Good morning, everyone. Welcome to day four of the Morgan Stanley TMT conference. Last day of the conference, we're gonna end it strong. We're super happy to have the management team from MongoDB join us. CEO Dev Ittycheria, and CFO Michael Gordon. Dev and Michael, thank you for joining us.
Thanks for having us.
Great to be here.
Before we get into the discussion, let me go through some disclosures. For important disclosures, please see the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. To kick off the conversation, I was looking at my model. You know, we were fortunate enough to lead the IPO way back in 2017. The revenue from fiscal year 2017 to the year that you just completed has increased by more than 11x. Incredible success. One of my favorite, you know, themes when I talk about MongoDB to clients is that you guys have been one of the most stress-tested companies in the markets in terms of, you know, cloud, competing with hyperscalers, partnering with hyperscalers.
It's been a very impressive track record. We did this year see some slowness. I wanna like recap some of the themes from the earnings call last night. In particular, you guys called out some consumption weakness related to Atlas. Maybe either Dev or Michael, how was Q4 different than what you saw in Q2 and Q3? Were there any themes across market segments or geos or verticals that drove or contributed to the weakness?
You wanna take this? You wanna start summarizing the quarter?
Yeah. Maybe I'll just start with the quarter. You know, we had a strong quarter. We delivered 50% year-over-year Atlas growth, 36% growth of the overall business. We delivered 10% operating margins. In general, our new business activity was also very strong. Retention rates were incredibly strong. Overall we, you know, we feel like we had a good quarter. What we did call out was that in December and January, we did see a slowdown. We had expected a slowdown, but it lasted longer than we had expected.
What we talked about is that, you know, when we see the February data, the consumption seems to be trending back to what we saw in Q2, the average of, you know, post Q1 slowdown, pretty much what we saw the last three quarters of last year. That was essentially the kind of the highlights. I don't know if you wanna give any more color.
Yeah. No, I think the only other thing is, as we said in the December call, we did expect December and January to be slower. As Dev mentioned, that just was more pronounced. It was across industries and geographies. Obviously the holiday season is a relatively global phenomenon. There isn't a lot of particular, you know, interesting undercurrents or differences or divergences.
Makes sense. Then in terms of Q1, it calls for a flat to sequential decline in Atlas as well as in the Enterprise Advanced. Can you just sort of walk us through why that's the case going into Q1, Michael?
A few things just for folks. On Atlas, there are really three things that affect the sequential view of Atlas from Q4 relative to Q1. First, which we've talked about and folks who follow the company well or closely know, is that there are fewer days in Q1. That's about a 3% headwind right there. Secondly, in the call last night, we talked about several million dollars of unused commitments that we saw in Q4. This is the cumulative effect of the slower consumption that we saw earlier in the year starting in Q2, so Q2, Q3. That's revenue that we typically would have seen from consumption there. That creates a several million dollar headwind as you look at the sequential from Q4 to Q1.
Lastly, when you think about the slower growth that we saw, it means that the starting ARR of Q1 is lower than it would have been, that presents a headwind. That's why we're looking for Atlas to be flat to slightly down. That's atypical. That's sort of not the normal trend, wanted to give you the piece parts to help understand that. On Enterprise Advanced, that's a more typical dynamic. Q4 is our strongest selling quarter, that Enterprise Advanced has the term license component under ASC 606. We typically expect to see declines from Q4 to Q1 sequentially. This year is no different in that. The renewal base is much smaller.
We tend to not sell a lot of new customers in Enterprise Advanced. It tends to really be more of an upsell motion, and the renewal base is smallest in Q1.
Makes a lot of sense. Then, Dev, for you, in this sort of weaker environment, can you just sort of walk us through your thinking on how the team is going to operate or invest from a strategic perspective? What areas are you sort of pulling back, and what areas you're sort of doubling down to put the company to position to come out stronger out of this downturn?
