Varonis Systems, Inc. (VRNS) Earnings Call Transcript & Summary

All right. Welcome, everybody. While we get situated, I'll read the disclosures real quick. For important disclosures, please see the Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. I'm Meta Marshall. I cover cybersecurity here at Morgan Stanley. We're delighted to have Varonis, Guy Melamed, CFO; Brian Vecci, CTO.

All right. Perfect. Maybe just to kind of start with in terms of -- just those who are kind of less familiar with Varonis or kind of data security in general, just kind of giving an overview of the company.

So where I'd start is almost all aspects of cybersecurity touches on data. Nobody breaks into a bank to steal the pens, they're after money. Nobody gets access to an account, an identity, an API these days, an agent, your infrastructure unless they're really after data. Companies face when it comes to data because they have so much not just in data centers these days, but in various cloud platforms, all of the hyperscalers, the applications, all of the AI workloads that they're building. They face a lot of regulatory risk. What do I have? Where is it? Are all the right controls in place, a lot of reputational risk. They don't want to be on the front page of the Wall Street Journal because they got breached. And they need to do more with less. They need to be able to automate fixing all these problems. And that means they face a lot of operational risk, not just in protecting data, but in making sure that they can do all this effectively. So we have an automatic security platform that answers all those questions, helps them implement controls safely, quickly, automatically, and if you're a CISO, what that means is, if you're a Varonis customer, we stop breaches, we stop you from getting fined. We make it really easy to automate all of the workloads that you need to automate from a security perspective. We minimize how long it takes to detect and respond to threats. If you do all that really well as a CISO, you're doing a very, very good job and you can deploy AI safely and quickly, which is at or near the top of everybody else's priority list.

I can always give the nontechnical answer. Sometimes it's easy to think of us. We're trying to do to cybersecurity, what the credit card companies were trying to do to fraud detection. And when you think about credit card fraud, all of us, when we travel, we get this text message, is this you? Yes, no. And the credit card companies have become really good at identifying what is abnormal. And the reason they have become so good in detecting what is outside of the spectrum is because they know everything about us. They know what we buy, where we travel, who we're eating with. And by having that full set of information, they can identify if there is behavior that is out of the norm. We do the exact same thing with data because we see who opens files, who moves files, who deletes files. And by having that full visibility, we can identify if someone is opening way more files than they usually open or if they're deleting files that they shouldn't even have access to. And whether it's someone from the outside that's taking over their credentials in order to get access to data or that employee is about to leave and now they're kind of getting ready for that next job and they want to take some of that sensitive information. It doesn't really matter what the reason is. We can help those customers be protected. And as Brian mentioned, there are really 3 things that we try to do. One is ensure that you don't have a cyber attack and you can be protected. Second thing is that you won't have any fines that are related to sensitive information that you're not guarding the right way. And the third thing is just the implementation of AI. We can help the organization be better protected towards that implementation because it's becoming really scary with the data that's going out there.

Got it. I mean I wanted to kind of jump to that next. We've highlighted before that data security becomes increasingly important with AI. You're having an expansion of the attack surface, different types of data that you didn't even think were critical before, are critical before. In the past, you've talked about AI being a net positive for you because it's just easier than ever for bad actors to use AI to create attacks. Just what are you kind of seeing in terms of the marketplace in terms of either attack velocity or just -- or customers being concerned about kind of AI security?

