MongoDB, Inc. (MDB) Earnings Call Transcript & Summary

All right. Good morning, everyone. I'm Chris Merwin. I cover software at Goldman Sachs. And very pleased this morning to be joined by MongoDB. we have both Dev Ittycheria, CEO; and Michael Gordon, COO and CFO. Thank you both so much for being here. We appreciate it.

Thanks for being here, Chris.

Thanks for having us.

So maybe to kick things off. This time last year, this conference actually happened live at the Palace Hotel in San Francisco and that was obviously just before things started to shut down. And since we're not sitting at the Palace now, and we're doing this on Zoom, I thought we could start off by asking you guys what you've learned from this year about work from home, what you think has worked well, what strategies will persist and which ones won't as we move forward in a post-COVID world.

Well, first of all, Chris, it's nice to be here. Thanks for having us. I think we've learned a lot, frankly. For example, if you had told me a year ago that I would see Michael once while we're running a global public company and would it go well, I think you'd be crazy. And frankly, we've only seen each other once in the last 10, 11 months, which was both kind of both fascinating and scary. And so what it's really shown is that we can operate a global company while working from home. I've seen very few of my leaders face to face, but we think we've run the business quite effectively. In terms of product and product development, obviously, we've been able to continue to drive our product agenda and deliver products and support our customers. And then when it comes to go-to-market, which always gets a little bit tricky because you're actually trying to engage with customers, one of the advantages was that we had an insights sales organization that was actually used to engage with customers remotely through Zoom and through phone. And so our field organization leveraged the experiences and best practices from the insights sales team to learn how to engage customers who typically you'd meet face-to-face all the way from just meetings to demos, to showing the technology, et cetera. And that has gone far better than we expected. We've also hired a lot of people. We've hired last year, including contractors, over 900 people. And we radically changed how we onboard people. Obviously, joining a new company is disorienting in many ways, but doing it remotely when you don't meet your boss and don't meet your teammates face-to-face can be quite challenging, but we really invested in making sure we onboard our people so that they can be productive as quickly as possible and feel connected to our mission and to the culture. And clearly, there's been places where there's been more stress. I think parents and young children have found it more difficult to work from home, people who lived in cramped working conditions found it -- let's say living with a bunch of roommates and there's not a lot of private space. They've found it challenging to work from home. But in general, I think we, as an organization, have done quite well, and we're fortunate that we're -- we work in an industry that -- and have a business that can continue to thrive in this pandemic.

Great. So I guess one of the other themes from 2020 was around digital transformations and clearly, we saw a big pull forward there across many categories of software. And so as we move through this year, and we see customers tick through their priorities in terms of system migrations and modernization, shifting those systems to the cloud, I imagine we'll need infrastructure to support those deployments. So should we think about a new wave of demand for the NoSQL database market as well? Just curious how you're thinking through that.

Yes. We definitely believe that COVID has fundamentally forced a lot of companies to prioritize digital transformation and just moving to more modern application stacks. Some of them may have done -- been more reluctant because for culture reasons or for regulatory reasons, but now every company knows that they need to be digital by default. I'm sorry about that. And so the long story short is that the changes have forced organizations to basically look at modernizing, and MongoDB's intrinsic value is all about letting organizations innovate very, very quickly. And so what does that mean? That means that allows developers to move very, very fast in building applications, it allows the line of business to ship products faster and roll out new features faster to the market and it allows CXOs to basically think about seizing new opportunities to respond to new threats. Especially in an uncertain world with COVID, that has become the forefront. So these changes are truly structural, they're not ephemeral. And so we believe that it's just putting front and center the need to do modern application stacks, being able to run your business in the cloud, being able to scale elastically and being able to innovate very, very quickly, which all play to our strengths.

Great. I guess going back to this past quarter, one of the things that stood out was that you accelerated your customer growth obviously in a pretty challenging time. Can you talk a bit about some of the changes that you made in your go-to-market motion, both, I guess, with your self-service and direct channels that were drivers of that acceleration?

