, they will do well over $8 million, and I expect they could grow as high as $20 million next year. That's the type of journey, and I don't use the dollar amounts, but just the progression, how they grow. And that's pretty typical. And that's why you see our net revenue retention rate is so high.

Yes. Yes, absolutely. So I guess, is there a way to think about your mix of greenfield versus replacements just in terms of that land motion?

I don't have it from a dollar standpoint, but we do -- as we -- it depends upon the customer itself. Like we have over 300 companies that have really built their business. We're foundational to running them. I would say those are greenfield opportunities, and some of them are quite large customers today. But we do a lot of on-prem migrations. And then once they get up and running, and it could take 2 to 3 years fully on, it sounds like then there's more workloads that come that they start moving on to Snowflake. We continue to come up with new use cases like one of the things that we talked about, and I would say this is a greenfield opportunity is an AT&T spoke at our Summit Conference, and one of the big things around, what they're doing is a lot of their ServiceNow data putting into Snowflake for analytics on their CMVB. And I would say that was a greenfield opportunity.

Right. Right. So I wanted to shift to the recent quarter, which you reported, I think, about 3, 4 weeks ago, really strong results, triple-digit revenue growth and I think actually an uptick on that dollar net expansion rate. And the guidance would raise -- was a little bit higher than we'd seen in the prior quarters. So maybe just help us understand what is driving the business now? Do you feel like you're kind of hitting an inflection now that you're a couple of quarters out of the gate? And you're seeing all these market tailwinds around cloud adoption?

What I would say is Frank has been with the company a little over 2 years now. And one of the first things that he really wanted to do was the company was very successful in what I'll say, commercial, smaller enterprise. He really want to go after the biggest companies in the world. And it takes time to ramp those guys. You can see our progression in the Fortune 500. And I don't think that's an accurate -- you really should think more about just large enterprise companies, and it takes time to ramp those customers, and we're seeing those customers starting to ramp now. And so it was really broad-based consumption that really drove our outperformance in the quarter and supported our ability to raise our full year guide well ahead of our Q2 beat this quarter. And I just think it's -- you just need to be patient with how long -- people underestimate how long it takes and how difficult it is to do these migrations. And if it were simple, many people would be doing them. And we have seen with some of our competitors, many, many failed Teradata migrations. To our knowledge, we are the only one that has been successful in doing them to date.

Interesting. interesting. Yes. And I think people sometimes forget that just because you signed a Fortune 500 customer that quarter, you may not get that revenue for several quarters out until that consumption.

And what you also need to understand, too, is our average deal size really hasn't changed. It's just under 100,000 is with that initial Cap 1 on average. And it's even to get to the customer to consume that full amount, it usually takes on average, 9 months to start consuming at that contracted rate. And then it's the follow-on year that they start then really ramping their consumption on Snowflake and it just continues as they move more workloads. And a lot of people think you just do a Teradata migration, you move everything. That's not the case. It can be, but most people start moving workload by workload out of the on-prem system, and that's why it could be a 2- to 3- to 4-year journey to fully migrate everything. And you see that with the ramp in consumption.

Yes. Yes. So I guess as we think about that net expansion rate, that did tick up this quarter, north of 160%. How should we think about the sustainability of that? Because on one hand, it does sound like you have this huge runway of large migrations and Fortune 500 customers that have kind of just signed up and will continue to ramp those use cases going forward.

Yes. Well, clearly, as we get bigger in our installed base of customers grows, I think it will be hard to maintain at the level we're at, but we think net revenue retention will remain best in class for a very long time. And it's supported by both the mechanics of new customer wins as customers typically land with a small initial deal size and expand based on actual consumption. I don't see that changing, quite frankly, with customers. And the potential for enterprise accounts to expand consumption with additional workloads and use cases. Our reps are really focused on our enterprise accounts to helping customers find more use cases and workloads that can move into Snowflake. And that's part of the reason why, too, we talked about the verticalization. In order to do that, you have to have reps that understand the businesses of these different verticals and the business value we drive. It's not about just speeds and feeds in the technology. It's about what value are you going to drive for them.