When you ask a customer about not just AI security, but AI in general, if the word data isn't in the first 15 words of their answer, they're probably just making it up and they don't know what they're talking about. AI security and data security are incredibly tightly related, and it's not just because it makes the job of the attacker easier. It makes the job of the defender even more complicated than it's ever been. Guy talked giving you an example of somebody using your credit card and a credit card company detecting fraud. Well, with data security, we need to make sure that not only is the data monitored, but the attack surface is much bigger because it's not just one user, it could be one user that's interacting with 1,000 agents. What are those agents actually doing? What infrastructure is out there? So for AI security, which is really what you asked about, companies generally want to do a few things. They want to inventory what AI workloads are out there. What models are they using, whether they're foundation or fine-tuned? What agents are out there? How are models and agents, end users and applications and data stores and code libraries all interacting with each other because that's what happens in an AI world. I prompt one tool or one application or one agent, and now it's talking to MCP servers. It's talking to large language models. It's talking to other agents. It's talking to my code libraries. It's talking to my data stores and other applications. Usually some combination of all of the above. So in inventory, managing the security posture are all of the pathways, the configurations, all of the things that a smart person is going to take advantage of is everything set properly? Are all the access controls set properly? Are you watching how things are using? Do you have the right guardrails in place so that people don't put sensitive information into prompts or that models don't return sensitive information? Do you -- can you prove that you're compliant? Because we're talking about security, security and compliance are very tightly related. There's the EU AI Act. There's the NIST RMF framework for AI. There's all these controls that companies now want to put in place so that they can safely deploy it. So this is what Varonis does, especially with the recent acquisition of AllTrue, we can now offer those capabilities in a depth and a breadth that nobody else can, which means if you're concerned about AI and you're not talking to us, you're missing a trick.

Yes. I mean -- maybe what you described is like this is already what Varonis does. And so -- but investors are always kind of obsessed with what is the AI security versus kind of what are the other pieces? Is it just -- it should be a tailwind for the entire business? Are there kind of areas where we should see -- or product sets where we should see kind of more attention be paced to with AI?

I'll give a qualitative answer. I'll let Guy talk about anything quantitative. But we just announced the acquisition of a company called AllTrue that does AI inventory, AI posture management, AI red teaming, AI compliance monitoring and LLM gateway and guardrails. That is additive to the capabilities that we had for securing underlying data stores and infrastructure and monitoring and securing AI tools like ChatGPT and Microsoft Copilot. We now have the ability to do inventories in new places to scan and secure models, to scan and secure code libraries, to scan and secure and monitor agent usage and the interactions between them. Those capabilities really matter, and that is -- we announced the acquisition a few weeks ago. And customer response -- I think someone from AllTrue said that we've done more demos in 2 weeks than they did in the history of the company before that. The demand is off the charts.

Okay. Got it. Data security is also just an area where we've seen a lot of start-ups, particularly kind of labeled as DSPM. How do the increasing number of competitors influence either just customer sales cycles or the journey of customers finding their way to Varonis?

It's nice that there's lots of other investment -- there's lots more investment money out there doing marketing for the problems that we solve, which wasn't always the case. It's kind of like we were screaming into the void for a lot of years and now everybody has finally caught up that data security is a really big issue. So it's been additive. It's really helped us get into a lot more conversations. That said, it means that we're in a lot more fights. The DSPM market has been -- it was created -- I would argue that we created it 20 years ago, and it didn't exist, and we were the only ones in the space for a long time. But even now, the players that are out there are only doing pieces of what we do because if you're a start-up, you're really only going to do this in the cloud, and you're really only going to do this. And by this, I mean the visibility that they offer in cloud stores that are relatively easy to connect to. We go into deepwater, and we don't just do this in the cloud. We do this on-prem. We do this not just in databases. We do this in files, and we offer far more mature capabilities. And by that, I mean, we automate things that the other DSPM tools can't. We detect things that the other DSPM tools can't. To answer your question, we are now in many more fights because there's nobody that touches us in coverage anymore. Varonis for files, sure, Varonis for databases, sure, Varonis for the hyperscalers, sure. Varonis for SaaS applications, yes, Varonis for AI. So there is nobody that's going to do anything related to data security or privacy where they offer out an RFI or an RFP. We're not -- we're going to be part of all of those conversations now. We're in a lot more fights because there's a lot more players in the space. But when we execute our sales process properly, our capabilities and the matureness of our platform means that our win rates have remained really high.