Yeah. We've turned that inside the company as raising the bar, and that really means is, essentially having a higher bar for new investments, a higher bar for performance, and a high bar generally in expectations, how we run the business. From a R&D point of view, you know, we're very committed to our roadmap. We are evolving from a database company to true developer data platform, and that's really resonating with customers, especially as customers wanna consolidate vendors in this environment. They love the fact that they can consolidate more and more workloads. I would argue that we've essentially won the NoSQL market. Rather than using a bunch of bespoke point solutions, customers really wanna kind of move to one platform with one elegant interface and where all the data sits. That's one.
On the go-to-market side, we obviously assess the different channels we have. We have a high-end enterprise channel, we have a mid-market channel, and we have a self-service business, so we are constantly assessing the effectiveness of those channels, and especially even teams, subteams in each channel. In areas that we're seeing good results, we're doubling down. In areas that we're maybe wanna see better performance, we're pulling back. We're orienting, and we talked about this on the call yesterday, really focusing on acquiring new workloads, either through the acquisition of new customers or the expansion of existing accounts. Then on a people side, obviously we've grown very quickly.
We recognize that, you know, not every person was a perfect hire, and so we're, you know, assessing and looking at our team and asking our team to raise the bar. Now that's not necessarily a cost reduction exercise, it's more of a raising the bar exercise. We are also slowing down growth in coming into this year, given, you know, what we see in the macro environment.
You know, a big topic for investors, over the last year, and certainly over the last several months, particularly as we see the results coming out of the hyperscalers is, you know, what's the sort of state of cloud and cloud investments? When you look at some of the key elements that drive growth for MongoDB, new workloads, new applications, re-platform or modernizing existing applications. From your perspective, Dev, what do you see in terms of cloud transformation projects, digital transformation projects, cloud migrations? Are those slowing down materially? How do you see the, you know, priority of this category spend, which has been really strong over the last couple of years?
Yeah. What I wanna just make clear is our slowdown in revenue is really a slowdown of existing workloads growing slower than we have seen in the past. Our new business is still strong. We still see customers building lots of new applications. Frankly, you know, the talk and excitement about, you know, generative AI is spawning a whole new genre of applications. We, you know, we don't see the level of software development slowing down. In fact, you know, I would argue that almost every company's business strategy is expressed through the software products and services that they build, in terms of new products or how they wanna run their business. That's super important. There is some talk about, like, seeing the hyperscalers and their slowdown. I think you have to remember, hyperscalers offer lots of services.
Mm-hmm.
Not all those services are mission critical. you know, for us, you know, people do view, you know, MongoDB as a mission-critical platform. Many customers told us, "If we're turning you off, we're shutting off the lights and closing the doors," because we're like the backbone of their business. we feel like given also the next genre of technologies, that software development's gonna be even more important, and developers are gonna drive that agenda.
Yeah. You mentioned AI, a later question, let's tackle it right now. There's obviously a big excitement around large language models, generative AI. I think MongoDB, as a company that's focused on developers, the acceleration of code development with these types of technologies. Walk us through what you think, you know, some of the knock-on effects for MongoDB could be if developer productivity is gonna materially increase. Any other comments you may have on how AI impacts MongoDB's growth equation.
Yeah. I think it affects our business in three ways. One, as I said, you're gonna see this boom of new software applications. I mean, if you talk to any venture capitalists out there, they're getting flooded with people, you know, with a bunch of AI companies trying to raise money. Obviously, every enterprise company is also, like, trying to leverage generative AI. OpenAI just lowered their API costs, so they're really democratizing the access of their large language model. You're seeing that also happen with Stability AI and people like Hugging Face. I think that's just gonna spawn a whole genre of applications, and developers are gonna be the center of that.
The second thing is I think you're gonna see is the developer productivity dramatically increase through things like code generation, you know, automation of test cases, automation of documentation, automation of comparing algorithms to see which one works, and a whole bunch of other ways that developers can increase productivity. That's gonna mean that the velocity of new application development's only gonna increase, which is gonna be a help for us. The third area I see is that obviously, with the large language models like OpenAI, everyone has access to that data.