Right. So as our Atlas business is growing last year, one of the interesting trends that we saw was, one, people love Atlas. In fact, our NPS Score, which we announced in the Q3 earnings, was 74. The typical SaaS product is about 30. So people really love Atlas. And what we realized was that rather trying to force customers into some sort of commitment, typically an annual commitment, before they can start using it, we said, why don't we just let them get them to use the platform earlier? So we made some changes at the start of this fiscal year, which was pre-COVID where the incentives for salespeople were not such that they could only make money, forcing customer and commitment, that they could also make money by just getting customers on the platform. And that paid huge dividends for us, especially during COVID where customers may be more reluctant and want more flexibility in terms of the economic relationship they have with their vendors and we suggested get them to our platform and our net expansion rates on Atlas are very, very high. And so getting more customers in platform, acquiring more customers more easily displayed huge dividends for us. And we think not only has it helped our business this year, but we think it's laid the foundation for durable growth over the next couple of years by allowing us to acquire more customers. And so those customers, we expect to grow over time and we're going to continue to encourage customers to encourage sales organizations to acquire customers quickly. And we find that this has really a nice flywheel effect between our self-serve business and then our sales business because sometimes customers may choose to engage with us in the self-service channel. And because we get so much instrumentation about what they're doing on their use case, we can very quickly triangulate from a sales organization to help those customers, we think, are most likely to grow. And then that pays huge dividends because it allows us to really add value and enable customers to grow much faster than they would normally. So we're really pleased with the changes we made, and we feel like we're just going to continue to double down on those changes.

Perfect. But I guess like taking a step back, when we look at the sources of the demand you're seeing, can you give us a sense of how much of that is coming from net new workloads like brand-new applications that need to be supported by Mongo or migrations off of apps that we're trying to use a relational database for that purpose? And how has that mix been evolving?

Yes. So when we went public, we -- and this was 3 years ago, we talked about how 30% of our new business was coming from migrations of relational workloads to MongoDB. Due to the growth of Atlas, because when you're choosing running Atlas, by definition, you're already using MongoDB, the percentage came down, but the gross dollars actually increased. What we have seen of late, especially as we introduced new functionality like asset transactions, which provides the highest form of data guarantees, and as customers really view us as not as just a niche database, but a true general-purpose database, that we're increasingly seeing a larger and larger percentage of the wins being migrations of relational workloads to MongoDB. I want to be clear, we're very pleased with also doing -- supporting customers with their new workloads. There tends to be a typical half-life of any application between 7 to 10 years. And so as applications get retired, there are people building new apps, and as people struggle with their existing apps with this -- because the applications are brittle, they can't add new features, whether the application cannot scale or the cost structure of that application is too high, it forces them to consider replatforming. So we're seeing a lot of success across all those dimensions. And as MongoDB becomes a truly viable -- is a viable, mission-critical data persistent platform, you're seeing more and more customers migrate off relational to MongoDB.

Great, great. And you touched on Atlas a couple of times. And certainly, the growth there has been very, very impressive over the last few years really. I mean, I think it's gone from 20 -- like 1/4 of your revenue in fiscal '19, it's almost 1/2 today. So in terms of that business mix shift and the growth we've seen there, can you maybe break down some of the factors for that? I mean maybe COVID played a part in modernization. I'm sure it's playing a part, but just trying to unpack the acceleration -- the growth that you're seeing with Atlas.

