Box, Inc. (BOX) Earnings Call Transcript & Summary

Steven Enders from Citi Software Research Team. And with us for this session, we have Dylan and Aaron from Box. So thank you both for being here.

Thanks for having us.

So I guess maybe just to start off. There's a lot of changes in Box over the past few years came up after the pandemic, a lot of changes on the product side and go to market. Maybe kind of walk us through what's different with Box today versus 3 or 4 years ago?

Yes. So maybe I'll point to 2 major elements of our strategy evolution on the product and go-to-market side and then maybe Dylan, can you talk about some of the bottom line kind of operational efforts as well. But basically, kind of the pandemic is an interesting point in time, but if you just sort of go back arbitrarily, again, sort of 3, 4, 5 years. Box was largely a platform that helped companies sort of store and share and collaborate on their most important files in a very secure way, but we were really squarely just solving that, the kind of collaboration aspects of content management. And I think certainly accelerated by the pandemic, but as a part of our strategy, kind of the whole time, we have been pushing to really expand the breadth of the platform to really power the full set of use cases that companies have around their data and their unstructured data. So that gets into workflow automation, really advanced data security and compliance, things like threat detection, data loss prevention, ransomware prevention. Then we got into the e-signature market because we felt like there was a lot of workflows happening around Box that we weren't necessarily powering, but we were just integrated into that were these more kind of transactional workflows where you want to get a contractor document signed. So we got into the e-signature market. We've seen a growing range of use cases around how companies want to be able to collaborate in distributed work environments. We've got into this product area called Box Canvas, which is sort of virtual whiteboarding and remote collaboration. And now, obviously, with AI, this is really being even kind of further accelerated where we can start to tap into all of this unstructured data that we have and help customers and companies make sense of all the information inside of their enterprise in new ways, and I'm sure we're get into that in a little bit. But basically, sort of job number one, expand the platform, get into multiple sort of key adjacencies to kind of core content and collaboration, each of which have faster-growing characteristics from a market standpoint than the kind of core content management market. So we wanted to enter into faster-growing markets, consolidate more value, more data, more workflows that our customers were doing around their content, all in our platform. And then we complemented that with a go-to-market motion that enabled us to get the full value of the Box platform into the hands of our customers. So -- and this is really our multiproduct selling and kind of suite strategy that we've been very focused on. So instead of selling one kind of core product, i.e. Box and then having some add-on products like data governance or data security we wanted to sell the full suite of capabilities to our customers. And that really reinforced Box as a platform to our customers, and it certainly accelerated our growth rate in the past couple of years. where we could both improve our total average contract value, retention rates, net retention rates and then even gross margin because of the kind of premium nature of these SKUs. So that kind of really brings us to today, which is we're going to keep doubling down on basically that kind of core strategy, add more value to customers, get into key adjacent markets to deal with unstructured data, and then bundle those products together and deliver them to our customers in an integrated fashion. So that's kind of been driving the top line. And obviously, we've made significant bottom line improvements as well.

Yes. And just to build on that, I would say, so 4 years ago, Suites weren't in the market at all. Currently, Suites customers represents close to half, 48% of our total revenue, and that's driven, both -- all the things Aaron talked about on the product side as well as that shift in the Suites have driven higher pricing of about 20% on a per user basis. Since we launched Suites, much stickier, full churn down from 5% annualized to 3% annualized over that time period. And then on top of the customer economics and the benefits, what Aaron talked about, we've also been very focused and over this time period to put in a bunch of really important initiatives to improve bottom line performance. So going back, again, about 4 years ago, we were breakeven. This year, we expect to deliver 25.5% op margins. And a couple of the key areas that we've been executing against are that public cloud migration strategy that Aaron touched on to run our operations more efficiently in the public cloud. We're at the very tail end of that now, also stood up our first engineering center of excellence at a lower cost location in Poland a couple of years ago. That's been scaling really nicely. And then just a lot of efforts around cost discipline across the board to drive that.

Okay. So I think it's a great way to start off. A lot to dig into here. I guess maybe just to start, I just want to talk about the last quarter a bit and what you're seeing out there from a macro perspective. I think billings was a little bit softer and then a guide down on revenue and billings for the rest of the year. I guess, what was different than what you expected from a seat growth, expansion, contract perspective on the deals that came in, in the quarter and the pipeline that was different than maybe what you were originally expecting in 2Q?