Right. Right. And sticking on that topic of workloads and use cases. I think you referenced an uptick in data lake migrations in the quarter. Could you just talk about, I guess, number one, what do you think is driving that? Is it something going on with the incumbent technology? And then two, how is Snowflake kind of evolved your data lake use case kind of non, I guess, nondata warehousing use case that drove that uptick?

Yes. What I would say is data lake functionality is actually core to Snowflake. This is something that's in our roots. It's been combining big data, cloud and traditional database capabilities. We've always had aspirations to do that, but you first need to get your data warehouse into the cloud. And then you can kind of adopt the data lake strategy. We expect more customers to rely on Snowflake for data lake workloads as we introduce support for unstructured data later this year. We have some very large customers that are -- have been for some time using us as a data link. This is not something new. And I think the uptick in our data lake migrations shows our ability to serve as customers' primary data platform because with -- as we improve the performance and efficiency of our system, we become cheaper that you can use this as a data lake for all of your data. Think about the compressions -- their new compression technology that we rolled out makes storage so much more efficient. We talked about we're always working on improving performance. We talked about we have a new chipset coming up next year that we're going to be rolling out in the cloud providers that we've been working with them that will further improve performance that makes us cheaper and more economical for people who use this as a data lake for their primary data.

I see. Okay. And I guess, just competitively, we often get asked this question, just what -- maybe give us a sense for what you are seeing among the legacy providers? I know Teradata last week had an Analyst Day and threw out some pretty large cloud numbers. We've seen some acquisitions in the Hadoop space. What's kind of the lay of the land? And how much of that do you think is -- those changes are helping?

We -- I'll talk to Teradata specifically. We never see Teradata in a competitive situation with moving from an on-prem to the cloud. And because at that point, the customer has already made the determination. They want to get away from the cloud. Our biggest competition seems to be with customers may kick the can down the road another year before they do their big Teradata migration so they may renew with them for another year. But we actually are not aware of any competitive situations with Teradata. We compete with Google, Microsoft and Azure, period, on these big on-prem data warehouse migrations. We don't see -- we replace Hadoop left and right. Don't see Cloudera competing with them in any way because once again, the customer has already made the decision they want to get off on-prem. It's really the 3 hyperscalers we're competing with there.

Right. Right. And I think in the past, kind of going back to the comments around the time it takes for large enterprise customers to really start to ramp in revenue. I think you've talked about 9 to 12 months. Is that a reasonable time frame in some of these more complex migrations? Or do you think that there's things you could do to maybe shorten that going forward? Just give us a sense for how you're thinking about that.

These on-prem -- it's not just a matter of transferring the data. The way the on-prem systems store data with proprietary code, you have to first translate the data. And then you move it, there, it can be very cumbersome. You have to understand these systems that are being replaced are 10, 15, 20 years old. They're getting feeds from hundreds of different sources across the enterprise. And then there's -- could be thousands of different reports that are being written on those things. So it's a cumbersome process to do these migrations. With that said, at our Investor Day, Deloitte talked about that they're accelerating migrations by up to 50% when customers adopted automation tooling that they had and apply various best practices. But there -- it's cumbersome, and it takes time to do, and I don't see that dramatically shortening. We're working on it a lot. We're developing tools for Teradata. We have those done. We're working on other technologies as well for other on-prem systems we're replacing. As well -- and likewise, the GSIs are all trying to develop these as well, too. It's just -- it's a hard process. It takes time.

Yes. It's probably also kind of a people and business process challenge, too.

Correct. It's a huge thing. And you don't just give it to a third-party GSI to do. It's like any migration. The users have to do all kinds of user acceptance testing to make sure that the data that they're seeing in the new reports is actually the correct data that's being pulled the same way it was before. It just takes time.

Yes. Yes. Absolutely.

I've been seen in a year, I've done a lot of ERP migrations, and they haven't shortened in time. They take the same amount of time and the biggest bottleneck is usually internal resources within your organization that have to be involved in the process.

