Snowflake [SNOW] Conference call transcript for 2021 q1
2021-05-26 20:56:06
Fiscal: 2022 q1
Operator: Good day, and thank you for standing by, and welcome to the Q1 FY '22 Snowflake Earnings Conference Call. . I would now like to hand the conference over to your speaker for today, Mr. Jimmy Sexton, Head of Investor Relations. Thank you. Sir, please go ahead.
Jimmy Sexton: Good afternoon, and thank you for joining us on Snowflake's Q1 Fiscal 2022 Earnings Call. Joining me are Frank Slootman, our Chairman and Chief Executive Officer; and Mike Scarpelli, our Chief Financial Officer. During today's call, we will review our financial results for first quarter fiscal 2022 and discuss our guidance for the second quarter and full year fiscal 2022.
Frank Slootman: Thanks, Jimmy. Good afternoon, everybody. We reported strong Q1 results with 110% year-on-year growth to $214 million in product revenues, reflecting strength in Snowflake consumption. Remaining performance obligations grew 206% year-on-year to $1.4 billion, indicating strength in sales across the board. While maintaining a net revenue retention rate of 168%, we also generated $23 million of adjusted free cash flow in the quarter. We are expanding our geographical scope in all three major theaters. Both EMEA and APJ have breakout bookings quarters. EMEA grew more than 200% and Asia Pacific grew more than 300% year-over-year. At the end of Q1, we had 104 customers with over $1 million in product revenue, an increase from 77 in the previous quarter. During Q1, we had key enterprise wins, including Datadog, Equifax and Walgreens Boots Alliance.
Michael Scarpelli: Thank you, Frank. We saw continued momentum in Q1 with another quarter of great execution at the start of our fiscal year. Our Q1 product revenues were $214 million, representing 110% year-over-year growth. Our remaining performance obligations were $1.4 billion. Media and telecom, technology and financial services customers drove the outperformance, and we saw meaningful growth from our health care customers. Our strong RPO results reflect more multimillion-dollar relationships with particular strength in the telecom and technology sectors. Of the $1.4 billion in RPO, we expect approximately 54% to be recognized as revenue in the next 12 months.
Operator: . Your first question comes from the line of Brent Bracelin with Piper Sandler.
Brent Bracelin: Mike, maybe I'll start with you and a quick follow-up for Frank. Consumption in Q1 here was very strong, $35 million increase in product revenue and another triple-digit product growth quarter. As you look at kind of Q1 consumption trends, any seasonality clearly didn't show up. Was that surprising? Any other surprises just as you looked at individual customer consumption trends that drove the outperformance in Q1?
Michael Scarpelli: Remember, our revenue is based on actual customer consumption. It's not a ratable recognition. So we did have -- as I mentioned, we did have some outperformance in some specific customers. And part of that is tied to certain projects they're working on. And just because a customer consumes a certain amount one quarter doesn't mean they're going to consume that same amount next quarter if they're not running those same projects. And so we are very happy with the revenue outperformance in Q1. And our guidance reflects where we think we're going to be for the balance of the year.
Brent Bracelin: Got it. And then, Frank, just following up for you here. If you think about kind of the data sharing opportunity, it clearly seems to be catching on like wildfire, so much so you're starting to see the creation of even new open source projects that are also tied to data sharing. As you think about the 27 net new $1 million-plus customers, do you think data sharing is now contributing to these $1 million-plus customers? Or do you think a bulk of the big spend is still tied to kind of replacement opportunity around data warehousing and the data share drivers still to come?
Frank Slootman: Well, Brett, the answer is really both. Customers need to walk before they run. So a lot of their initial focus is on workload transitions and so on. But everybody has a very, very keen view in terms of where they want to be, where they're going to be over time. And you're exactly right, data sharing is absolutely essential. It will really enable data sciences in order to create context around data, in order to enrich data to really fully reach that potential. I mean it's taken us some time to really evangelize this whole idea. But as you said, it is resonating very, very aggressively in the marketplace. We have a lot going on in our vertical markets that are very specific to those contexts, very specific to the unique circumstances of vertical industries. So we're really excited. We're very well positioned for it. And we have really developed our Data Cloud and our Snowflake Data Marketplace to the point where it's completely operational. And hope we'll see you at our Investor Day because not only we want to talk about it, we're also going to show you a bunch of stuff and see how far we've come in that area.
