Good morning. Thank you for joining us here. Had an insightful little keynote there with the CIO panel. Good news is, a lot of investment in data, so this is very topical here. Very pleased to have MongoDB. Today we have Michael Gordon, the Chief Operating Officer and CFO, and Serge, the SVP of Finance. Welcome to Nashville. Thanks for having us.
Thanks for having us.
Absolutely.
Great to be back here.
Listen, let's start maybe with this idea of we've seen a slowdown in the overall spend environment. It's been challenging. But as I take a step back and think about the role of a database, I like this idea that behind every good application lies a database. So let's talk a little bit about the business, the drivers of the business. What portion of ARR is driven by new apps versus migrations off older apps, and how would you compare that to kind of maybe what you saw in the 2021 IPO?
Yeah, so take a step back. First of all, great to be here. Thanks for having us. Unfortunately, there's even an app, a database behind the bad applications, not just the good ones. But, so it's a huge market. And, you know, IDC's numbers are north of $80 billion. J ust to try and kind of frame it up, it's a market that's actually also growing in double digits, which normally you'd say, like, databases have been around for decades. Like, how is a market that's been around that long, you know, growing at a, you know, well in excess of kind of GDP clip? Shouldn't a more mature business grow in line with GDP?
Fundamentally, it's really because databases are at the heart of every application, and you know, companies today are using software, specifically software that they build, to try and drive competitive advantage, right? To improve their end user experience, and things like that. Each of those applications needs a database. That's what creates the opportunity for us. So that's really, you know, kind of the core of the opportunity. We're, you know, closing in now on $2 billion of revenue, but still tiny, kinda, you know, low- single-digit share of this huge market. If you think about, you know, the market overall, to kind of your question, you can think about it in terms of new builds, and you can think about it existing.
I think the simplest way for me to think about it is, if you look at the IDC projections, there's anywhere between kinda $10 billion-$12 billion new being created every year, right? New spend every year, right, which again, just sort of highlights how big the opportunity is. And then there's, let's just for simplicity's sake, we'll call it kinda $80 billion in the installed base. Not every dollar of that $80 billion like issues an RFP every year, right? If you have an application that's working perfectly fine, you're not gonna bother to, you know, change out the database. Databases are extremely sticky.
So when you think about it, if we just, like, roughly assume that there's, you know, an average 10-year application life, which probably on balance is coming down, just given the rate and pace of technology innovation, but just for simplicity's sake, we'll call it 10 years. So that's about $8 billion of existing that's gonna re-platform every year. So you add that to the sort of $10 billion-$12 billion of new, and, you know, you wind up with sort of $18 billion to, you know, plus billion new kind of transactional in any period. And if you look at sort of our incremental market share, it's been increasing over time, but it's still kind of mid single digits. And part of that is just because of the footprint coverage.
Yep.
We can talk more about that. In terms of the relational migration piece, at the time of the IPO back in 2017, we said about 30% of the new business that we were winning were relational migrations. That has continued to go up in dollars, but down in percentage. I think our last disclosure on that was about 20%. It varies between 20% and 30% over our time as a public company. And part of that is because Atlas, which is now 71% of our business, which is our database as a service offering, tends to be new builds, new applications.
Also, if you went to, like, a big bank, or a big telco or someone like that, by the time they're taking an existing application, re-platforming it, moving it to the cloud from, you know, on-prem, everything else, at some point, you could probably get at least half the people in the bank to call that a new application, so like, the distinction gets a little bit fuzzy at some point, so like, I wouldn't like overly push it, but that's kind of the setup, so why don't I stop there, and we can go anywhere you wanna go?
So the way I look at it, $18 billion-$20 billion dollar jump ball-
Yep
Every year across new and migrations that you can drive net new ARR every year. How do you increase the attach rate to the new part? Obviously, the mix has gone up since the IPO, but what are the things that you can drive a higher attach rate to net new application builds?