When companies say, "I have this corpus of internal data that I have, if I can use that data to train my models to give me a competitive advantage over others, then I need to be able to access that data." You're gonna see a lot of people start modernizing their legacy stacks because a lot of the data is trapped in legacy platforms, and it's not so easy to use. I think that's another advantage for us. We view AI as a real accelerant for our business.
as an operational database, Mongo owns a lot of that, stream-.
Yeah. I should also.
For.
I mean, said another way, for the first category, view us as the picks and shovels of the gold rush. Like, people are, you know, a lot of AI companies are building on top of MongoDB. They're doing the transformer work, they're doing all the, you know, training of the models, but they're using our, you know, massively scalable data persistent platform to essentially build those next generation applications and companies.
A long-running theme for MongoDB has been the ability to the document model to address more and more use cases and more and more workloads. I remember at last year's Analyst Day, I think you called out time series, graph, and search as additional workloads that can be subsumed or consolidated onto the MongoDB platform. What's sort of the state of play when it comes to those initiatives? Is there anything that the team needs to do on the product side to drive more of that standardization and consolidation of those workloads?
Yeah. You started this session by framing that we've, you know, we've been battle tested or battle scarred, right, in terms of questions. The first big question when I became CEO is, can MongoDB really run mission-critical applications? We answered that question. The second question was, can you really be a general purpose platform? We answered that question. Just look at the size of our customer base, look at the variety of use cases, you know, across every industry, every geography. The more recent question, you know, when we went public was, can you really build a cloud business and go head-to-head with the hyperscalers?
Obviously, we proved that in spades given that Atlas is now 2/3 of our revenue. Our platform strategy is based on the fact that the document model is truly a general-purpose way to run a wide variety of use cases. It can support key value stores, it can support ACID transactional, you know, transactional intensive use cases. We've built out search, so people don't need to have a bespoke, Elasticsearch or Solr implementation. They can run everything on MongoDB. Time series is getting a lot of attention, and we're seeing a lot of demand from customers to run time series workloads on us. You know, we also have mobile capabilities and doing very sophisticated enterprise mobility use cases with Atlas Device Sync. The list goes on and on and, you know, graphs.
I believe that all these kind of point solutions are gonna really struggle, especially in an era when customers wanna consolidate vendors and they see a way to kinda do that with MongoDB.
Is there any sort of examples of that coming out of the Q4 call about, you know, enterprises moving towards that type of motion of standardization and consolidation? I think there was.
Yeah. There's a number of use cases. We called out a bunch of use cases where customers were basically consolidating, you know, vendors. You had to think about it, the tax of... You know, if you try and use a net new technology for every net new use case, you could end up with 10, 15 different bespoke, you know, database engines, right? How do you learn all that? How do you integrate all that? How do you synchronize the data? How do you even back up all that data? The cost and the tax of managing all those complex bespoke solutions just becomes overwhelming. When you look at most of the data's gravitating to JSON, developers are the one driving the agenda. People want a highly performant and massively scalable platform.
They want a platform that people already know and recognize, and so that really positions us well.
Yeah. I think when you look at, like, the database lineup or the data management lineup from the hyperscalers, it is that 15 database lineup, which is, you know.
Yeah. I don't know if any customer wants to use 15 different databases in the enterprise.
It makes total sense. Let's talk about Just sort of sticking on product and in this era of focus on ROI and cost, the relational displacement opportunity. I remember at the time of the IPO, Michael, you sort of called out, you know, 25%-30% of net new business was driven by relational displacements. Last year, you guys came out with Relational Migrator. To what degree has relational migrations been a driver of growth in recent years, and how big of a focus can this be going forward?
It continues to be a big opportunity. If we just take a step back, right, the market per IDC is $84 billion in 2022, growing to $138 billion in 2026. Our market is a little bit different. Not every dollar of that $84 billion is doing an RFP every day, right? You have an application that's working fine, but there's about $12 billion of growth in the market every year. We're exceptionally well-positioned to go tackle that as the leading modern general-purpose database. There's also this sort of $84 billion that will, given You know, we can have a debate around application life cycles, but, you know, every year there's an opportunity to compete for a portion of those, and that's where you see people modernizing, right?