And just to add to that, the vision for Varonis for many, many years has been to provide the customers one pane of glass, one dashboard that can allow them to be better protected on whatever platform they're using. And when you think about -- and we get this asked a lot about kind of supply consolidation and how -- where is the space going? And when you look at kind of some of the acquisitions that we have made, all of them are targeted to provide way more visibility on way more platforms. The e-mail security, we got a lot of questions about that. Are you going into that space? It's part of the MDDR offering. This is not a stand-alone offering, and now we're just focusing on e-mail security. That's not the intention. That's not the case. But when you think of e-mail together with the MDDR offering and when you think about the AllTrue acquisition with the platform that we have to offer, it makes it way more interesting and 1 plus 1 is not equal 2. That's been the premise of how we thought about kind of the road map, and I think we're kind of moving in that direction.

Got it. I'll circle back to some of the kind of platform building or the acquisitions that you guys have done in a little bit. But Guy, maybe just recapping Q4 earnings just in terms of -- you had some higher conversions than you were expecting. Just kind of level setting where you were encouraged kind of coming out of Q3.

It's a very good question. Going into Q4, when we announced the end of life, we had about $180 million of non-SaaS ARR that we wanted to convert. And I know the #1 question we got throughout the quarter was how much of that number will be converted. We were able to convert roughly approximately $65 million of non-SaaS ARR in Q4. I think part of the reason that number was so high was because of the end-of-life announcement. It generated urgency within our customers, and it allowed us to have that type of conversation of be moving and what's the path to move to SaaS. And then now we are able to give kind of a range for 2026 on what we expect will be converted. And we put a bear case scenario and a bull case scenario, and I think we fall somewhere in between. But in a very, very high level, that $180 million going into Q4, 1/3 already converted, 1/3 is expected to convert and another 1/3 is going to churn. And that 1/3 is mostly focused on kind of the federal and state type customers. So that vertical and those type of customers are the ones that are probably going to be impacted the most. That would have happened no matter what. If we would have dragged this transition over several years, those customers eventually would have kind of churned. We're just kind of speeding things up because I think when you think about the risk of the organization and the fact that the difference between the on-prem subscription value proposition and the SaaS value proposition is night and day. Some of the other companies that have kind of stuck with the transition for many years and weren't rushing to get customers over didn't have that big of a difference between what they were offering on the on-prem and the SaaS side. For us, there has been a leap in terms of the functionality in SaaS. And in orders of magnitude, it's a better product than the on-prem subscription. So kind of dragging it doesn't help the customers and it doesn't help us, and it doesn't help the financial kind of structure and model from a leverage perspective. So when we look at where we are going into 2026 compared to where we were when we announced the end of life, we now have guardrails of what we expect to see on the conversion side. And I think they are much better than what they were when we announced the end of life in Q3. So I think part of it definitely has to do with the announcement of the end of life and the urgency it generated for our customers.

Got it. There's a lot of additional disclosures you gave around your business in Q4. Just where do you think investors should be focused on from a metric standpoint as we head into this calendar year?

Everything and anything we got asked -- no, no, everything and anything we got asked in 2025 was how would the business grow post transition? Are you only growing because of the conversion? Can you give us metrics about SaaS growth, excluding conversions? And we opened the [ kimono ] and gave everything in Q4. And then the first couple of questions we get is why is total ARR x percent growth, which is counter to everything we were asked about. So I think that the focus is exactly where investors wanted it to be. We're giving the numbers. We're giving the conversion number on a quarterly basis. We're showing what is the growth ex conversions. And I think that looking at a business or a part of the business that will be 0 at the end of this year, which is the non-SaaS ARR is not the focus. So honestly, like we have the base -- the bull case and the bear case, getting another $5 million of conversions will have an impact, but they're not as significant as how the business is growing on the SaaS side. That should be the focus. That should be where investors that have asked us all year long for that type of information and now we've provided it, that's where the focus should be.

You should always know investors always want something slightly different than what you...

Yes. We gave everything you asked [indiscernible] I think we were trying to be as transparent as it could be on this thing.

