Couchbase, Inc. (BASE) Earnings Call Transcript & Summary

This is on? You get the award for the best dressed CEO we've seen. I've seen only 10 CEOs in the last day or so, but this definitely. Do you agree, Matt?

Yes. Oh yes. Very good. I'll take it. If I leave here with nothing else, I have one win.

No. It's so good to see. I think the last time we met face to face was before the pandemic. And so, you're looking great. Thanks for coming. For those that I am somewhat familiar with the origins of Couchbase, but I do remember the first time we met, you walked me through how the two entities merged, and you created this new company and the reason for its existence. If you just can walk back to the founding of the company and when you got involved in it, that would be great.

Yes. So we are very fortunate to be an Accel company. And a guy by the name of Kevin Efrusy, who's one of the longstanding partners at Accel, met one of our founders, Steve Yen, at an interesting time in the evolution of technology and databases. Kevin at the time was heavily engaged with Facebook, having sourced that investment for Accel, which I think we can all agree was successful. But he was working with the technology teams who were scaling the Facebook backend. And Steve, our founder, amongst others, had been around technology as an entrepreneur himself and really understood the limitations of databases. And he said, "God, I would love a world where databases didn't suck." And Kevin captured on this. He said, "Well, that's pretty interesting because I'm hearing a lot of 5-letter words coming out of the team at Couchbase. This whole sharding thing, how the database is slowing down progress." And from that very time, Couchbase set out on a mission to build a database at scale that didn't suck, or set another way that opened up opportunities for application development for a future world where the database was not a limitation but an enabling platform for application innovation. And we've been fortunate to be executing on that vision for the last almost 15 years now.

That's good. So as you look ahead for Couchbase, what are your goals for the company 4 to 5 years? And maybe as a precursor to that, to us, or those of us that may be listening on the webcast, how would you describe the positioning of Couchbase, the reason for existence of Couchbase, the category that you're trying to create in an otherwise crowded market where, as you said, databases are very -- I won't say what you just said because I'm on a public webcast, but there's a reason why database category has been a very tough one, but you obviously do something really well. Explain to us that positioning.

Yes. I think about the morning that I've had from the time I woke up and until I'm in this very seat, and my guess is I'm not alone that I've probably touched dozens of applications over the course of my day, getting caught up with messaging, maybe booking some travel, checking in on some financial services, so on and so forth. And when I think about things at a very high level, I say applications are the gateway between our human world and the digital world. And literally everything we do in our lives in the digital world runs through an application. Applications do not exist without databases, full stop. And so, I think you cannot think about a database company without thinking about the applications that database is built to support and power. At Couchbase, we've been very ambitious to say, let's take on the most mission-critical, high-performance enterprise applications. So the ones that are touching the most amount of data that are being accessed by the most amount of people. So if you're ordering something online, if you're applying for a credit card, if you're accessing streaming services, if you're an aircraft carrier getting thousands of passengers from point A to point B, all of these things require massive datasets that need to be accurate and reliable in real-time to power the very applications that we're dependent on in our lives. And so, simply put, we are modern database for enterprise applications which require a level of scale and performance and distributed computing that other applications -- or I'm sorry, other databases, simply aren't architected to support. Now, when we think about the future world, applications need to be highly personalized, highly customized, they need to be available at any point of the network topology, anywhere in the world where we pull up our phone, we want that application to work. And I need my application to work differently for me than your application working for you. It needs to know who you are, it needs to be answering questions on your behalf that you're not even asking. You start to layer in GenAI and allow your imagination to well, how are these applications going to evolve over time, and then you ask yourself the question, well, what about the database powering those applications, the attribute of scale and performance, distributed computing, cloud to edge, those are totally aligned to the very thing that we've been building for the last 15 years. So we think that enterprise, mission-critical, applications of scale that matter, that has been our North Star, and that's never been a more exciting time than it is today.

So what are your goals if this were to become a bigger category? What do you think Couchbase looks like 5 years from now, what would you like, more importantly, for Couchbase to be?

We had our Financial Analyst Day not too long ago, and we talked about these 3 phases, the path to getting public. We were really focused on building a great foundation. That's everything from our core architecture, go-to-market model, we needed a set of financials that we were proud of, and that we are in this phase of inflection, which were taking the company to the next level, which was really the addition of our Capella offering, building on top of that foundation. We very much see a path to $1 billion software company and beyond. And what I talk about internally is that's inflection, Capella plus AI. And what I don't think people are thinking about when they think about the great market transition that is AI is what's happening at the data layer. And fundamentally, there are going to be databases powering applications in an AI world, and we believe that we are going to be one of the most important ones to do so. And so it's an incredibly exciting time to think about how strategic we can be to future enterprises and the business that we're going to build along that journey.