Brent Bracelin: Looking forward to it. Thanks, Frank, Mike, for another great quarter here.
Operator: Your next question comes from the line of Keith Weiss with Morgan Stanley.
Keith Weiss: Excellent. And very nice quarter. Frank, I wanted to continue the discussion on edges and the percentage of customers that you're kind of adding, the edges going from 10% to 15% and the amount of edges up 33% quarter-on-quarter. Is that something we should think about as going to directly impact consumption? Have you seen that the more edges the customer have, the more data consumption is going on within their data warehouse? Or is this more of a construct about stickiness? Once you're getting granted all these partners, you're never going to stop using the Snowflake solution. How should we think about the like the financial impact of this growth in edges?
Frank Slootman: Well, the answer is both, obviously, it's going to make things incredibly sticky. But for us to designate something as an edge, it has to have a minimal amount of time in terms of durable consumption. So there's very much a consumption dimension to these edges that we track and mods or otherwise, we don't consider it an edge in our world. In other words, they have to be durable and stable because a lot of data relationships are transient. They exist for a period of time. They exist for a project, maybe a trial, whatever it is. But that's not -- that's also a metric that we follow where we really look for, what we'd call, stable or durable edges. So they exist over periods of time. They drive consistent consumption. So the answer is absolutely both to your question, yes.
Keith Weiss: Got it. And if I could sneak in one follow-up for Mike. The 27 customers getting to that $1 million-plus level this quarter, that's a real eye-popping number. I think that's more than you did in the first three quarters all of last year. Anything in particular getting that motion going faster? And on the other side of the equation, is that a number we could expect to see on a go-forward basis? Or is that kind of too high of an expectation to have for the remainder of the year?
Michael Scarpelli: Well, I'm not going to guide to $1 million customers. All I will say is there's a number of -- as there was going into this quarter, we had a number of customers on the cusp of going to $1 million, and we continue to see a number, and I think it's going to be very strong growth. But what I want to remind people is, when we land a customer, it takes many times, 6 to 9 months -- I think it's closer to 9 months -- before a customer actually starts to consume at their contract rate. And so a lot of this is the impact of really focusing on larger customers over the last 1.5 years that we're starting to see that pay off. But what I will say is of that 104 customers, only, I think, about 25% of those are actually majors, and then the balance is enterprise customers. So think about Fortune 500. Only about 25% are Fortune 500. The others are across-the-board end customers. And my point there is even small companies can be big consumers of Snowflake. And there's a lot of them, as I said, that are just on that cusp. So we expect that to continue.
Operator: Your next question comes from the line of Derrick Wood with Cowen.
James Wood: Great. Congrats on a great quarter. Frank, I'd love to hear about what's causing the breakout and growth in international regions? I know you guys had some leadership changes. I'm sure you're feeding more headcount there. So how much is kind of your own efforts versus other factors like market awareness or growing cloud acceptance or anything else you'd call out?
Frank Slootman: No, it's exactly what you just said. I mean we just needed to properly operationalize ourselves in these geographies. As you know, it's market by market. We have to have the correct leadership in place. We have made a lot of leadership changes in these regions that we're very pleased with. When you have a great product like Snowflake, I mean the impact of that is going to come fast and furiously. So I'm personally going to invest a bunch of time in Europe, given my own background, because I think the opportunity is tremendous. So we're excited that we actually see these regions coming online and contributing and we expect that to continue. We're very happy with the latest changes we made in Asia Pacific as well. We have very high expectations of Japan, obviously, ANZ and there's other markets where we're going to be starting to open up as well.