Yeah. I think there are two kind of key things. Like, first and foremost is sort of the, you know, product, right? And we have an amazing product, and, you know, relational databases, which are the vast majority of the spend today, you know, so were built and pioneered in the seventies for a completely different era, and so we've come along, you know, with not just what we think, but developers generally view as kind of a better mousetrap, right, and so having a great product is the first place to start. Obviously, that's not a stagnant effort. We have to continue to invest in the product and have had a lot of success there. Second is really developer mindshare, and so developers love working with MongoDB.
And then the third thing is footprint coverage for sales. And so I think this is an important distinction, especially for investors who are newer to the category. There aren't a ton of database companies to invest in. And so when we, you know, went public, we were the first, you know, database company to go public in over a quarter of a century. So there aren't a ton of comps, right? It's not like cyber or something where, like, there are, you know, security, where there are, like, a million companies out there. And so one of the key distinctions of our market is companies tend to adopt or pick databases application by application, right? Kind of going back to your opening, you know, comment.
And so when Piper, you know, decides to run a you know, an application on MongoDB, they don't, like, kick out every other instance of Oracle, and like everyone else, it's sort of like an application-by-application process. And so that changes the dynamic a little bit, you know, unlike, ERP or a CRM, right? Like, you know, you the sales and trading team's on the same, you know, HR system, as the research team or the banking team or the private wealth team or, you know, whatever, right? And so that's just sort of different, than what you see in a lot of software. And so the biggest challenge is the awareness, right? What we're offering is a paradigm shift, and some people are still steeped in relational.
And so it's the awareness, the education, and then the selling. As popular as we are, you know, our salespeople don't get to, you know, kick back, put their feet up on the couch. You know, once, you know, we sell a workload into account, we need to continue to sell additional workloads as you go, development team by development team and division by division.
The only other thing I would add, but importantly, when we are in the room, our win rates are very strong.
Yeah, for sure.
That kind of gives us one of the key reasons why we have confidence to continue investing.
So let's talk a little bit about the growth profile of the business. Obviously, really big opportunity out there, to drive a higher attach rate, high win rates. But if I look at the overall kind of market, we've seen, industry growth rates across the cloud slow. We had kind of two decades of 30% growth, got spoiled. But now we're looking at an industry that's gonna grow about 13% across the whole Cloud 100, if we look at the industry. My question for you is: How do you think about your growth profile here? It has slowed. What are the levers you're thinking about that could maybe sustain growth and/or maybe start to reaccelerate growth? Serge, we'll start with you.
Yeah. I'll quickly hit it. Obviously, you can add anything you want to. But I think about growth in two different ways, right? I think about it in terms of there's an existing set of applications or workloads that we've already won, and then there's the new business that we win, right? And we talked about this at our Investor Day, but that set of applications that we've already won, when a workload starts out, it tends to start slow, it tends to grow quickly for the first couple of years, will tend to grow, but at a you know slower rate over time. And so if you won no new workloads, just the organic kind of you know growth rate of that you know installed base would slow over time, right?
And so the first thing, when you think about what does it mean in terms of driving, you know, higher growth rates, is you need to keep adding new, not because what you acquired isn't attractive or compelling, but just, like, naturally, it's not gonna keep growing at as fast a pace as it does in the first couple of years of an application's life. And so the latter part, this kind of the new business piece, is really a function of the size of the sales force, you know, our footprint coverage, sales productivity, and all those kinds of things. And so that's sort of the combination.
The last thing I'd say about the installed base is that growth in the installed base, while it's affected by the kind of tenure or whatever you wanna call it, of the applications in the mix, is also very closely tied to the end user activity, of those applications. So if you think you have an application, could be a customer-facing application, could be a consumer-facing application, could be an internal application, you know, whatever it might be, the ways in which the end users interact with that database, right? The underlying reads and writes that you see, the transactions that they create in the database, will drive the underlying consumption, of that database. And so some of those things, as we've called out at various points in time, are affected, by the macroeconomic environment. We kind of see that activity.