Yes. So it's a perfectly logical thing, at least in my mind, to ask you about the role of GenAI. There are some that are positing that GenAI is such a radical change in the architecture of how applications are going to be built, that if I had to start a SaaS company today, I wouldn't build it on middleware, database, and business process logic, et cetera. I would straight away have a RAG LLM learning from data. And application design is going to be -- not that I believe in it, and we're all vested in the things that we grow up with and believe. But what do you think about how the Couchbase architecture would fit in that GenAI world where applications are written natively in a different way? Or maybe you have a different viewpoint. I would love to hear your thoughts on that.

Well, first thing I would say is we're incredibly excited by AI, because you'd be hard-pressed to think about any application that isn't going to be augmented, improved, enhanced with the benefits of generative AI. And so, I think it's incredibly exciting. If I allow my imagination to go, and I think about connected homes, or an operating room of the future, or autonomous vehicles, or the future of travel, and I think about the experiences that would be that much better than we have now, there's certainly an element of generative AI that's going to curate experiences that wouldn't be possible without that. But for the entirety of my experience to be fulfilled, it's still going to be dependent on data that we're dependent on today. How much money do I have? Where am I located? Who am I traveling with? Where am I going? And so I think the magic of AI is going to be a combination of structured, semi-structured, unstructured data being brought together in near real-time to unlock experiences that would otherwise be impossible. And so, when the discussion of RAG and this goes on, people say, well, AI apps. I like to say, well, help me understand what application you're talking about. And I'm yet to come up with applications that don't still depend on data that we're leveraging today. And so, then I go back to the data platform, well, how do you get all of this data together in a data store in near real time? The computer science behind that is very, very complicated, and you have to have architectural tenets in your platform that are designed for that very thing. So some of these next-generation technologies that are enhancing things on the margin aren't doing so with the platform extension that I think is going to be required to power applications on a go-forward basis. Now, we benefit from countless conversations around the world on a daily basis with enterprises that are on the bleeding edge of innovation and really understanding where are they using this technology to drive their application agendas. And fundamentally, Kash, we think if we can build the most dense, high-performing data platform for operational use cases and then get really creative on how we bring the right LLM to the data, as opposed to the other way around, it's going to unlock sparks of innovation that are pretty exciting.

Got it. Why don't you talk to a little bit about how the data layer needs to evolve to support AI applications? You touched on it a little bit, but there's a big debate out there around RAG versus fine-tuning, where vector search fits into this. You guys just announced your columnar capabilities as well, so talk to us a little bit about the Couchbase story and the evolution of the data layer to support AI.

Yes. So I'll talk about a few things. AI is a powerful tool. Bringing capabilities to the hands of developers, we think, is a pretty obvious thing that's going to benefit Couchbase. So we have our copilot IQ inside Capella. People get access to the full power of Couchbase through natural language. They can get an application up and running, so I think that's a big part of it. Vector search, we're embedding into our platform. We announced vector search for server. We also just announced vector service at the edge. We do think that vector is an important extension of our platform and a capability. I think some companies are talking about that as its own product. We do think it's a service as part of a platform. Scale, performance, TCO are all important vectors of innovation on vector technology, which we're aggressively innovating on, but we think we're going to be able to vectorize everything inside the dataset and open up different types of search beyond just vectors, so hybrid search, which is really important for enterprise applications, and then thinking through a cloud-to-edge architecture with the various data formats is paramount to our strategy. So I think we're taking a multifaceted approach to this evolution, but we are very excited about the platform we have in place to go forward along that.

You raise an interesting point around standalone vector databases versus the approach of taking your operational data and having vector embedding sitting alongside it. What are the advantages and disadvantages of both sides of that coin?

I think of an application, and I think of this concept of application anatomy, the application needs to be powered by database technology. And vector is a sliver of database technology for a subset of applications, but there's still other dependencies on that application. So would it not be more powerful to have a platform that spans a broader set of that anatomy? And when we think about our competitive positioning, we want to be the operational data store for high-performing applications. When you layer in vector search, that gives us more of that anatomy. With the announcement that we made on columnar, if you think about a JSON application, the analytics stack around that is relational by definition. You've got ETL data. You've got to put that into a different analytics technology. The developers are beholden to waitx to other personas in the account. We've removed all those barriers and opened up the ability to do analytics with our platform while the application is running in real time, all in the hands of developer. It's yet another example of expanding the platform to support a broader set of that anatomy.