Yes. Just to remind people, I mean, Atlas grew 61% year-over-year in Q3. So it's definitely a very fast-growing business. The reason why Atlas has become so popular is that what organizations have realized is that there's real value in focusing on differentiated work versus undifferentiated work. And typically, when you deploy a distributed database like MongoDB on-premise, you have to take on the onus of knowing how to provision, configure, manage, back up, update your environment as you're continuing to build on MongoDB and then as we can add new features and capabilities, update and upgrade your existing MongoDB [ limitation ]. What customers have realized, there's a lot more value for them to focus on building great applications that truly transform the business versus investing in DBAs and other support staff to maintain and manage the infrastructure. So the value of Atlas is that all that magic work is outsourced to us. We take care of it and we, obviously better than anyone else, know how to manage and scale MongoDB, and customers can really focus on differentiated work that really has a competitive advantage to the business. And that's been a big reason why customers choose to go to Atlas. The second reason is that you don't have to prebuild capacity for peak usage. Within the cloud environment, you could scale your demand as your business scales. So if you are a very cyclical business, say, you're an e-tailer and obviously, we just went through the Christmas and holiday shopping season, you don't have to prebuild capacity and then it's unused the rest of the year. You can scale your environment in Atlas and then scale down very, very quickly. We also offer global coverage. So you don't have to have data centers around the world. You literally can run Atlas across 80-plus regions across all the 3 major cloud providers. And most recently, we allowed the capability to run a single application across multiple cloud providers. And why is this important? One for resiliency. We've seen some pretty high-profile cloud outages recently. So you can run the app across different cloud providers and know the application is going to stay up. Two, geo coverage. Sometimes certain cloud providers only have one region in a particular market. And so now you can have redundancy and better coverage in a particular market. And third, you can also leverage best-of-breed capabilities across the different cloud providers, say, one cloud provider has better machine learning algorithms, another cloud provider offer some services that you need for your app. It's -- you can easily do that. And then last but not least, customers also like the fact of having optionality to keep vendors honest, and they never want to be locked in, especially in the database space where a lot of people have bad memories about an incumbent vendor holding them hostage. So for all of those regions, Atlas becomes a truly attractive platform to run apps on.

I would just add from sort of a modeling and a business perspective. While all of those are great reasons, not every company is ready to embrace the cloud. And so when you think about sort of the trajectory and where things are going, what we've tried to help people understand is we really run the business on a channel basis. And the products just sort of come up because there can be a large financial services company, maybe a regulated company, to Dev's earlier point, that can't embrace the cloud or can't be all in on the cloud for whatever reason. And so the ultimate mix shift will really depend on the migration of large enterprises in terms of their cloud adoption at a macro level. So I think that's probably important to understand just in terms of the underlying adoption of our products because we really are focused on having salespeople sell -- increase the penetration of MongoDB within their respective accounts regardless of how a customer wind up consuming MongoDB because they may all be at different points in their cloud journey.

Got it. And that's actually a good segue to my next question, which is that, obviously, we've seen unbelievable growth from Atlas. I mean, we've seen a little bit of a slowdown for on-prem. I mean, how do we think about the trajectory of that business? I mean, as you said, I mean, you're going to have some customers who just aren't ready, right? They're not ready for a migration to the cloud. They're going to be on-prem for a long time. You might even have net new customers who want to be on-prem. But curious how we think about the divergence of growth for each of those businesses and anything you can share about your expectations longer term for the business mix as we move forward here.

Mike, do you want to take them? I think Mike is on mute.

Yes, yes. I'm sorry. I just -- I think I'm unmuted now. Sorry about that. I just accidentally hit the mute button. So yes, I really think it goes back to this channel piece. And so if you think about the 2 channels as the direct and the self-service channel, the self-service channel is really almost entirely Atlas at this point. And so the adoption there is fairly straightforward. I think in the direct sales side, we've talked about how there are really kind of 2 main buckets within direct sales. You've got the mid-market team and the enterprise field team. In the mid-market team, that also is predominantly Atlas at this point. The product-market fit there is exceptionally strong. Those folks tend to be earlier adopters of technology, et cetera, et cetera. And so ultimately, the real product mix depends on your assumptions around enterprise client, enterprise customer adoption of the cloud, right? And so I don't think we'll be meaningfully different from there other than the sense of new workloads tend to be easier to put into a cloud setting, and people tend to hold off in terms of migrations, but that varies very much customer by customer.

Perfect. So one question on competition. Obviously, there's some private competitors in the NoSQL space and, I guess, certainly, the hyperscalers as well, but can you talk a bit about some of the reasons why you continue to win, I guess, against the competition, both from the large cloud providers and AAA, those SQL vendors?