Yeah, the only, the only other thing I would add is, we've been very happy with our ability to win new business sort of across the board. Q1 was a little bit of an outlier. We can talk about that, but generally speaking, new business. What's been the challenge and sort of the constraint in our growth over the last two years is that we've sort of seen two phases of the slowdown in the growth of the underlying base. The first was roughly two years ago, when, you know, the interest rate environment changed, the investment environment changed. We saw usage growth across the portfolio slow down, and with it, as Michael was saying, the consumption growth slow down, and then again, in Q1 of this year, not quite to the same extent, but still we saw usage year-over-year growth slow down.
And, again, it feels like a slightly softer macro environment again. And so that's the that's been the increased headwind that we've been fighting, 'cause there's nothing we can really do about the growth of the existing applications. Where we spend most of our time is really all of our time around how do we win even more new workloads to offset that trend.
As you think about the levers on the sales force side, go-to-market, is it just a more challenging environment? Is sales productivity a little more challenging given this environment, or are you seeing some good productivity numbers and-
Yeah, we haven't seen the challenges that Serge referenced. We had some internal operational issues in Q1, but beyond that, like, we've been able to successfully sell, you know, really in all environments, even in a more challenging macro environment. And part of the way that I sort of think about it or sort of mentally visualize it, if you think about sort of like an ocean, there can be all sorts of turbulence and chop and froth at the top. But when you're a small, you know, share player, you know, down at the bottom, some of those headwinds can exist, but you can still, you know, as long as you're executing well, continue to win, despite that chop, just because, you know, you're an insurgent, right?
Who's gaining share, and you know, you don't have that giant installed base that the legacy players do. And so I think that's been important. We also talked about, to your question about sales productivity, on our last earnings call a couple weeks ago, about how productivity was up year over year.
Okay.
And so, you know, I think the team has done a good job and has been able to execute well, despite, you know, generally a more challenging, you know, macro environment. But fundamentally, you know, we think there's a big opportunity, you know, as a small share player, as long as we execute well, we should be able to continue to win new business.
Serge, question for you on Atlas.
Mm-hmm.
This is a growth engine for the company. It's been a fantastic growth engine. It's now a $1.3 billion business for you. Is it just usage that you kind of need to reverse here in consumption or reverse to drive some stabilization and/or re-acceleration there? Are there other, I don't know, product factors that could help drive an acceleration on the Atlas side? Walk us through, like, the net levers as you think about that business.
Yep.
It's been the key growth engine.
Mm-hmm.
How does that growth engine sustain and/or improve?
Yeah, so there's two things of which the first one is usage. Usage has, we've seen a slowdown that we think is largely macro-driven. We don't control that, but certainly there are states of the world in which you can assume that it can get better, but we can't plan for that, we can't bet on that. It just will be what it is. The one that we focus on is acquisition of more new workloads. Product does play a role, so things like Search or Vector Search or some of the other things in our portfolio mean that we are winning more workloads than we otherwise would have, because those incremental features might be the difference between us and somebody else.
But there's many other things that sort of sit in the bucket of, like, how do we acquire new better, which is, you know, app modernization. We're getting increasingly excited about the opportunity to use AI to get more of those legacy relational workloads on top of our platform. AI apps themselves, we've launched a program called MAAP, to sort of build an ecosystem that makes it easier for customers to start building applications. And then a number of what I would call more operational tactics in terms of how we configure our sales force, where we deploy our people, marketing tactics, all of which are there to kind of increase our ability to win new workloads, to supplement whatever happens on the usage side.
I tried to hold the AI questions. We have this-
You did a stellar job.
I did, right?
Yeah, amazing.
You know, I tried to save it for the 20-minute mark. We're not quite there yet, so but let's dive into AI. And one of the things that I think, just intuitively, as you think about a database storing transactional records-
Mm-hmm
-to going to now storing conversations. Every question you ask the database-
Mm-hmm
You're gonna be storing. Those early customers-
Mm-hmm
-that are now kind of deploying, basically, AI. What do you see? Is the usage, consumption pattern, is there anything you can kind of give us that is a little bit of a tea leaf relative to what could happen as more of these AI apps-
Yeah
-Are built and run on Mongo?