Yes. So I'd sort of characterize maybe the first half of this year, and then we can kind of point a little bit to the second half and some of the trends that we're seeing. So the first half of the year, with Box, we started to see the very slightest initial impact of macro headwinds in Q3 of last year and then more pronounced in Q4. And so that has sort of caused us to sort of forecast a certain degree of growth and dynamics into this year. And what we saw in the first half were obviously some variance to that initial view. Mostly kind of on the margin, but enough that we wanted to adjust our guidance. And basically, with the dynamic being that in this kind of environment, we've had healthy logo growth, we have very healthy deal volume. But one of the dynamics is with a sort of a seat-based model is where we can take on one department or 5 departments or the whole company. It's really the discretion of the customer of kind of how much they wanted to play Box and for what use cases, there are certain situations where we'll go into a deal, maybe in our pipeline that looks like it's a $100,000 deal just to kind of arbitrary number, and through that sales cycle, either due to the customer's budget or ultimately, the use cases they take on that deal might come in at $75,000 or something in that process. And so we've been trying to kind of calibrate our full process of how we kind of look at the pipeline and how we see our forecast given some of the dynamic that we're seeing in the customer base. But overall, I think every company on the planet is dealing with some degree of macro headwind and they're making IT decisions based on, okay, what are the most mission-critical things that they can invest in. Fortunately, we remain mission-critical in a key number of areas, hence the 3% full churn rate. But certainly, again, we might talk to a customer that was going to expand Box into marketing, sales and R&D, and due to budget pressures, they just choose one of those departments or one of those use cases and then they'll expand more over time. So that's sort of the dynamic that I'd say, maybe in the first half of this year. We have seen some degree of stabilization in terms of sort of not seeing the same kind of quarter-on-quarter degradation that I think would characterize maybe the first half of the year or starting in Q4 of last year. So that's certainly a positive sign. And then some of our internal KPIs, I think, point to some pockets where we're seeing more -- maybe a slight inflection from the original kind of bottoming out and we'll point to some of the kind of billings dynamic that we expect to see in Q4, which is certainly stronger than Q2 and Q3 of this year.

Okay. So I guess couple of things there. You mentioned stabilization happened in 2Q and into August. I guess, what exactly is it that has stabilized there? And then secondarily, as we think about the outlook, I guess, how scrubbed is the pipeline at this point? Was there like another cut taken to better understand those dynamics? Or like how do we think about conservatism that's being accounted for now?

Yes. So I would say that not a fundamentally different approach to how we're thinking about kind of guidance philosophy. That said, yes, absolutely, as we've learned more about this environment, see more of these deals play out, have taken a more conservative view of what happens in some of those deals, especially if there are kind of a range of options from a deal sizing standpoint. So we've taken a more conservative kind of deal-by-deal view of things in light of the environment. And just overall, what we've done is, as Aaron mentioned, we are seeing the biggest impact on just the rate of seat growth. And so roles really the most pronounced pressure -- amount of pressure through and into our expectations for the back half as well, just to be prudent as we're not expecting for there to be any sort of recovery anytime soon in the way we're approaching that.

Okay. So when you think about the 4Q, I think it does imply an acceleration in billings from 2Q and 3Q. I guess what gives you the confidence in that happening? Like what are you seeing either in the pipeline or just in trends in general that gives you that confidence?

Yes. So actually, on the billing side, for Q4, I would say from a -- what's baked into our expectations and guidance, nothing too different from the Q3 numbers, which the dynamic is more around, in Q3, there are just some other optical pressures around payment durations in particular, where we had very strong payment durations in Q3 of last year, including one large customer multiyear prepayment. So delivered even on a constant currency basis, 20% billings growth, which is kind of related to also just -- it was really tough comparison. Aaron mentioned, we started to see kind of later in Q3 some signs of the macro impact in our business, more pronounced in Q4. So it's really a combination of just a cleaner pump. And is absolutely kind of validated a lot of the bottoms up kind of scrub pipeline analysis that we do.