Yes, Chris, we believe that we really win for 3 reasons. One, we believe that we have the best database of data persistent platform. And the reason for that is that it's -- the fundamental tenet of our architecture is the document model, which more easily maps to -- more naturally maps to the way developers think and code and has really proven to be the most productive way for developers to work with data. And so that's point number one. Point number two has been designed from day 1 to be a highly distributed and scalable database so that as your workloads grow and your [ performance ] needs to grow, MongoDB scales with you. And so we truly are the most performing platform out there. And then the second reason is that we have incredible developer mind share. There's been a lot of vendors who've tried to become in and say they're going to displace Oracle as the new database king, but they never really have the developer mind share. We have had -- we have enormous developer mind share. Just to put that number in perspective, our software has been downloaded from our website alone 130 million times in total, and it's been downloaded over 55 million times this year alone, which is more than in the first 10 years of the company's history. So that gives you a sense of how popular MongoDB has become. And we truly have a large developer community, and we've built that community over the last 10 years. And the third reason I would say is platform independence. There was a study that came out that said that 85% of enterprises are already using services from multiple cloud providers, and that number is expected to grow to 98% over the next 3 years. And the fact that MongoDB runs anywhere, both on-premise as well as the cloud and across all the major cloud platforms, and as I previously said, you can run an app across different cloud providers, it really gives comfort to organizations who know that they never will have to be worried about being locked in by using MongoDB. And so that allows -- that gives customers a lot of confidence, and that's the reason why we win.

Perfect. Actually a couple of other questions coming from the audience. One is sort of asking your view on the Oracle autonomous database and how Atlas would compare to that, and anything you could share about price in terms of differences there?

Yes. So Oracle basically has announced the autonomous database a couple of years ago. Frankly, I have not seen, and I don't know any of our sales team have seen them show up in any deal that we've competed on. And Oracle more recently announced that JSON, which is the underpinning of the document model, is the most modern way to build applications and they try to add some JSON capabilities. But the reality is that they, at the end of the day, are still a relational database where you have to stuff data and tables in rows and columns versus a document model where it's more natural to -- the way developers think and code. And so we feel very comfortable about our competitive position. And if our competition is now encouraging customers to think about JSON and the document model, we'll take that -- we'll have that conversation any day of the week. And in terms of pricing, we have never lost a deal to Oracle on price. I'll just leave it at that.

Perfect. That makes sense. I wanted to touch on, I guess, the relational side for a second. Obviously, with Snowflake, I know there's 2 very different things, right? SQL, your guys are NoSQL. But I guess you have Atlas Data Lake now. And I'm just curious if, at any point, we start to see the lines were a little bit between you and Snowflake or just curious how you think about that.

Yes. Let me just try and unpack a couple of things that you said. So one, when you think about the database market, in the old days, there was only one database. And you ran that -- you had a database for both running your transactions and your operations as well as running analytics to understand what's going on in your business. What people realize, because relational databases were single-node systems that you could quickly cap the capacity or really bring the application down to its needs by running too many queries. So this whole genre of data warehouses came to play where you would offload your operational data into the data warehouse and run queries to run -- get insights on your business without impacting user performance. And so we are an operational database. Snowflake is a data warehouse just so that your audience understands the distinction between the 2. And so that's very important. The second point is that our primary persona is the developer. So developers use operational databases to build apps. And the most complicated thing about building the application is thinking about how you store and retrieve data. And their persona or data warehouses is most used by business analysts and now maybe data scientists to run sophisticated analytics and trying to get insights to the business. And so we view these things as very complementary. Where there is some overlap is around like real-time analytics. Data warehouses are not really designed to be real time because by definition, they're batch oriented. You have to move data from your operational data store to data warehouse. And where we see analytics is like as people build applications to instrument the business, for example, like a financial app, giving market data or a gaming app where it has leaderboards in terms of what's happening or say, a supply chain application providing real-time inventory levels, et cetera, that's built into the application, and we're embedding more capabilities to allow people to instrument their business through their apps. And so there might be some overlap around the analytics space. And because MongoDB has a distributed database, we're well positioned to do real-time analytics because we can dedicate some -- I would say a 5-node cluster of MongoDB. 3 of those nodes could be dedicated for users and 2 of those nodes could be dedicated for running real-time reports. And that's where the overlap is. But we do not see Snowflake as a competitor. In fact, they're more going after the data warehouse market. Now we have announced this notion of a Data Lake. And what that is, is the fact that the cheapest place to store data is your object stores in the cloud, like something like S3. And because S3 is unstructured, it's -- we are well positioned because data warehouses like Snowflake need a structured schema. Because S3 is, by definition, unstructured, it allows you to create the data no matter what the data type and how that data is structured. So think of that as like going after the old Hadoop use case where people were trying to use Hadoop as a way to kind of get access to all this data they're putting in this Data Lake. The problem with Hadoop was that, one, it required very expensive skill sets like data scientists to use. It required tons of customization, so -- and it required a lot of time to value. While most people don't have a lot of money and don't have a lot of time, so that's why the Hadoop market never really took off. With MongoDB's approach, you already know the MongoDB query language. You're already using something that's already in place and you're only paying per use. So it's a variable cost structure. So it's a much more competitive and compelling offering. And this is an area that we'll continue to expand in as we go forward.