The opportunity set there is to move from the legacy relational technology to a modern scalable database that addresses a bunch of the things that Dev was talking about earlier, not just AI, but a whole range of use cases. That really is the core of the opportunity. We tend to see that most in the Enterprise Advanced product that continues to run at about a quarter of the Enterprise Advanced business is relational displacement. Obviously as our business grow, the dollars that we are displacing are increasing, and that continues to be an important part of the overall opportunity. Specifically on Relational Migrator, as we announced at our MongoDB World event last year, that's a sort of first phase. It's not meant to be a self-serve tool.
Relational migrations are complicated, part of what Relational Migrator does so successfully is it helps demystify the document model, right? If you were trained 20 years ago, as a developer, SQL was the standard. That was the only way to do things. It was inconceivable that a document model could deliver asset support, right? Helping people understand and do sort of the schema visualization, to see how would this work in a document model, helps sort of debunk some of the concerns, it's helpful in that way. We're obviously continuing to invest in that.
Right now it's field facing for our internal teams to work with customers, over time we'll continue to invest in that 'cause it is a big part of the opportunity.
You know, we mentioned the hyperscalers a couple minutes ago. Wanna get a status update on the sort of balance between them as a competitor and them as a partner. Particularly, I think, you know, one of the debates out there is, in a slower growth environment, do the hyperscalers get more aggressive with the cloud infrastructure ecosystem? Dev, maybe we'll give us your take on what the sort of status of your relationships with the hyperscalers.
Yeah. I would tell you that our relationships with all the three U.S. hyperscalers have never been better. I would say people may remember, you know, the year after we went public, you know, Amazon introduced their clone version of MongoDB. One of the things that's important for investors to realize, unlike most other open source companies, we have a very restrictive license that limits what people can do with that, and what one thing they can't do is take our free version and go compete with us, which they have done for other businesses. They essentially had to imitate, you know, our features and functions, but they actually built it on a relational back end. Our CTO, Mark Porter, actually was part of that development effort, we know.
We have a lot of inside information on, or knowledge I should say, on what happened. With that approach of building on a relational back end, there was massive feature and performance trade-offs. Obviously it's a big market. They saw so much MongoDB running in their clouds. They said, okay, they wanna go after that business, which would be natural. What they realized over time, one, is that they could not compete head-to-head, and if they ended up losing that workload to another cloud, all that business that also was generated around that workload would go with it.
Internally started realizing for every dollar we saw, they saw multiples of that dollar in terms of total spend with the underlying storage and compute and all the other ancillary services a customer would buy. Long story short, they started recognizing, why are we competing with MongoDB? You know, it's a truly a win-win relationship. Now, we're not naive. They still have their clones, and, you know, we have, you know, our sales people are trained on how to go compete against those clones, and we have enough reference points where people have switched from the clones to MongoDB.
Mm-hmm.
The same situation happened with Azure, where Azure had their own clones. They had an API for Cosmos DB. We have over time basically got them to see the fact that they just can't compete head to head. We just recently inked a deal where they're now incentivizing their sales teams. They're doing product integrations. They're even incentivizing their customers, you know, if they use Atlas, they can draw down against their commitments. It's a much more, you know, a friendly arrangement. Google has always been friendly. They don't have a competitive offering to us. That's been really good from the get-go. I speak to either I or our team speaks to the senior leadership across those three organizations, you know, frequently, if not, you know, weekly.
That's a great update. In the database market, it's been siloed for the last couple of decades between the world of analytics and the world of transaction interactions where you get more of your business. In the past, there's been movements to bring these two worlds together to have this sort of unified platform. One of these, it seems like there's another attempt at that going forward. You see companies like Snowflake try and move into the operational OLTP database market. Do you see convergence happening between analytics and/or data warehousing and the operational database and how do you feel like MongoDB is positioned for that convergence?