Right. Right. When we think about the demand environment, I know, from a COVID perspective, you've commented in the past that you weren't really negatively impacted. But curious if you've seen any incremental tailwinds in economies or industries that have begun to reopen, such as travel or retail or oil and gas. Just curious what you're seeing.

We did say and we still believe this -- the pandemic was really net neutral to us from a revenue standpoint with some organizations, they expanded their use of Snowflake while others slowed down depending on the industry. Notably, many of the customers that grew dramatically during the pandemic are continuing to expand their use of Snowflake today. So I think we might have accelerated some of the movement for some companies to get on Snowflake, but those guys continue to expand. We're also seeing negatively impacted industries open up in terms of their migration planning. They haven't gone away. And so net-net, I think it accelerated some of the movement to Snowflake, but we don't see it slowing down as the economy. And I don't expect when COVID, if and when it goes away, it will go away at some point in time. I don't see that slowing us down in any way. I think the whole cloud adoption is here to stay by people.

Right. Right. And we got a couple of audience questions come in, they're kind of both on the same topic interestingly. But just if you could touch on your positioning relative to Databricks, your view of where you are today and when and how you could potentially intersect down the road.

Yes. So we coexist in many accounts. Databricks is really more around data engineering, data science. They do a lot of the ELT, the extract load transform with our customers that we see that we don't really do that. And so we do not compete with Databricks on enterprise-wide data platform. We just don't never see them in a big Teradata replacement or big on-prem replacement just does not exist. Where we intersect is more around some of the stuff we're doing with Snowpark, with our app development where data engineers, data scientists will be able to do more in Snowflake, especially as we roll out Python capabilities that we've talked about. That's more so. It seems like they want to compete with us. We were happy to be a partner with them, which we were. We had brought them into many, many accounts early on.

Yes. Yes. So maybe to touch on that, I think Snowpark was one of the bigger highlights both from the Investor Day and in some of the events that you've held in the past. Just give us a sense for does that target different personas in an organization? Does it open up new workloads? How do you see that kind of expanding the opportunity for you?

It really targets the app developer, the data engineering, the data scientists, and it definitely opens up a new class of people. Now -- and by the way, we started working on this over 2 years ago. It just takes time to build these things in. We had chose to first support Java and Scala, but there seems to have been a big movement. And we saw this coming, where Python seems to be one of the most common ones that data scientists and data engineers, and hence, why we're rolling out Python. We've been working on that for a while, too. It just takes time. And the issue with Python is a little bit of the Wild West where there's many different open source, there's many different. We're going to have an enterprise class Python library support. So that you know your -- there's -- from a security and other things that everything is being done properly, and that's why it takes time.

Right. Right. So sticking on the analytics front, I mean you talked about earlier just how Snowflake has really enabled kind of new category of use cases beyond what you've seen traditionally in the data warehouse side, kind of this idea of democratization. I'm curious, on that analytics and insight side, how are you thinking about the democratization in there? Because on one hand, you have both partnered and invested in companies such as DataRobot. I think there's certainly a number of kind of fast-growing companies in this space around providing kind of low-code tools to business users to do analytics and machine learning processes. How do you kind of think about your role in that market?

We are going to be an open platform, and we want to let our customers choose best-of-breed technologies for them, and we're happy to partner with all of these guys. And as you said, we have various investments in many of them. Some we don't have investments in, but we have tight partnerships with them. And for us, it's all about them driving consumption in Snowflake. And all of these guys are driving consumption in snowflake. And I think they like working with us because we're not trying to compete with them. Unlike some of the other people out there that partner and then they want to try to compete with you. We are very, very -- we're going to -- we want to remain an open platform and let customers choose best-of-breed technologies that they want to use. And at the end of the day, we got a revenue from compute. That's where we make our revenue, and that's the value of Snowflake, our software. And any of these partners that are going to drive compute, we're going to partner with them.

Yes. And I guess, related to the earlier question on data risk. Do you think that openness of the platform is an area that you'd point out as a differentiator?

I would say that I know we are a very open platform, and we partner well with people and partners like to work with us.