And Matt, maybe shifting back to the competitive landscape. You guys are a nonrelational database. There's other cloud-native vendors out there in the market, both the CSPs and some of the other players in the market. What is Capella's right to win relative to some of these other players out there?

Yes, look, we think about, from a product perspective, a concept of sustained differentiation with everything we do. So our job is to understand unmet, underserved needs on behalf of enterprises and build great products and services to fulfill those needs. If we're doing our job, we need to stand up against any database out there and prove that we have utility that they don't for whatever reason. And so it's not lost on us that every application owner is going to evaluate the full suite of products out there, inclusive of ones that are available from AWS and Microsoft and GCP. But the combination that we have in Couchbase Server of scale and performance, SQL compatibility, different search technologies, vector, cloud, those are unique and differentiated platform capabilities that people choose Couchbase as a strategic go-forward platform. Now we're going to continue to innovate there to make sure that we enjoy those advantages, but Couchbase Server feeds Capella and will go head-to-head with any one of those databases. Now, I think customers understand that, and so I'd be remiss if I didn't point out how important those partnerships are. Because when we sell Capella, customers benefit from drawing down cloud credits. It's a big part of the value proposition. It's really important for us to have integration with those clouds at increasing levels of capability. So we spend a lot more time probably thinking about the partnership side of that than the competitive, and maybe it's so innate in us in building unique features that's what we do every single day that it comes pretty natural.

You mentioned Couchbase Server feeding Capella, right? So maybe we could talk a little bit about that dynamic. Capella is in its early days, but it's growing quickly. So what initiatives do you have underway to accelerate Couchbase Enterprise's migration to Capella? Is there migration tooling? Are you working with partners? Talk to us a little bit about that.

Yes, we want to make sure that enterprises have a seamless path that want to go from enterprise to Capella. And so, we are building migration tools. We're ensuring that we have a partner ecosystem to help with that. I think there's dynamics on software versioning. So if, let's say, a customer who's been with us for some time, maybe not on the latest version of Couchbase Server, that can create a step that is required to get them on the appropriate version to make the eventual move from Server to Capella. But we've got services capabilities. We're working with customers on those initiatives. So we're not going to let the technology step of enterprise to Capella be a barrier. And quite frankly, I think we have all the capabilities inside the company to be able to say that with confidence at this point. At the same time, we're not in a position where we're pushing customers there before they're ready. And we want to make sure customers have choice and that Capella is an industry-leading experience once they get there.

And in an end state, when you think about maturity of Capella versus Enterprise, do you anticipate that Enterprise will still remain a reasonably large portion of the overall business.

Yes, I think so. I think we see a path to Capella being 30% to 50% in the medium term. But there are enterprises that are still making decisions to manage on their own and deploy in certain circumstances. So we feel that it's very strategic to have both. The fact that there's compatibility between the two deployments and customers can live in multi-hybrid edge deployment simultaneously with enterprise running here, Capella running a subset of applications over there. Again, as a strategic partner to enterprises, that's a capability that they want, and we will provide that to them.

So picking up from that point, how increasingly are your sales people landing Capella as the point of entry with new accounts?

Yes, look, we had a great quarter on new logos, best in company history. So they're doing some things right. I think we are getting much better at understanding an initial developer's need for their database, curating a great experience for them to understand what those capabilities are and get up and running in the production. I think we're enabling our sales teams on areas of particular success so that they can replicate that, starting small and the power of letting the technology grow. So I think we've learned a lot in some of the differences in traditional enterprise to cloud-based adoption. But at the same time, we're still landing enterprise logos. So when I step back, the most important thing is we're understanding how to articulate the value proposition of our database and data platform, how that aligns to application agendas, and then ensuring that customers have the appropriate way to consume that technology as they go forward.

So it's not segmented by use case or profile of the customer, intensity of the workload, complexity of the workload. It's more customer preference.

Fortunately, we can support in both deployments. And the fact that we are such a high-performance, broad platform, we satisfy so many different verticals and use cases, I think it's more, Kash, about understanding the preferences of the development teams and what they're trying to do that really is the bifurcation between Enterprise or Capella out of the gate.

So if I'm an Azure developer, so I would tend to go with the cloud version of Capella.

Sure, yes.

Or if I'm an on-prem developer who is used to...

Or maybe I'm in a particular geography or in a particular line of business where there's complexity on cloud for regulatory situations, and I need to stand up my own data center with a particular set of attributes. And so, we play in use cases like that, where it would be inappropriate for us to try to push Capella. It's just you won't have product-market fit.