James Wood: That's great. And maybe one for Mike. I mean numbers look great across the board. The one outlier was the Fortune 500, which looked like kind of a slower net add in Q1. Anything you could kind of speak to this number, maybe seasonality or thoughts around what to expect going forward.
Michael Scarpelli: Well, these large accounts are very, very long sales cycles, and you are going to see lumpiness in the additions. Obviously, Q4 was a strong quarter. And as one would expect, that's just landing a customer. That doesn't mean it contributed to revenue. As I said, most of those Fortune 500, we landed in Q4. We've seen virtually no revenue from them yet today. I can't stress that enough. And given Q4 is the end of a commission year for people and accelerators, reps do everything natural to close everything in the end of that commission year. So I fully expect we'll continue to close Fortune 500 the balance of the year. And it's all based upon when the customer is ready to begin that journey.
Operator: Your next question comes from the line of Kash Rangan with Goldman Sachs.
Kasthuri Rangan: Congratulations on the quarter. Two questions. One is you talked about never-ending replacement of data warehousing. If you could just expand on that, I'm fascinated because we seem to think -- at least I seem to think -- the data warehousing is just a small subsegment of the database market. But your comments seem to suggest that it's a longer-term, longer-tailed growth opportunity, that very core of the business. And one for you, Mike, I know that for your guidance, it doesn't look like you're assuming a big flow-through of this beat. Is that again because you do not want to forecast increases in consumption revenue? Is that why the revenue forecast for the upcoming quarter is conservative?
Frank Slootman: Yes. Kash, Frank. Yes, we're actually super early innings on replacing these legacy on-premise data warehouses license. And actually, you see that in some of Teradata's number that they're actually hanging on to their business, one of the reasons it's been. We think Snowflake has really been the only company that's been successful in transitioning these legacy systems. We've not seen it done successfully by the public cloud companies. So most of that opportunity is still there and it's still coming. That's why I think it is, for all intents and purposes, indefinitely. So much of our business is actually not coming from those sources. But our expectation is that it will continue to contribute materially to our business for a very long period of time.
Michael Scarpelli: On your question, Kash, on the beat, we did flow the beat through to the full year guidance, plus about $1 million more, but we also had the headwind going to the full year. As I mentioned, we introduced new storage compression technology that we literally just rolled out. And based upon the early feedback of that, it's going to take about $13 million of our revenue away from the company because the economics of our storage is so much better for our customer, with that compression being much higher than we were expecting it to be. Switch is a good thing for our customers. And longer term, it's going to drive more -- it's going to cause customers to put more data in Snowflake, which will ultimately drive more consumption.
Kasthuri Rangan: Wonderful. And Frank, if I could just follow up very quickly. The cloud-based -- the hyperscalers have their own data warehouse cloud offerings. Why is it that Snowflake has been able to keep them at bay? I know they are partners of yours. That doesn't get in the way of sometimes you compete in this market. Any structural barriers and advantages that you guys have against the hyperscalers that have their own cloud-based data warehouse?
Frank Slootman: Well, I can go on and on about what all the reasons are. But for the purposes of this call, one of the things that makes Snowflake completely different is that our founding team started with a clean sheet of paper. They obviously were deeply steeped in database technology over a very long period of time. And they were looking to absolutely not carry any legacy forward that they didn't like and really reinvent architecturally for cloud scale computing, which is very, very different from on-premise. So it was incredibly different, very, very innovative. As a result, we're not straddling the on-premise and public cloud environments. We're only in the public cloud. And it's very important. I mean you look at a lot of the public cloud companies. They have carried, not only architecture, but actual code forward from our private environment. They've tried to evolve that and adapt that. The thing that we always say is say, look, it's hard to catch up when you're not sitting on a good architecture. You're only going to get further behind. Architecture matters. It matters a whole lot. And this is really what the strength of Snowflake is rooted in its core architecture, something you should never lose sight of.
Operator: Your next question comes from the line of Kirk Materne with Evercore ISI.