Perfect. So coming back to Atlas for a minute here. Can you talk a bit about what the net expansion rates look like for Atlas customers relative to your on-prem? I imagine they're a lot higher, obviously. But as customers expand consumption and even use cases, I'm just curious like if you could give us a sense for what that looks like.

Yes, sure. I think the easiest way to think about this, we go back to my earlier comments, and we tend not to think about it on a product basis but on a channel basis, right, because what you're really thinking about is how is a customer expanding. And therefore, you're really thinking about, okay, you acquire a direct customer through sales and how do they expand over time, that expansion can be an EA customer who's continuing to add EA workloads. It can be an Atlas-only customer. It can be an EA customer that adds Atlas workloads. And so when you're really trying to think about sort of the unit economics and the customer economics, just really what those net ARR expansion rates and things like that feed into, it's, at least from our perspective, more appropriate to think about it in the context of a customer rather than of a product. And so what you'll see is we've had very high, over 120-plus percent net ARR expansion rates in the direct sales channel consistently. What we've also talked about is if you look at the self-serve customers, right, which are sort of Atlas-only, they have even higher expansion rates. Obviously, the dollars are different, right? We're talking about self-serve customers, on average, about $6,000 a year and direct sales customers north of $100,000 a year. So very different customer profiles. And the self-serve customers tend to start off small, but their expansion rates of those cohorts are even more positive.

Makes sense. And one maybe more modeling-focused question. As Atlas continues to grow as a percentage of the mix, I mean, how do we think about that impacting your gross margins over time?

Yes. So Atlas is gross margin dilutive today. And I think the key thing to call out, and we've talked about this in a couple of different settings, is we -- that being said, we've executed quite well on our gross margin plan for Atlas. And so if you had told me that we'd be at, whatever, 44% of revenue in Atlas and 72% non-GAAP gross margin 3 years ago, I don't think we would have been able to achieve that. So we've certainly done very well. We shrunk the gap meaningfully in terms of how dilutive Atlas is. That said, it does include the underlying infrastructure cost that is a cost that we bear or incur related to the consumption. And so Atlas is and will be lower gross margin, but we've a nice job of shrinking that. We've still got more to go. Atlas is growing faster, fastest-growing product. And so we'll probably continue to have some slight margin degradation or erosion, but we're trying to sort of describe that visually as sort of a shallow U rather than something that just sort of falls of a cliff.

Sure. Makes sense. I wanted to read one more from the audience here. The question is, what are your top 3 use cases and how has that changed with COVID?