Yes. Okay. Okay. Makes sense. So one of the other things that I thought was interesting from the last earnings call, just related to the Data Cloud and I think in data sharing specifically, you talked about -- threw an interesting stat, 20% quarter-over-quarter growth in stable edges, is I believe what you referenced. I guess, number one, how should we think about stable edges? And is there a tie-in to future consumption or revenue from that?

Yes. So the first is an edge is a data relationship between 2 parties where they're sharing data. And we didn't think that was that exciting because you could just access data once and then say that's an edge. So we came up with this concept of stable edges. It's our definition. It's something that we've come up with. And the stable edge shows more of an enduring relationship that's going on for some time. And it's really, we look at -- though to be qualified as a stable edge, you have to consume more than 40 credits per day consistently over a 6-week period. And if you've been doing that every day, over a 6-week period, then we consider that to be a stable edge because that's some -- that's a data relationship that people are actually relying on for something within their business. And we're very happy with the growth we're seeing in our stable edges. Yes, I would love every one of our customers to do some type of data sharing, but I don't think you're going to get 100% and you're not going to get 100% stable edges because some people, with the nature of their business, they just don't want to share their data, so...

Right.

But what I will say is the more stable edges you get, the more stickier platform and that whole network effect.

Right, right. Okay. That's interesting. And just in terms of the opportunity for data sharing, how do you kind of size it? Or think about it on a relative basis to traditional data warehousing?

What I would say is it's a massive opportunity. It's really hard to size. We're really creating it because no one was really doing it before. But think about all the stuff that was being delivered through APIs and whether you think about any data provider that was providing data through APIs. Why do you need to do those APIs? It should just be done through a data sharing capability, much, much more efficient, more [ governed ], a much better way to do it. So I think it's a massive opportunity, and I think it is bigger than the whole data warehousing market and the opportunity.

Yes. I guess on that point on APIs, I mean, even that kind of gets me into some more transactional use cases in addition to the analytics use case, right?

Yes. But I'm not saying we're getting into the transactional use case directly. I would say, as our performance improves, and what we're doing a lot of those things that were being done in transactional databases can be done in Snowflake.

Yes. Okay. Okay. So if we -- maybe just -- we can shift gears to the international opportunity. It feels like this is a market, especially when we look at the percent of revenue coming from outside the U.S. is still relatively small compared to what we see at more mature software companies, could you just kind of compare and contrast what you're seeing in your various theaters? And how should we think about some of these really emerging markets like the EMEA? Is that comparable to where the U.S. was a couple of years ago? I know the adoption patterns of cloud can be a little bit different from a regional perspective.

Yes. No, EMEA, obviously, we started out in the U.S., and the U.S. is such a big opportunity that the U.S. growth has outpaced everything. With that said, EMEA and APJ both had great quarters last quarter and their businesses are growing very nicely. We have -- and this is, by the way, why I want to get away from this whole concept of Fortune 500 because that's such a U.S.-centric term. It doesn't include all the -- we have landed a lot of large accounts in Europe that are just growing. I talked about that big retailer in Europe. They will probably be one of our top 10 customers next year. And we see a lot of opportunity in Europe. We hired a new leader in Europe, almost 2 years ago now, I guess, it will be. And not quite now, not quite 2 years ago, it's about 1.5 years ago, and she moved from the U.S., and she's doing a phenomenal job of building the enterprise sales force. We had a, I would say, more of a commercial sales force in Europe where she's been transitioning that, and we're landing a lot of accounts. Frank is actually spending a lot of time in Europe meeting with these large customers. I would say, APJ, we hired a great guy who was at VMware for, I think, 15 years who've built the whole VMware APJ region up. And he is doing a phenomenal -- he's been with us now, I don't know, 9 months or something like that. He's doing a phenomenal job with the new Cap 1s we're landing and also the profile of salespeople that we're bringing in there. So we're very optimistic and we've landed some big accounts in APJ, and we have a lot in the pipeline.

So it's fair to say that you're -- just by virtue of numbers, you're hiring faster in terms of sales capacity internationally than the U.S.?