Stewart Materne: Frank, I was wondering if you could maybe just expand a little bit on your comments on the verticals. It seems to me, when we start thinking about the data sharing opportunity, landing some of those key beachhead clients in each vertical is going to be really important to sort of build out beyond that. And I guess, along those lines, I was just kind of curious where you think you are in each of those 6 verticals. There's a couple that are maybe ahead of the game, on sort of building out sort of real market knowledge, having salespeople and understand the intricacies of each of those verticals versus maybe the ones that you still need to do some work on.
Frank Slootman: Yes. That's kind of a broad-ranging question. And again, we definitely come and see us at Summit because we're going to showcase a lot of customers in these verticals. You can hear directly from them in terms of what they're doing and how they're doing it. But our largest vertical, and this may actually surprise some people, is actually in media, in streaming content and advertising. It's actually not that much of a surprise because they are a digital-to-consumer business in a lot of way for companies like Hulu and Disney and Comcast and NBCUniversal to render businesses. And they are very sophisticated, very advanced. And obviously, because of the dominance in advertising of the likes of Amazon, Facebook and Google, this is very, very important to them. How to develop their businesses in Snowflake is really a key enabler for that for many reasons. And so that's just one example of a business that is extremely active with Snowflake at this point because of the strategic challenges that are going on over there. Second largest vertical is actually financial that might not be large financial institutions. That may not be much of a surprise because they tend to be -- for most software companies, health care, life sciences is huge. Retail, obviously, in consumer packaged goods is very, very big for us for similar reasons because digital transformation is so large in that part of the world. Of course, the advertising side, a very big thing over there as well. Public sector software companies is a very large component of our business because many, many software companies out there, and you'll see a lot of that at our Summit event, are re-platforming on top of Snowflake, Snowflake becoming a core part of their stack. So there's a lot of moving parts in our world. We're really excited about progress we have made as a company that we no longer sell exclusively on architectural distinction. We really crossed over into the customers' business, really enabling their challenges, their outcomes. And it makes us a much higher value-added partner than we historically have been when we were just focused on shifting and lifting workloads from on-premise to the cloud.
Stewart Materne: Yes. That's really helpful. And Mike, if I could just squeeze in a really quick one. Given your consumption model and we're heading into the summer season, is there any way to think about in terms of seasonality from just consumption over the summer? Over the years, we've seen sort of bookings trends, obviously, come down a little bit in say the third quarter. Is there anything we should consider in terms of your full year guidance on that front?
Michael Scarpelli: First of all, it's factored into that, but there is no -- you see more seasonality on weekends, obviously, because there's not as many employees in the office, but a lot of our workloads still run on weekends, too. So it's not like it drops to 0. So clearly, around holidays, you do see some. But it's not as profound in terms of people taking vacation in December that you see a big drop off consumption. We really haven't seen that.
Stewart Materne: Congrats.
Operator: Next question comes from the line of Karl Keirstead with UBS.
Karl Keirstead: Okay. Maybe two for Mike. Mike, any noticeable change in the revenue mix between Snowflake on AWS versus Snowflake on Azure? And then I've got a quick follow-up.
Michael Scarpelli: I will still say AWS is still our biggest. But Azure from new bookings, as we get into large enterprise, that continues to increase as a percent. But still, AWS is our biggest from new bookings.
Karl Keirstead: Okay. Makes sense. And then, Mike, just as a follow-up, interesting on the storage compression change and the delta in terms of the revenues. I'm just wondering whether that's symbolic of any change in pricing strategy. Maybe you, Frank and the team have a greater willingness now to pass these types of cost savings on to customers to drive future growth. Or was this change a little bit more onetime as just opposed to a real high-level pricing strategy change?
Michael Scarpelli: Well, I'll be transparent. I learned a lot in this process. So this is something the company has historically done about every two years. There's a big focus on new compression technology for storage. And the impact of it was bigger than we would have thought. And we only knew that once we actually got real live examples from customers. And our philosophy has always been to pass that on to customers. But there's other performance improvements as well. For instance, we're working on a new chip technology that will dramatically increase performance or improve performance. We do expect that to have an impact, and that's more of next year you'll see that. And we have always done that and continue to improve the performance of our product that goes directly to the benefit of our customers.