Okay. All right. That's helpful. I do want to take questions from the audience. So if there's anything out there please raise your hand, and we'll make sure to get to you there. I do want to talk about Suites because that has been a really strong point. But maybe it was a little bit softer in this past quarter. So I guess how are you feeling about Suites at this point? And is there maybe a saturation point as we think about the penetration in the customer base at this point?

Yes. I mean we're extremely bullish on Suites. So I think we're certainly not satisfied with the quarter-on-quarter maybe Suites kind of TAV or revenue base growth. But again, kind of most of that is just correlated to the general seat growth dynamics that we've seen across the business. So we believe somewhat temporary as a phenomenon. But in terms of upside that we see both in seat growth within our Suites customers as the kind of maybe macro, let's say, levels off, we see some more normalized comps and customer growth and the -- just amount of unpenetrated part of our customer base where we know we can continue to sell Suites into. It's hard to get an exact figure, but there's probably another kind of 20% to 30% of the customer base that Suites is relevant for, kind of plus or minus even from there. So we see substantial upside in our Suites sales motion from the kind of current 48% of our revenue base that has Suites. And so I think job #1 right now is we're going to make sure that we get our current kind of top-tier plan, Enterprise Plus in the hands of all of our customers. At the same time, we expect in probably the next fiscal year that we will have an additional kind of plan that has even more added value within the platform, and that's going to become kind of a key investment area for us. So we have a lot of innovation in the pipeline that we know sort of represents even more value to our customers, and we want to make sure that, that value is appropriately priced into the product plans.

Okay. So I guess with the new higher price tier potentially coming in the next fiscal year, what would that look like? And how would it be different from, I guess, how you're thinking about things today?

Yes. It's a lot of specifics for where we're at right now in announcing it. But I would just say, if you look at what we've commented on around our platform, we're going really deep in data security and compliance, going really deep into workflow automation, going very deep in AI and kind of broad content management, data management use cases. So in each of those areas, in our product road map, we sort of see the next, let's say, concentric circle of more value we can offer our customers. And we think that there's an appropriate level of sort of substance there where we think it makes sense to offer a new addition of the product here. But it won't be like a major surprise relative to what we do today. It's going to be even more advanced ways to work with your content and get the most out of your business information. And certainly, with how much AI has, I think, taken the world by storm, you'll imagine kind of components of our AI strategy will also make their way into higher tier additions as well.

Okay. So I do want to ask about the AI bundle. But before I do that, I do -- I think one of the questions we get, especially coming off of last -- is Suites has done really well, and I think there was a view that maybe it changes the monetization angle, more modules, maybe be a little bit more insulated from like a seat-based solution or that being an impact. I guess, how should we be thinking about the impact that Suites actually has from like a monetization angle because it does seems like it is more tied to seats than I think maybe people were expecting?

Yes. I mean I think we are still a seat-based business model in the sense of you buy Box primarily for a certain number of employees or contractors or partners within your [ expectation ]. So kind of the big multiplier effect is the number of seats that you have and then really kind of price per seat is what the Suite element drives. So we still have the dynamic where consistent seat growth and seat retention is a very important part of our business model, but we've seen a huge benefit from the upward kind of emotion from a price per seat standpoint that Suites is driving, and that has helped quite a bit. But within our Suites customer base, if you look at the actual economic -- economics that we see there, higher net retention rate than the average, higher price per seat than the average, higher gross margin than the average. So all things get better, the more customers have Suites and we're going to continue to push on that.

Yes. And then before -- the customers moving into Suites, not penetration, is not necessarily directly tied to seat growth, but tends to be pretty correlated for a couple of reasons. The first of which is, in a lot of cases, when customer is just looking to go bigger with Box, they hit that one where they recognize the value, and they have the budget for it, that will often mean, okay, now I can bring in all these capabilities and open a Box to a new set of users, that's one example. So that's one way that it's not even because of the seat growth, but a lot of the same budget dynamics, if you're just in a really challenging environment, you might, for the same reason, say, okay, I'm only going to expand by 1,000 instead of 3,000 seats. I may hold off a little bit to move into Enterprise Plus is one example. And the second is just once you roll out Suites, that brings in so many additional capabilities that, in a lot of cases, those customers say, okay, now, these customers were before maybe Box didn't see as mission critical or not as valuable to what they're doing day to day, now that we have X, Y, Z capabilities, it just makes a ton of sense, and that's why it often leads to see expansion or is correlated with it, even though they are two kind of different levers.