As a percentage, we're growing it more, but absolutely, we're still -- there's still so much opportunity in the U.S., too.

Yes. Yes. Makes sense. And I guess thinking about the U.S. for a moment, give us an update on the Federal business. Where you are? How you're thinking about that opportunity maybe this fiscal year as well as next?

The Federal and public sector, so first of all, we told you we had FedRAMP moderate. We need FedRAMP high. Federal government, most agencies are wanting FedRAMP high today. We are in the Azure and AWS GovClouds that they have. But we should have our FedRAMP high next year, and we are working a lot of opportunities. I think, next year, we should have some nice new logos in the Federal space. But from a revenue standpoint, it won't really start to kick in until 2023 because of that ramping it takes to get there. With that said, we do a lot of public sector. We do a lot of state and local stuff we've been doing in the U.S., and there's a number of opportunities internationally as well, too, we're working on. So I would just say the whole public sector is upside for us.

Yes. Yes. For sure. So I wanted to ask you about the kind of the hiring environment as well as kind of how Snowflake's thinking about the future of work. A lot of our companies have referenced a challenging hiring environment as it relates to the great resignation that everyone talks about. Some companies, they even had to do kind of onetime salary increases. What are you seeing on the hiring front? How is just overall morale now that we're 1.5 years into the pandemic?

Yes. I think morale is very good. I'm not hearing anything negative. I will say it helps when you're a public company and you're highly valued, that helps with morale. Let's not kid ourselves, that helps a lot. But we continue to hire. We have pretty aggressive hiring plans. We'll add 1,200 people this year. I feel good that we'll get there, but we are selective in where we hire. I will say that the Bay Area is extremely competitive for engineering talent, hence we add a lot of engineering talent in Seattle, but we're also building out Berlin and Poland and Warsaw. We're in Poland and Toronto for engineering talent. And we're going to go to where the talent is, but we want to keep it to core development centers. And the salespeople at the engineers, it's about -- they want to work on interesting things where they can make a difference and do stuff, and we think we can offer that to people. And the nice thing about a Snowflake is you can see the impact of your work directly in our results because we just do one thing. And salespeople, at the end of the day, it's about where they're going to make good money. And so you have to offer a competitive package you can get, and we're getting very good people so.

And I think you're in an interesting position, just given how early you are in the company's life cycle with some of the dynamics around future of work. But obviously, I know I think you guys continue to be a remote primary company. How are you thinking about that long term? And especially as you look like at areas at really high cost of living, high real estate costs,too, how do you think about the long-term mix of in the office versus remote and potentially that being a differentiator to your financial profile?

Yes. So I do expect we will return to the office. When that is, to be determined, but I don't think it's going to look the way it did in the past. I will say people are much more open to some level of remote work. That doesn't necessarily mean full-time remote work, but more flexibility and where you're going to go into the office where there's a purpose. I have to tell you, I was totally against remote work. I used to be. But what I realized is during the pandemic is we haven't missed a heartbeat. Productivity is up, I think, in my people in the organization. And quite frankly, I've actually enjoyed -- I do not miss jumping on a plane. I get to sit here on my ranch right now in Montana and it's much more enjoyable than sitting in the Bay Area, quite frankly. So much so I never want to go back to the Bay Area. So we're going to be open with our employees. I do think, though, it is super important. I learned that engineers need to work in groups when they're working on projects and you can't -- assume just doesn't do adequate that personal interaction. So there needs to be some in person in the office, but that doesn't mean 5 days a week. I think we're going to be doing a lot more hoteling in offices, which is going to keep our real estate footprint down. But the office is not going to go away. Not for -- I don't see that happening.

Yes. Kind of striking that right balance between in person, right? Well, Mike, we are out of time. Thank you so much for finishing us off here as part of our global tech conference. It's great to see you in your beautiful ranch house. I'm in a conference room here in a hotel. But thank you for the great discussion, and thanks to all the investors for the questions and for participating in the tech conference.

Thank you for having me, Tyler. Bye now.

Absolutely. Take care.