Yes. So one of the things that I want to really stress for everyone is that we are truly a general-purpose database or data platform. And so what that means is that we're not one of these like niche database that people only use us for a particular type of use case. So when you look at our customer base, and we have over 22,000 customers today, some of them are some of the most sophisticated, demanding customers can imagine both at the high end as well as cutting-edge start-ups. They're using us for a wide variety of use cases. They tend to be in 3 big buckets. They tend to be in systems of engagement where they're either front office or customer-facing applications. It could be revenue-generating applications like e-commerce platforms, content management systems, gaming, mobile applications, et cetera. The second category, what I call is system of record applications where you're the single source of truth, whether it's financial applications, it's supply chain applications, it's billing systems, et cetera, that people are using us for. And the third area, which is more nascent is around systems of insight, where people are building and instrumenting their app to be the -- basically giving people insight on what's happening in their business. And so in the early days of MongoDB, we tend to be more focused on systems engagement because that's where the highest rate of change was in terms of adding new features. Think about your mobile app. Every morning when you wake up, some mobile app has been updated over the night because of the rate and pace of development is so high. We were very well suited for those use cases. As people got more comfort around MongoDB, they started deploying us deeper and deeper into their enterprise. So as I said, billing systems, financial management, we have treasury risk management system -- treasury and risk management systems. A lot of banks are using us for trading platforms, wealth management systems, et cetera, that MongoDB is used for. And then you're seeing this convergence of real-time analytics and people using that as well. So I want to stress -- I know sometimes people view us as a NoSQL database, which in some people's mind is like a niche database. We truly are a general-purpose platform, and we're going in and people are using us either for new applications that truly run their business or replacing the existing relational systems with MongoDB.

Perfect. One other benefit of Atlas is you can see what your customers are doing, right? You get kind of data back from them, which is, of course, a huge benefit. So as you've had more and more of that data back -- come back to you, how has that informed the product road map? And can you talk a little bit about -- anything you can share about plans there and recent features you've added that have had a lot of success?

Yes, sure. So you're absolutely right. Because it's a cloud service, we can see what customers are doing. Obviously, we don't have access to the data, but we can see like avenues in usage, and we can get a sense of how will they -- one, a couple of things that we found is that sometimes customers don't always know how to best leverage MongoDB. So we can very quickly come in and help people better model their data because sometimes they come with a relational mindset. We can help them understand how to say, set up indexes when we see slow-running queries and saying like, you can get much better performance if you put an index in place. So we have tools like performance adviser that allows customers to get recommendations, what indexes to build. And then we can also help them with like optimizing their configuration for performance and scale depending on how quickly their application is growing. So that really helps customers. Sometimes you actually help them downsize their cost because sometimes they haven't modeled their data correctly. And what we don't want is them to get some surprise bill and then things like MongoDB is so expensive. So in some ways, we actually engender more trusting customers who are saying like, "Hey, not configured MongoDB correctly. You're overspending," and then they love that, and that's why they grow so quickly because then they realize they're going to get so much value over time. In terms of new capabilities that we've announced, one of the areas of new capabilities announced was Atlas Search. One of the things that we found with many customers is that they were co-mingling both our -- the OLTP database. Atlas would like to say something like search oriented, say, from Elastic. And people realize that the [ dual home ] data between those -- the 2 different [ cells ], we didn't make a lot of sense. And so with Atlas Search, they can kind of have one unified data set and be able to address all their search needs as well as having a general-purpose data platform. We introduced Realm and Realm allows people to have the same benefits of MongoDB but on the mobile client. And as apps become more and more mobile first, especially with the advent of 5G, you're going to see the sophistication of mobile apps goes to the next level, which then generates another problem, which is around synchronizing data from the client to the back end. And so we're building capabilities, very sophisticated synchronization capabilities that will become GA shortly, where people can sycn data from the clients on the back end and from the back end to the clients. So think of like field service workers who are doing a lot of work, updating data in the field would be a natural use case and those kinds of capabilities will be a natural place for MongoDB to expand into. And then, as you mentioned, Data Lake, and then there's another use case associated with data we call online archive. One of the challenges customers have had is if they archive data, it's very difficult to create that data in real time. And what we've allowed is as your Atlas implementation grows, you can easily archive your Atlas data in a way that -- to say, S3 but [ still create ] as it is still online. So online archive becomes a very powerful tool to really drive your cost structure down while still getting all the benefits of seeing your data in real time. So these are capabilities that we've built. We're seeing a lot of traction from customers and it's allowing us to address the broader sort of use cases for that.

Perfect. One more question from the audience. When it comes to competing with AWS or Azure, do you find them more competitive at the high end of the market or the low end? At what point do customers become more concerned about being leveraged to a single cloud provider?