I just thought about this. It's a different sales force that sells Capella or is the same sales force that's got "Enterprise" sales?

No, no. We have a fully integrated organization.

Okay. So I could sell both.

Yes.

So the incentive structure, how do you make it neutral for a sales person? Because the ACV of a Capella deal is presumably not going to be as big as the ACV. Maybe the TCV makes it even bigger of a difference, right?

Yes. Look, there are significant benefits to Capella. And so, we are wise in ensuring that the incentives for the sales teams are appropriately aligned to unlock that. So it's not necessarily about neutrality. What I want to make sure is that we don't have conflict. But selling a dollar of Capella Couchbase is better than selling a dollar of Enterprise. We want to make sure that we don't get at odds with customer needs. And I think we've proven with our incentive structure that we can strike that balance. Now, a big part of this as well is once a customer goes down the Enterprise path, they're not on that path forever. So a big part of this is enabling them to migrate. We've got some major, major, some of our biggest customers that within the last year have told me Capella's not the right offering for us for the various reasons, who within the last year have come back and said, actually, we've changed our mind. We want to get deployed on Capella as fast as possible. If we had pushed that too early, we would enjoy the same strategic relationship. And so I think it's an eventuality that a lot of these are going to move to Capella, particularly if we continue to do our job and make that the amazing experience that it is. But if we focus on the Couchbase core platform differentiator and have two great modes of consuming that, I think we're pleased with the balance and the trajectory there.

Got it. I remember asking you many, many years ago, what's the biggest scale application running on your platform? You talked about that use case and I happened to have witnessed that. So with Capella, what is the biggest scale deployment? If you cannot mention the company name, you can anonymize the name, but what is the biggest at-scale production deployment of Capella?

Yes. We have a company in the financial services space, payments, and they moved a seven-figure Couchbase deployment all the way into Capella. The number of people that are moving money around the world on Capella would blow your mind. It's staggering, the scale and reach. And this is what I don't think people fully understand about Couchbase.

That was going to be our last question, but it's okay. It's perfect. What do people don't understand about it, including blokes like me, what do we not understand?

We have customers that quite literally their business would stop if our database wasn't performing. Some of the largest companies on the planet, we are running their revenue engine. Like literally, they're not taking customers' money if the Couchbase deployment goes down. We have travel companies that the world stops if the Couchbase application isn't working. We have deployments in the healthcare space where next-generation operating rooms are running on Couchbase technology with mobile solutions in a way that if power goes down, we're still up and running. Patients' lives could be impacted. So when we talk about enterprise scale and performance and the requirements that come with that, I don't think there's an appreciation for how big these applications are and how complex it is for enterprises to pick strategic platforms that they're going to bet on a go-forward basis. Now, what you would have said to me before is, well, that's interesting, but how many of those applications are there, right? And we would have debated that over time. What I would ask you now, though, is with AI coming and the way that data is growing, are there going to be more or less applications that are going to be dependent on data in real time that are going to impact our lives if we get it wrong? And what I don't think people have appreciated is that we have carefully protected this architecture for mission-critical applications with knowledge and the ability to see that more and more applications are going to need those requirements than they did before. So maybe a subset, but we've been playing for the medium and long term. And I think AI is going to be the single greatest catalyst for application innovation, which is only going to reinforce the needs of these attributes that support these applications.

You'd be a good spokesperson for the team, so you sound very good and articulate. But do you believe, as much as enthusiastic as you are about generative AI, do you think we need a killer application or two or three to really validate this theme? How far do you think we're away from...?

I think they're going to come. And I think what I've learned about being at a database company is we are not the only ones that need to think about innovation. If you think about, again, just the applications that we use and how companies are innovating with gaming and connectivity and social media applications, software developers are insanely creative. They need a tool that can allow their imagination and capabilities to unlock their very innovations. I have no doubt about AI being a world-altering technology. What we study and are paranoid about is ensuring that we have the capabilities that are going to enable that unlock. And we have customers that are deploying AI at pace in production with things like chatbots and improving existing customer interfacing applications at some pretty interesting levels. We have another customer that was running an application with next-gen AI. It had to do with a promotion on the revenue side. In test, it hallucinated. Had they rolled it out, they would have been beholden to the offer, which would have cannibalized their business. And fortunately, they put a stop to it. Now, eventually, are they going to roll that out? For sure. But they got to make sure they get it right. Now, again, we don't need to perfectly predict which way and at what pace, so long as we understand the data layer. And I do think back to what I said earlier, there's less of an appreciation for how important this layer is to unlock the applications, whatever they may be.