Okay. All right. That's helpful. Maybe we can switch gears a little bit, unless there's any questions in the audience, and we'll pivot towards the AI discourse. I know you launched Box AI, the collaboration out there with Microsoft Copilot for enterprise content. I guess what has been the early customer feedback that you've heard so far with those solutions out in beta? And I guess, are there any verticals, end markets, use cases that are resonating a bit more at this point than others?

Yes. And just to be super precise for anybody new to the story. So there's Box AI, which is a kind of our AI offering where we embed things like OpenAI and others into our platform. And then we also announced totally separately in integration with Copilot, where we can feed data into Copilot for Microsoft's AI experience. So I'll focus primarily on the Box AI front. So we believe we're in an incredible position to leverage large language models, probably more so than the vast majority of enterprise software. And the reason for that is that what large language models are extremely good at relative to other forms of technology is they're very, very good at understanding the large amounts of unstructured data in the form of words. And where do you find lots of words, you find them in documents, more so than literally any other data type just on the planet. If you take a contract, a movie script, a memo, an earnings transcript, an invoice. All of this data is represented by, again, sort of strings of words that large language models are better than any other, again, technology that we've seen at understanding those words and then letting you kind of interact with them in a natural language way. So for us, this is a profound shift in how we can work with our unstructured information. We just released a report with IDC where they found that amongst kind of general enterprises, but 90% of our data is actually unstructured. So we spend a lot of time thinking about, and Enterprise spend a lot of time thinking about their structured data, what's in the Oracle database, what's in the SAP ERP system and rightly so incredibly important data inside of those systems and in kind of relational database systems. But actually, 90% of their data is in the form of documents and contracts and marketing assets. And all of this information is sort of floating around and being shared by individuals and employees and created by employees and shared with partners and customers and colleagues, and we've never been able to programmatically at scale, understand that data. It's always requiring a human to kind of open up the file, look at it, understand it, read it, share it to be able to do anything with it. And now with AI, we can now do that at a completely different scale with automation. And so we announced Box AI, which basically connects the unstructured data in Box with leading large language models, starting with OpenAI, we have announced a partnership with Google, where we'll be also adding some of Google's AI capabilities into Box as well. And then over time, we've built a platform-neutral approach. So we'd like to be able to offer up any AI model in the future. There's no particular preference that we have other than kind of just quality and performance that we want to offer our customers. And then we're going to make it really easy to interact with those models in our platform. So imagine if you're a lawyer and you're looking at a legal document or contract, you want to be able to quickly understand what are the clauses of this contract that might be the highest risk ones. And I need to kind of quickly look at a 50-page contract and understand that or you're looking at an invoice and you have tens of thousands of invoices coming into Box every month, and you want to be able to automatically extract key data fields from those invoices and then have that invoice be routed to different people as relevant, or imagine if you have a large sales team and there's lots of product catalogs and somebody needs to ask a question like what's the price of this product that we're selling in this particular region, how do you go find exactly the right document that has that answer or if you could just ask that question across a large set of data and instantly get an answer back. So those are the use cases that Box AI is going to be able to enable. And again, I think if you take any one of those use cases and compare them to probably a lot of the AI conversations that we're hearing in enterprise software, these are some of the most potent ways that we can bring AI into an enterprise context. So we're really excited and our customers are incredibly excited as well. We've started rolling out Box AI in a design partner program to a select set of kind of beta customers. And then throughout this fall, we'll be expanding that even further, and we'll have some announcements at BoxWorks, our conference in October that we'll share more about where that's going.

Okay. All right. That's great to hear there. I know we're going to hear a lot more of BoxWorks, but just like conceptually, how do you think about how the product set changes from AI? How do the use cases that you think about that could incorporate AI, longer term? Just how do you envision the product evolving to meet this?