Yes. So our win rates against Amazon and Microsoft -- because they do have -- offer clones of MongoDB. I think we've talked about this in the past where they basically try to emulate MongoDB with a relational database, trying to emulate the [ cadence of ] MongoDB. Obviously, there's some severe performance and feature trade-offs you have to deal with. So our win rates when we compete head-to-head are really high against them. Our issue is that because there's such large organizations that have such broad reach, there are places where they're winning business that we're not even participating in. So that's frankly the things that we worry about most is like how can we quickly expand our reach, how can we generate awareness. That's why the self-serve business becomes a really effective tool for us to enter markets where we meet on the same half feet on the street. And then our insights sale organization doesn't need to be in every geo. They can be in different theaters but still have a broader geo coverage. And so that's the way we're trying to expand our reach. But clearly, there are definitely areas and opportunities that we just don't see as a virtue of their brand and their reach.

And I guess to address that, I mean, from -- in terms of investments, is that a function of investing more in direct sales channel? Just curious how you think about investing behind growing awareness, not only with the self-serve tool, but just, in particular, among larger companies as well.

We as an organization have made a pretty concerted effort to try and expand our reach in terms of feet on the street around the world. We've talked about the fact that the challenge -- the blessing of the curse of this -- of having a big market is that because it's big, you can grow fast. But because it's big, it takes a long time to have the geo coverage you want. So sometimes you just need to show up in deals to win those deals, right? And so this is just an element of physics where if you're not in deals, you're not going to win certain deals. So that just takes time for us to do and we're operationally constrained. We can't add 1,000 salespeople overnight because you'll have to make trade-offs in terms of quality and training, and that's not something we want to do. So we are also investing in kind of go-to marketing efforts. We're using marketing tools and digital marketing to just generate awareness, help drive people to our self-serve channel, drive people to our conferences. Obviously, with COVID, they've been more virtual. But as the world opens up, we'll invest again heavily in in-person conferences because that's a great way for us to engage and educate the market and grow the MongoDB community. And if you look at any data by almost every objective measure, we are today the most popular modern database in the world, but that doesn't mean that we're standing pat. We want to continue to grow the community and continue to drive more awareness and recognition on MongoDB.

Perfect. And I guess on the partner channel, I mean for enterprise customers, I mean, I know there's a lot about building awareness here, right? And that's, of course, the case at the low end. But at the higher end, with presumably large organizations with very large IT organizations, how important is the partner channel to drive that awareness to get you in more deals where -- again, your win rates are very strong.

Right. So we have typically 3 types of partnerships. One is, I would say, the cloud providers themselves because obviously, everyone's running around their platforms, and we're both a consumer of their services as well as we work with them to win deals and drive awareness of Atlas. And I would like to remind everyone that they actually funded part of the development of getting Atlas on the platforms because you saw how a platform MongoDB is. The second group of partners is the system integrators. They end up being a natural complement to us because in many times, customers just don't have their own development capacity to move as fast as they want. So they hire system integrators to build apps. So those system integrators end up choosing MongoDB as the underlying part of the tech stack, and so they drive more business for us and obviously, they have broad reach. And the third area is ISVs. As ISV -- new ISVs build their apps, they have to choose a database. And so naturally, when -- the more times that you choose MongoDB, the better, best for us. And sometimes existing ISVs, say on-prems in ISV, I think that they have to replatform because their architecture has gone on so brittle that as we think about replatforming, we become an obvious choice for them to use. And when you step back, there's really 3 objectives that we're trying to accomplish with our partner channel. One is expanding our reach because by definition, they will give us -- bring us access into business that we may not actually see ourselves; two, build more credibility with customers when you have a large SI or cloud provider endorsing us, it really builds credibility with customers that MongoDB is the right choice. And three, close deals faster, right? Like sometimes you may use a partner's paper because they already have an existing agreement in place. Sometimes by virtue of their relationships inside the accounts, they can help deals accelerate because they know what's been budgeted and who the right decision-makers are. So for those 3 reasons is why we built out our partner channel, and we've had a lot of success with it so far.

Perfect. All right. Well, I think we're just about up on our time here. But I just wanted to thank you both so much for joining us today. I really appreciate it. Great to you, and we look forward to chatting soon.

Thank you, Christopher.

Thanks for having us.

All right. Good to see you. Thank you.

Take care.