I asked Dev yesterday from MongoDB that, would you ever consider becoming an applications company? And he said, no, we don't have that dominant share of the database market to be able to doing something like that. In an earlier instance, he quipped saying that, no, man, hallucinations are going to get in the way. We cannot take that risk now because we work with large, prestigious companies, as you do work with. So you think hallucinations are a stumbling block for the development of large-scale GenAI?

I think it's a factor, but I'd say any emerging technology, there's things to overcome, and we will be better at it tomorrow than we are today. And I think the pace of innovation with LLMs and generative AI is probably exponential in nature. And so, it would be unfortunate for us to take these limitations and assume that those are going to play out forever. If instead we say, well, what happens when those converge to something near zero, what's possible and where does the risk tolerance, marginal cost, marginal return unlock? Our innovators are well down the path of assuming improvement there and unlocking what's possible.

Got it. Any questions?

I had one.

Yes, go ahead.

Matt, just on the Capella ARR component, you guys set a pretty wide range, 30% to 50% of ARR in 4 to 6 years from the December Analyst Day. So when we think about exceeding the upper bound of that range, what has to go right for Couchbase?

Look, I think it's a function of how quickly customers want to migrate. We need to do our job and ensuring that the experience is fantastic. We had a customer, one of our larger Capella ones, who's been a legacy enterprise customer. We were early in market with this Couchbase managed service before we eventually got to Capella. They were on the whole journey with us. And this person got up in front of our sales organization and said, "Look, I've been doing databases my entire life. And I've always had this dream that the database would just work." He said, "Now you all have been telling to me for 8 or 10 years telling me that it'll eventually do it, and I really didn't believe it until now. But my dream is realized. It just works. I focus on building applications and it works. It scales. It's doing what you said it would." That to me was validation that we had the experience where it needed to be that customers now are making a choice beyond technology and what deployment model and business model works for them. I think it's an eventuality that we will get to the upside of that. Question is over what time period. And some of that is outside of our direct control. But what it's ultimately going to come down to is how fast does the base decide that that's a better deployment model for them.

And in terms of what you can control, talk to us a little bit about the traditional migration path from Couchbase Enterprise to Capella. What does that look like today? Do you think you can shrink that timeline to incentivize more customers?

I think we're seeing it. We had a great quarter on Capella, both in terms of percent of customers and percent of ARR. And what we're seeing now is we don't have to move the entirety of the deployment. We can start with a single application. And so, I could take you through 1 through n, every customer that has any ARR with us, and we know exactly where they are in their plan or not right now or lack thereof on moving to Capella. What we need to do is make sure that we are aligned, capabilities, services, application deployments with the pace that makes sense for them. So back to your question earlier, do we have all the capabilities in terms of migration tools and services, get their installed base up to the appropriate version, I think we're prepared for the potential to be on the upside of that and beyond. But we've got to execute.

What is the uplift from a revenue or ARR standpoint or whatever metric you feel comfortable sharing, when someone goes from Enterprise t Capella? How should we think about that? Is that a captive opportunity?

I think we generally say $1 turns into somewhere at $1.50 to $2.00.

Okay. So maybe more of like an operational question then. We've seen companies go through a somewhat similar transition. If you look at Informatica, they're undergoing one, and the stock did show pressure, but investors have really appreciated how they've operated thus far. Have you considered maybe taking a more suggestive approach around the migration to capture that captive base? Or is it more so just letting customers migrate as they're comfortable and following that?

Yes. I think life is a balance. I think we're trying to get that balance right. If I have a Fortune 10 customer who has invested in Couchbase for many years, their business is running on the deployment, and it's working perfectly fine. If I showed up to them and said, you have to move over to Capella, we're going to end of life, that over time we are going to prove to not be the trusted advisor and strategic partner that we need to be over the many years that we plan to be doing business. And so, I'd say we are handling those discussions as if we were on the other side of them, which is here's the value proposition that's in front of you. We believe it's a better TCO. We're going to help you with that experience and doing it in a partnership-based approach as opposed to moving people. Now we are very convicted that the offering is compelling, and we can do things with TCO. So I think most of this for me is about finding the right time and pace for these migrations, and customers have different variables that drive different outcomes there. But I'm not overly concerned about our ability to get over there. I want to make sure that we're doing it while showing our value as a company beyond just the offerings that we provide.

Okay, great. Thank you so much.

On that note, we're out of time, but thank you so much, Matt, for coming.

Thank you. Great to be here.

Good entertaining conversation. I love databases, by the way. I've grown up around databases. So this is fun, always fun for me.

All right. Thanks, guys.

Thanks so much. Round of applause. Thanks a lot, Matt.