Yeah. We see primarily two types of apps being built.
Okay.
One is what I would call the assistance and automation, and those are kind of bots and/or conversational assistance, things that take off sort of human effort, low-level human effort, and relatively easily replace it with AI. And the second is what I will call analysis more broadly, and that's, you know, intelligent document retrieval, helps you find the answer faster. That's where some of the, you know, RAG architectures in particular-
Mm-hmm
-become helpful. And that's where we're seeing both. Those are the two major flavors. And then there's a bunch of other interesting stuff that's on the margin, but if you think where the numbers are, those are the two first buckets of applications that we're seeing. We get this question a lot, which is: What's the usage intensity of an AI app gonna be?
Yeah.
Can you tell me what's the multiplier of an AI app versus a regular app? And I'll maybe offer two thoughts on that. One, for sure, it's too early. So, we most of the activity that we're seeing right now, frankly, is on the startup side. And they're, you know, these are exciting new products, but they're still looking for their, you know, product-market fit, and ultimately, we don't see a mature application that we can go benchmark against other applications. But then maybe in a more speculative bucket, that may not be the most important thing. The most important thing, ultimately, we always think, is the popularity of the underlying application. And maybe to make a parallel, we have a number of video games being built on our platform.
Mm-hmm.
But they're not all equally popular. One of the most popular video games in the world is built on our platform, and the ARR that we win from that is multiples larger than the average video game ARR on our platform. And that's not a function of the fact that that particular game uses the database differently, it's just the fact that everybody's kids, including my own, play it. So some apps will be more popular than others, and that will ultimately be the bigger driver of ARR that we get from them than the usage intensity. But obviously, we're watching the usage intensity, and when we have something that we think is, like, actually statistically significant, we'll definitely share.
Maybe it's helpful just, you know, given that we, you know, approached the AI topic, just to take a step back and think about, like, what are the kind of the key ways that it affects us or how we think about it.
Mm-hmm.
And I think there are really, like, three key buckets. I think the first is everyone talks about code assist tools, right? And we can debate how much developer productivity that's gonna drive, but it will make developers more productive. As a result, developers will create more applications to the... You know, opening comment at the beginning, every application needs a database, and so that just creates even more opportunity, right? So while we're not lacking for TAM, the overall market size should grow even bigger, right?
Yeah.
As a result of Code Assist tools, so that's kind of like a market size comment. The other two pieces are really more about market share, so the second thing I'd say is, in general, those modern applications, you know, or the AI-powered applications or whatever you wanna call them, will desire, will benefit from, will need a more modern database, and MongoDB is a very good fit for that. And so the result of that is, you know, we have the opportunity for kind of more share, as a result of, you know, our document model and our kind of, you know, better suited, kind of technological differentiation, for the modern world, and so that's helpful from, like, a market share standpoint.
And then the third piece is this large, you know, installed base of legacy relational applications were early on, but we're finding ways to leverage generative AI to help make migrations of that, more efficient, faster, cost-effective, et cetera. There's a fair amount of interest from customers, you know, around that. It's still quite early on, but incrementally, that also sort of creates another share opportunity for us when you kind of think about the overall market size. So I think those are kind of like the three big-picture, you know, kind of trends, or ways that AI has the potential to benefit us.
One of the things that really distinguishes Mongo from others in this kind of space, early on in the NoSQL kind of space, was developer mindshare. You've built a great reputation with developers. Just intuitively, as we think about ten years from now, when machines are writing the code-
Mm-hmm.
How do you create affinity with a machine? You've created affinity with the developer, how do you make sure that when machines create apps, that Mongo is the preferred app? Is it, like exposing documentation to these training models? I'm just trying to think through, like, the future set, where machines create the apps and not humans.
Yeah, maybe-
How do you drive affinity?