Operator: Our next question comes from the line of Gregg Moskowitz with Mizuho.
Gregg Moskowitz: Congratulations. I had a follow-up on secure data sharing. So as that ramps, naturally, your consumption is going to rise by a lot but so too will the complexity of all the data sets being shared in all kinds of intricate ways. And so I'm wondering, do you have any concerns about maintaining a high level of security and governance on the platform as that unfolds?
Frank Slootman: It's Frank. The answer is no, not really, because data sharing is completely integral to our architecture. In other words, we're deploying the exact same security model that we deploy for internal as we do for external security. So as I said earlier, architecture matters. And this is 1 product that is beautifully designed, doing exactly what it does. It is not a bolt-on. It's not a hack. It's just incredibly well implemented. So the answer would be no to your question.
Gregg Moskowitz: Great. That's helpful. And then a quick follow-up for Mike. So you mentioned that EMEA bookings grew over 200%, Asia Pac over 300%. Does this reflect acceleration? Can you say what the growth looks like in fiscal '21?
Michael Scarpelli: What I will say is both EMEA and APAC exceeded their plan. They had very strong quarters. We saw $1 million plus -- it was actually a multimillion dollar deal we did in EMEA with a big pharmaceutical company. And we landed a very nice customer in Japan as well. So very strong across the board, I would say, performance in EMEA and APAC.
Operator: Our next question comes from the line of Raimo Lenschow from Barclays.
Raimo Lenschow: Congrats from me as well. A quick question. Frank, as you look into the replacement cycle, and you kind of pointed out you're the one that actually does it properly, what do you think around the partner capacity to help you there? You mentioned Deloitte already, obviously, as kind of one big one, but I assume just getting off of Teradata is actually quite a big job with kind of a lot of extra work involved. How happy are you with the channel there? And then I have one follow-up for Mike.
Frank Slootman: Well, we're happy in places I highlighted. Our relationship with Deloitte, who is our lead partner, they went from a standing start a little bit over a year ago to $100 million of business, which is an absolutely ripping trajectory that they're on. It just shows here that demand for these migrations is enormous. We have a lot of engagement from all the other big names out there, whether it's Capgemini or Infosys, obviously, Accenture and so on. But they're all scrambling to certify, to staff, to provision. Our professional services organization is actually, by far, the best at this. And that's obviously a logical consequence of the fact because this is all we do. So we're really using our own abilities to help leverage them into the business. And it's not easy because the ramp is so steep for everybody, and we've got to make sure that we do an absolutely terrific job for our customers because these migrations are not easy, they're not cheap, there's risk involved and so on. And that's really the friction in the marketplace. So it's actually very welcome that the system integrators are leaning in as hard as they are, but the enablement for them to become really effective large grown businesses is sort of the day-to-day challenge that we have. But I view it as all good work that we do.
Raimo Lenschow: Yes. Okay. Makes sense. And then, Mike, the storage compression, et cetera, like does that impact gross margins? This might be a stupid question, but like the 72% is really, really high. So now everyone is wondering like how do you manage to get so quickly there and is there kind of more upside next year with more compression, et cetera. Or is it totally unrelated?
Michael Scarpelli: Well, what I will say is it does improve margins. And the way it improves margin is because storage becomes more efficient. Storage is a smaller component of the overall mix of the revenue, and compute is the real value of our software that drives more margin. And I will say we did roll this out in April, and you do see some of that coming into an impact on last quarter. But we did say at our IPO, if you remember, we thought we could get to the mid-70s. That might feel very good that we'll get to the mid-70s. It's going to take some time. And stay tuned for our Investor Day, and we'll talk more about that later. I would say the biggest improvement we've seen to date in the gross margin has really been the renegotiation of our contracts with our cloud vendors and the discipline in our sales organization around discounting and coupled with the fact that as we move into larger enterprises, we're selling more business-critical enterprise, which attract a higher contribution margin.