Yes. So I think to some extent, we're seeing for lack of a better term, like kind of random acts of AI in software right now, which is sort of you just like throw AI at a thing, then that thing will get better. And we have a pretty different approach, which is we're so architecture oriented in terms of our platform that we think about sort of where do you slot kind of high leverage capabilities that will have a very wide and positive blast radius, but one where there's a very kind of clear reason for that sort of slot to exist. And so for us, it's both incredibly simple, but then almost infinitely powerful, which is you take one individual piece of data, i.e., a document, contract, marketing asset, et cetera. When you run AI against that what are the things that, that opens up? And we've identified 5 or 10 different things. You can summarize information. You can ask questions of information. You can extract data, you can classify documents, you can automate workflows, but it's all on one simple organizing principle, which is tick the content, run AI on it and then create capabilities that leverage that architecture that make Box just generally much smarter. And so instead of saying, like there'll be 32 ways inside of our product where you'll randomly see an AI thing pop up and then users get confused or you kind of get fatigued about where to use this stuff, this is very clearly aimed at how do we help you automate and bring intelligence to just your unstructured data and make it more useful. So that will -- it's actually going to show up in a key -- in the most important areas of our product, but again, really embedded in how you work with your unstructured data in new ways. So again, security, workflow, asking questions of data. And then we have a -- we expect to have a pricing model that tries to balance both the premium nature of these capabilities that we're offering our customers. So how do we get more value as a result of that. But at the same time, I think our expectation is, if we fast forward 10 years from now, just to choose a really distant point in the future, we probably shouldn't be talking about AI discretely because software will be intelligent. So we also want to anticipate a future state where you wouldn't be at a conference saying, okay, I had to pay extra for a mobile app. So we want to kind of expect -- sort of figure out where these cost curves are likely going to head, and how do you have AI also be an embedded component of the platform and offer the most amount of value to our customers. So that's right now the balancing act, which is offer the greatest amount of value to our customers, but also understand that there's premium use cases customers have, that might even cost us more and we want to make sure that we're sort of capturing value as a result of those premium use cases. So you're going to see a kind of a balanced approach to how this gets priced and packaged in the suite.

Okay. So I guess if I'm kind of understanding that from a monetization angle, you're really focused on this driving price, building new plans and packaging to help kind of support that and probably less focused on new SKUs, at least that's kind of how you're doing it today?

Without -- I would say, because we don't want to preannounce anything inadvertently. I would just say we think about it as how do you have the widest amount of impact, but also capture the greatest amount of value. And then it's worth noting, ensuring that we don't have any kind of gross margin headwinds as a result as well. So that's kind of important to throw in there. Because the underlying GPUs can be expensive, so we do want to make sure we capture value as customers are using this.

Okay. All right. That makes sense there. I guess, similarly, how do you view, at this point, where it makes sense to incorporate AI, where it makes sense to build it out yourself and where it makes sense to partner, because I guess you mentioned like there's a lot of rapid change going on and new vendors popping up and new solutions. So how do you think about that part in the architecture component?

Yes. I mean, I think 2 things come to mind. So I think it was Bill Joy from Sun that had a famous quote, which was, there will always be more engineers outside of Sun than inside Sun, and that was sort of an argument for Open Source and Java back in the day, which is this idea that like you never want to be competing in a market where there's way more talent that is heading in a direction that you could otherwise be leveraging. And so AI is kind of like the quintessential example of that. I generally, other than for very narrow use cases, I'm pretty skeptical of any strategy that requires kind of a proprietary AI or LLM approach right now just because there's just simply too much R&D happening that you want to be able to capture all of that value and accrue it to your platform. So what's incredible about this moment right now is you have basically every large technology company on the planet and almost every significant start-up that could exist training and building some of the world's most advanced AI models. And so for us, that's incredible because -- and just remarkably, somehow they became kind of almost uniform API approaches to these AI models as well. So it's not like all of these companies are all competing on sort of having a scarce resource that they kind of keep themselves. They're actually all competing on something that's insanely expensive to go and build, but then opening up for as many possible use cases as possible. So that's great for us. Because then we can basically be in a world where today, let's say, GPT4 is the kind of the bleeding edge of AI. Google, obviously, is sitting around right now, trying to make something better than GPT4. And Facebook is probably sitting around trying to make something better than GPT4. And so for us, all of that innovation can accrue to our customers. And we have to make sure that that's easy to use and easy to understand for our customers in terms of how you configure and manage these AI models. But it's an insane boon for us. And why we very explicitly will not be in the business of kind of like large training runs, heavy CapEx. This is very much about offering this innovation in as variable of a cost basis to our customers as humanly possible. So that's where we're going to sit in the market and all of our value is going to be the layer of how do you integrate these AI models into your data. And that's actually a problem that, in many cases, actually even more complicated. It's just we benefit from 15-plus years on the platform, but that's where data security, compliance, permissions, having a vector database and index around the content really, really matter, and that's where we offer a tremendous amount of value and expertise.