Yeah. It's important to understand, there's two worlds. There's the OLTP, online transaction processing, and OLAP, online analytical workloads. Those are the two different worlds. They serve two different personas. We serve the developer persona. For our business, it's all about getting developers to use MongoDB to build the applications that run and transform their own businesses. OLAP workloads are all about getting insights in terms of what's happening in my business based on the, you know, the data that they're collecting. They're two very different worlds. You also have to remember, if you look at a DB-Engines, there's more database companies available in the market than there are days in the year, right? It's kind of, you know, I guess a large TAM attracts a lot of people.
There's only really five to seven companies that matter.
Mm-hmm.
I know there's some, you know, companies out there claiming that they're gonna try and do it all. I would tell you, a lot of VCs come to me about, you know, investing in database companies. Database companies require a lot of capital and a lot of time. They require a lot of capital because this is not some trivial product you're building. You need, you know, to spend a lot of engineering man-hours to basically build the features that customers want, and it takes time to do that. The second thing, it takes a lot of time to get developer mind share. One of the questions on MongoDB that investors had is: Can we really be, you know, really successful?
When I looked at the developer mind share that we had, and even today, you know, our software's been downloaded over tens of hundreds of millions of times. You know, we have 40,000 customers and millions of users running one. That's the reason we're winning. It's not, you know, because we have developers who really have preference of MongoDB over other alternatives. So, I think it's gonna be very hard for an OLAP vendor to move into the OLTP space because you have to get the developer mind share. The place that we think we are well-positioned for analytics is real-time analytics because your OLTP engine is the place that's generating the data. What products have I sold? What customers are buying? What, you know, what, what stocks are trading?
You know, how much am I billing this customer, et cetera? That data by definition is real time, so then you can use that real-time information to basically speed your time to decision-making and speed your time to insight. That gives us a, you know, very well position. Because we're a highly scalable performant platform, we can do that in a way without compromising the end user's experience. That's where we're focused.
Makes total sense. In this higher cost of capital environment that we're in, wanted to talk a little bit with you, Michael, on the topic of margins, and maybe sort of frame it. I mean, clearly the company's been investing for growth. Mentioned at the top of the call, you've 10x revenue in the last five to six years, which has been fantastic. I think next year you're guiding for sort of 5% operating margins. If we step back and look at like sort of the unit economics of the business and maybe at kind of your larger, more mature customers, what sort of margins are you seeing on them on your sort of renewal and expansion business?
In light of the fact that you have companies that are at scale, you know, at sort of 40% operating margins today, is there anything structurally that prevents, MongoDB, over the long term reaching those types of terminal margins?
A few thoughts. First, just back to the IPO, you know, which you brought up. You know, we had long-term target margins of 20%+ there. At the time, we were -36% margins, so we've made about 40 percentage points of progress towards that 55% that we're looking for. Still work to do, but a lot of progress over the last few years. When you think about the unit economics, I'd offer up a few thoughts. First, as Dev was talking about, and he was mentioning workloads, this $84 billion market we're going after really isn't even customer by customer, but it's more workload by workload, right? I think it's easiest to think about unit economics on a workload basis. That first workload is the hardest to get, right?
It's the most expensive. It's the most difficult. What we've seen, and you may remember we had some analysis on this actually in the S-1.
Yeah
... that talked about sort of the incremental progress, right? That second, third, fourth application were more successful. The velocity is higher. You start to win them in bigger chunks. The deals become bigger. That's really where you see a lot of the operating leverage when you think about the cost of acquiring that workload. Customers, you know, we're still pretty early on. You know, in the journey, our relative penetration, even a big, you know, large opportunities, most of our, you know, $200+ million+ customers, we still have very, very small wallet share of, right? There's still an opportunity. Yes, you do see improving economics over time.
It's not, you know, I wanna caution people, it's not sort of a full on like set it and forget it mode of like we've acquired the count, now everything just sort of comes to us. We do need to keep investing in it, but we see, you know, meaningful economies of scale.