Operator: Our next question comes from the line of Brent Thill with Jefferies.
Brent Thill: Mike, just on quota-carrying reps this year. If you're not giving a number, can you just give us a sense of are you going to keep the same trajectory of growth you added last year or trying to get more productivity out of the reps you added. Just any color around on the quota-carrying side would be helpful.
Michael Scarpelli: Well, we don't disclose quota-carrying reps. And what I will say is we're going to add about 1,200 net employees for the full year, and we do expect that we'll add about the same level into our sales organization this year as we did last year.
Brent Thill: So same number of absolute reps?
Michael Scarpelli: Correct.
Operator: Our next question comes from the line of DJ Hynes with Canaccord Genuity.
David Hynes: Frank, I want to ask about Snowflake Ventures. You called out ThoughtSpot, DataIQ. I know DataRobot's in the portfolio. So obviously, analytics and ML, AI are logical targets. Just curious like where else are you seeing interesting stuff happening in the ecosystem as it pertains to driving volumes to Snowflake?
Frank Slootman: Yes. Areas like governance. Obviously, governance is becoming such a huge aspect of data operations in large institutions. When I say governance, it really relates to security as well as probably to see compliance. So we actually acquired a company last year by the name of CryptoNumerics, and that is now actually flowing into our platform, and we have the ability to anonymize PI data and things of that sort. Super, super important that customers can really feel safe on our platform and they allow data on our platform, that it is fully governed. I mean it's been a big issue, but it's literally growing in importance quarter-on-quarter. And a lot of our large customers are really organizing themselves to maintain that kind of posture. There's areas in data cataloging. I mean there's a whole ecosystem around our platform. What we like about investing is not just the upside in terms of the investments, but we get to build closer relationship with these customers on the basis of the investments, much more seamless integrations, better customer experiences. And that's what we like to do. I mean we're very much an ecosystem-oriented company rather than we have one flavor, and that's what you're going to use. We want to make sure that the whole ecosystem feels like a very good experience to our customers. So we look for opportunities. We see a lot, and we're happy with the number of investments that we've done, and we're looking at new stuff continually, yes.
David Hynes: Yes. Yes. Obviously, some good ones in the portfolio already. Mike, a follow-up to you. So Raimo asked about gross margins as it pertains to kind of storage compression technology. I was going to ask -- if I think about the trajectories kind of you guys expand into, doing more with unstructured data, I think there are heavy demands on storage there, like how should we think about the trajectory of gross margins as we get to that mid-70s target? I mean is it fair to think that it could dip in the interim? Or what's the right way to think about that?
Michael Scarpelli: I don't see a dip happening in our product gross margins at all, but there is a limit to where you can get to. And when we're going through our IPO, people were asking questions. I did say I don't see us getting into the 80s. I can see a path to the mid-70s. We may, one day, be able to get into the high 70s. But given the storage component and the costs associated with the public clouds are in there, it's pretty hard to get beyond this.
David Hynes: Makes sense. Nice numbers, guys.
Operator: Our next question comes from the line of Brad Reback with Stifel.
Brad Reback: Mike, I think last quarter, you talked about investments in FedRAMP. Can you maybe give us an update where you stand? And do you think the federal vertical could be a contributor in fiscal '23?
Michael Scarpelli: So there's FedRAMP high we're working on. ITAR is going to be out mid this year, FedRAMP high end of the year. And obviously, if we didn't see a big opportunity, because there is a big cost associated with doing that, we wouldn't be doing it. And we have a very good pipeline within the public sector. And we're very focused on it. As of today, it's not a big driver of revenue. So there's a lot of upside there.
Operator: Our next question comes from the line of Tyler Radke with Citi.