Okay. With everything going on right now, and I want to bring Dylan in on this as well. How do you balance investing more, trying to put more into the product, putting more into the go-to-market versus trying to drive some of that margin? Because I mean, it does seem like there's a big sea change architecture shift that's happening now with AI. So how do you view the balance of growth versus profitability at this point?

Yes. I mean to say that the philosophy hasn't changed. I think one of the key things, if you go back to some of the multiyear efforts, where we've been on the journey to fundamentally kind of change our cost structure, a lot of that are areas that we can drive leverage without having to really sacrifice growth. So it's not an investor, an engineer or an account executive to go out in sales and talk to customers, or bank, but hey, how can we deliver this service to our customers that much more efficiently. They're getting a better experience, same experience with higher gross margins. How do we deliver this type of innovation and accelerate it as we've been doing. We've had some of the fastest rates of engineering hiring in the past couple of years that we have for many, many years, while still driving leverage in that line item of the P&L by really going big in Poland. And then certainly, there are other trade offs that we do make there. We are very committed to the long-term target model that we laid out, showing significant improvements in our overall profile, kind of both the revenue growth side of things as well as on the bottom line. But there, we more make the decisions looking very granularly at what is the return we're getting, especially on the go-to-market side, right? So every segment, every geography, how our new rep is performing, how does that compare with global averages, really tuning our bets, making trade-offs in that sort of way. And we do the same thing on the engineering side, based on what we're seeing as that evolves. So there's no like perfect -- there's a lot of nuance in it. I would say a lot of what we've been doing from a cost structure side has been in order to enable us to both invest in these critical growth areas while still driving it.

Yes. I think the only thing I'd build on that is I think the -- in terms of this is hard kind of to assess unless you kind of go under the hood of a company. But like let's take like 2 types of -- architypes of companies that you have Apple and Amazon. And Amazon is like, let's innovate everywhere and then hopefully an AWS emerges. And that's a great approach. And AWS did emerge, and that's insane and amazing, but it does mean that you're exploring and wandering in lots and lots of spaces. Thousands of engineers working on a project that may or may not work. On the other side, let's take an Apple, which is like these are like we are going to go super deep. We have very focused efforts. We're not going to do a ton of speculative R&D. This is not a bottoms-up research, kind of everybody -- lots of different things and have thousands people running in different directions. This is very much a kind of tops down in the sense of it's a leadership team coming together saying, what are the big bets. And I think we look much more like an Apple in the sense of we make critical bets. We don't do a lot of wandering in lots of exploratory areas. Part of that is just -- I think we care so much about the user experience, and we care so much about the kind of architecture patterns of the platform, but it's also because we have a tight-knit team where we can kind of right now at this stage of a company, we can make those kind of key selective bets. So I sort of somewhat joke about this, but 2 years ago, everybody was excited about blockchain. Well, we said no, we're not doing that because it makes no sense. So -- but lots of companies were like 3 dozen engineers on the blockchain project, and like we just don't do that. We decide to be very decisive about where is technology going and where can we have the greatest amount of impact and we put all of our resources on that. So we don't have the same kind of extraneous work that's happening in the company that I think some organizations do, and that's just because of the philosophy of our approach to innovation.

So I think we're running up on time here. Just going to do a quick scan, and see if there's any last questions before we let everyone go here. No. All right. Awesome. Well, I appreciate both of you being here.

Appreciate it.

And -- yes, I appreciate everyone being in the room as well. All right. Thanks again, everybody.

Thank you.