No, that's great. A great context. I do want to go to the audience, see if there any questions for the team. Before I get there, just a couple of follow-ups. On the 20% operating margin target, how should investors think about the timeline to get to that level? On the topic of stock-based compensation and share dilution, are you managing to a specific dilution level, and what would that be in calendar 2023 and beyond?
Yeah. On the target margin to 20%+, we've not given a timeframe. We've made continued progress against that. We said in our December call that we were expecting to make around 100 basis points this year, would be in fiscal 2023, and we'll be happy with that next year. We wound up doing more in fiscal 2023. In our guidance that we gave last night, we're able to kind of deliver the additional 100, so about 250, if you think about it that way. We're pleased with that pacing. We are trying to balance the long-term opportunity that we have, and so we're not trying to aggressively, or sort of over course correct, to emphasizing profitability at the expense of growth. We're taking a long-term orientation.
I think that's probably the critical thing. Do you wanna talk about stock-based?
Yeah. So we are obviously in a business where it's really important for us, excuse me, to attract the best people in industry. This would not be a really good business if all we're getting were B or C players. The thing is with, you know, so we index our compensation to the market, and the market is really set by the larger tech companies and, you know, more recently startups. Obviously people are starting to rationalize their behavior, but we have to play in that market. That being said, we are not doing any make-whole grants or broad-based special grants to our employees given what's happened with our stock. You know, we recognize that, you know, when investors feel pain, you know, employees should also feel the pain commensurately.
That we also do recognize that, you know, we are all, we're shareholders ourselves and we do care about dilution and, this is something that we're gonna constantly work on, you know, as we continue to build the business.
Appreciate the thought, Dev. Let's go to the audience and see if there's any questions for the management team. Right here in the front. Tom? Back.
Thanks for taking my question. Just wanna dig into the February improvement comment a little bit, since that does stand out relative to some of the, your peers at this conference in cloud world that still sound very uncertain on the near term and when there's gonna be a stabilization on some of the cloud optimization efforts that are going on in customers. Wondering if you think that's something company specific to Mongo or, you know, is it a kind of a broader market read? Thank you.
Let me address the point about optimization. You have to remember, one, we've seen no change in the dynamic, but you have to remember, we are highly aligned with our customers. When a customer builds an application, they want that application to be used because you know, they have scarce development resources, so no one's gonna build apps that no one's gonna use. They see value when that app is being used, and we see that revenue when that app is being used. We are highly aligned with our customers. Now we have seen in the past year some corner cases of some customers under severe financial duress who've done some things about re-architecting their, you know, clusters to basically sacrifice performance and resiliency and take on more risk.
Those are corner cases and that's not really stable, but, you know, in general, we don't see customers, you know, focused on like trying to optimize, you know, like many other companies do. We see the, you know, what we see is a high correlation with read growth with, you know, revenue growth. As they're, you know, as they're using their applications, you know, they're upgrading their clusters and that's been a trend for the longest period of times. Maybe you want to talk about February.
Just quickly, you know, February was rebounded from the pronounced holiday slowdown that we saw. February also, when you think about the expansion of existing cohorts was in line with really what we've seen, the average that we've seen over the course since the start of the turndown in Q2. We're pleased to see, you know, that recovery. Obviously it's still below pre-macro levels. Wanna appropriately, you know, make sure folks know that. When you look at our guidance, and this all relates to Atlas, when you look at our guidance from an Atlas perspective, we're assuming that that is the outlook, right? We're assuming that things don't get materially better or materially worse.
Obviously to the extent that they are better, that would benefit us to the extent that they deteriorate, that would be adverse to us. On Enterprise Advanced, just to kind of round out the guide view, 'cause I know we're coming up on time here. We had a very strong year. We faced a tough compare there. We expect to see growth, you know, in Enterprise Advanced, but just not as strong as we saw last year.
Makes sense. With that, thank you so much, Dev and Mike. You had earnings last night. You joined us on stage at 8:00 A.M. We really appreciate you joining us at the Morgan Stanley TMT Conference.
Thanks for having us.
Thank you.
All right.