Tyler Radke: Maybe to start with Frank, I think earlier in the call, you talked about how kind of the legacy data warehouse migration has been a really important go-to-market motion. I'm curious, given your focus on the Data Cloud, like are you changing at all in terms of the initial use case that you're leading with? Perhaps, given the success you're seeing with data sharing and data marketplace, do you find that those are easier to perhaps land a new customer with given that it's more greenfield and less of a large-scale legacy database migration?
Frank Slootman: Well, the thing that happens is when people move these legacy workloads to the cloud and you're running a platform like Snowflake, all of a sudden, you're finding out that workloads can run orders of magnitude faster. In other words, not like after 3 weeks, I populated my warehouse. I'm now populating the dashboard, running the reports. People are almost -- they've already moved on by the time they get data. When you accelerate that time frame -- we often talk about this whole concept of the time value of data. If you get data -- most of our customers are now on a 24-hour cycle, but that's now moving up, right, where people are seeing data in hours and minutes. And we're very aggressively working to event architectures where you get to sort of near-real time, not in the computer science sense of real time but in the visceral sense, it's near-real time. So you're now looking at data very differently than you did before where you have to wait for days and weeks for it to see it. Now it's showing up very quickly. That completely changes people's perspective what they can do with it, right? And that is what is so different, so different. That is enabled by the public cloud coupled with Snowflake architecture. That is what's driving the consumption of high net revenue retention rates because people are discovering completely new use cases and things they can do that were never a consideration in previous times. It's very important that this whole static batch-oriented way of thinking we have in data warehouses is completely different now. Okay.
Tyler Radke: And then a follow-up for Mike. I know that the current -- you don't really have a current RPO because, obviously, it's a consumption-oriented model. But in terms of your comments on the percentage of RPO that's expected to be recognized in the next 12 months, I just wanted to confirm, it seem like the growth in that number picked up quite a bit from the last few quarters. So number one, I just wanted to confirm the growth rate on that number; and number two, just understand kind of the factors that drove what would seem like a good bookings performance here in Q1.
Michael Scarpelli: Well, it was a strong Q1 and bookings performance. And we estimate that 54% of that RPO will roll off into revenue over the next 12 months, and it was 54% -- 55% last quarter. And I don't know what else there is to say about that, but that's how we see it rolling off to revenue.
Operator: Our last question comes from the line of Kamil Mielszarek with William Blair.
Kamil Mielczarek: Congrats on the great start to the year. I have a follow-up on the targeted 1,200 headcount increase this year. Can you update us on the execution to date? And should we expect that to be front-end loaded? And maybe could you provide the split between the first half and the second half of the year?
Michael Scarpelli: Well, as I said in the remarks, we onboarded 436 net employees in the first half of the year. And so 1,200 is definitely front-end loaded. We always try to add a lot of people into our sales and marketing organization in the first month of the year to get the benefit of our sales kickoff. But definitely, we're going to continue to add in the first half, and it's going to be more skewed to that.
Kamil Mielczarek: That's helpful. And you've delivered incredible RPO growth over the past -- it's been over 200%, I guess, for the past few years now. Can you provide an update on contract duration, how that's changing? Is that something being pushed by your sales team? Or is that just a function of more penetration at the enterprise level?
Michael Scarpelli: Well, last year was the first year we really incentivized our sales force to sell multiyear contracts. And what I would say is now it's getting more into the normal sales motion of our salespeople. It's more natural for them to be going in and asking customers to sign a 3-year contract. Historically, we used to sell 1-year contracts only. And so we still do have renewals. There are still some customers on renewals that only want to do a 1-year renewal because that's what they've always done. And there are some customers, especially ones where they're new customers, they want to do a 1-year contract because they want to get to know us more. Generally, those newer customers we're finding on renewals are doing the multiyear renewals with us as they're seeing everything they can do on Snowflake. But that's fully reflected in our RPO.
Kamil Mielczarek: Makes sense. And congrats again.
Operator: And ladies and gentlemen, this concludes today's conference call. We thank you for your participation. You may now disconnect.