Yeah, let me maybe say three things. First one is, we think that for foreseeable future, including ten years, a human will be involved. And the early evidence from some of the Copilot seems to suggest not so much that the low productivity developer becomes more productive. It's more that the high productivity developer becomes even more productive. So those people are still gonna be there, still gonna be exceptionally important. Maybe there's a future, you know, super far enough in the future where you can imagine a world where that doesn't exist, but that would sort of be, like, the first thing that we would say. Second of all, it's absolutely important for the machines to know how to code a MongoDB.
So we have partnerships with all the hyperscalers and all their copilot products. We give them our documentation. We are invested in making sure that their technology works. But there's a third thing I would say, which is, if you broadly divide our competitive advantage, I would put into two bucket. The first one is developer affinity and/or ease of use for the developer. So this is the idea that our product is intuitive to use, doesn't require translations into rows and columns, and that's some of the reasons why it became as popular with developers as it is. The second, though, is that our databases operates better. It is more scalable, it is more performant, it offers lower latency, and over time, you'll see with incremental features, we'll offer, you know, the unique security that nobody else has.
So even if you sort of imagine some world in which, like, the developer no longer matters, and therefore, you know, it's kinda like, the first element of our competitive advantages is diminished, we find it hard to see that world over any conceivable time horizon. But even if you're interested in that, the second thing still remains and only grows more important over time, 'cause all of us expect more from apps in our lives, and that requires a better database.
Yeah, I think about it, like, simplistically, and the developer affinity is really a result of having a better product and developers being rational about that. And sure, we have great, you know, T-shirt swag or whatever, but, like, that's not why they're picking MongoDB. Everyone likes swag. Everyone likes swag, right, exactly.
We have great swag.
But that, you know, the bots or the machines, whatever you wanna call them, you know, will be at least as rational, and we ought to do well with that. We'll end on the toughest question, which has varied from year- to- year, or you call competition the big bear thesis and concern-
Mm.
-or with the cloud natives, right?
Mm-hmm.
Two years ago. Open source, at least with the investor community, has really risen to the top of, like, that concern, right?
Mm.
Oh, is Mongo gonna lose share to the open-source community, Postgres?" As you think about competition, have you seen any sort of resurgence in interest in open source or not, relative to competitive environment?
So the answer is no. We have really not seen any change. And so, I'll address Postgres and sort of how we see the competition, but what appears to us that has happened is, like, we had a tough Q1.
Yeah.
And we weren't very happy about that, and we lowered our guide, and we sort of offered our sort of list of explanations as to why, but investors, it's their job, ultimately, to, you know-
Just compete
-to pick it apart and come up with.
Yeah.
Suddenly, Postgres has become this thesis.
Yeah.
Postgres has been around for 30 years. Nothing has changed in terms of the way we compete with them. Postgres, in its many forms, is a principal competitor of ours, just to be clear about that, but that's been the case for years. What I mean by that? Postgres is not a company, it's a technology. The way it's actually primarily monetized is through the cloud provider solutions. We've been telling you guys for years that cloud providers are our greatest partners, but also our greatest competition, and not just their, you know, NoSQL offerings, but also their relational offerings, which are all really variants of Postgres. Postgres benefits from being cheaper and slightly easier relational, but doesn't fundamentally solve the structural challenges of relational, which go about sort of like the developer productivity and difficulty of scaling and performance.
We have great win rates when we actually compete against them, but they do have this built-in tailwind of lift and shift, right? So, like, if I'm a company, I'm leaving to go to the cloud, if I one day in the process leave my legacy relational, expensive, proprietary databases behind.
Cheap alternative.
It's a cheap alternative, so it isn't surprising that they're gaining in popularity, but it's not at our expense and when we do compete with them-
Yeah
We like what we see.
Yeah, and we're just having the open source piece, obviously, that's at our roots, and so we, you know, big, big fans of open source and have benefited from that a lot ourselves. Out of time here. Thank you so much for.
Thanks for having us.
Thanks for joining us. Yeah. Appreciate it. Thank you.