Yes. Maybe I want to take it back to just discussions that have happened over the past couple of weeks around AI and security and just this idea that cyber can be vibe coded, which I think generally misses kind of the competitive moats that you guys have. But just how -- I can give my own response, but you can give a much more intellectual response.

I think you summarized it very well. The technical moats that we have, and they are many and growing means that there's -- listen, I spend my time in front of customers. I don't know anybody that's thinking about vibe coding security period. I don't think anybody that thinks that's a serious -- there are tools. There are certainly software. There are some types of software that are going to be -- that are under a threat. But the ones that are unique platforms in the ways that Varonis is, there's nothing that does everything that we do in the ways that we do it in all of the places, like the depth and breadth of not just the coverage and the automation, and we are using modern tools to help build our technology even faster. It's not to say that -- I think you summarized it well. I'm going to leave it at that. Yes, we don't see it as a threat to our business. We see that there are AI tools that are going to help us innovate even more quickly. But Varonis and data security as a whole, I don't think is under attack.

Okay. Got it. I want to dive a little bit more into AllTrue, which you just kind of mentioned. Just how does this fit into the strategy and particularly given you've done kind of a number of these like smaller acquisitions over the last couple of years?

I think AllTrue, which is now Varonis AI Atlas is it extends our visibility into places, and I'm restating some of the things I said before, but it's really important. The AI inventory now where it's not just the data stores and the supporting infrastructure, it's also the models and the code libraries and the agents themselves, including their interactions and behavior, along with the automation to fix those problems because that's key to what we do as well as the behavior analytics, connecting that into what we do from a threat detection standpoint and the best-of-breed data classification and posture management that we do already. It's a bit of a game changer because if you're any kind of organization that is either building, thinking about building or has built any AI workloads and at this point, everybody is in that boat, you need this level of visibility. And if you would ask me that question a month ago, that would have been my answer. Now because we have a month of runway of putting this in front of customers, and I've been able to see the reaction from CISOs who have said, not only do I need this, I need this yesterday. What I just -- what you just showed me is light years in front of what anybody else can do. And our road map includes connecting it into the Varonis data classification and behavior analytics, which lets us not just secure AI, but secure and monitor AI with the context for intent, meaning we know exactly what data and why it's important and how the data is flowing and we understand behavior in a way that's unique. So the breadth of visibility that AllTrue adds for us, the ability to do things like AI red teaming, I can throw the Varonis AI at your model now, and it will test for things like jailbreaking and bias and prompt injection. The ability to do so much more by just -- with less people, with fewer man hours, but just by extending the Varonis platform, it's a real game changer for us.

I guess, Guy, part of the decision to end of life was almost to just kind of focus salespeople around selling the broader platform. You have multiple other categories that you've talked about expanding into as you've done kind of some of these smaller acquisitions, AllTrue kind of being the latest one of these. Just how are you seeing kind of that process of getting salespeople to kind of focus on selling the broader platform now that they're kind of done with the conversion piece?

So in 2025, our commission plan was focused on obviously selling to new customers, existing customers, but also getting through a large part in dollar terms of the conversion. And we were compensating reps on the conversion side. So we talked a lot about the fact that the conversions were actually cannibalizing the time of the reps because they had to deal with a lot of the paperwork and the bureaucratic checklists related to moving customers from on-prem to SaaS. In 2026, back to basics, reps will not be making money on the conversions. They will be making money on selling to new customers and selling to existing SaaS customers. And I think that's an important component that generates the focus where it should be. And that's part of the reason that we just want to be done with that process being 100% of SaaS out of ARR by the end of this year. But having throughout this year, the focus on kind of the right elements.

Got it. I mean from a -- just from a technical standpoint, you're obviously talking about customers want all of these different areas that you guys have extended the platform into. But just -- is there early evidence of really being effective at selling kind of the greater breadth of products?

So keep in mind, the Interceptor SlashNext acquisition was closed in September. AllTrue was closed a couple of weeks ago. So it's still early stages. But from all the conversations that we're having with customers, we feel very good with those acquisitions and where they can take us in the years ahead.

What I'll also say is it's very easy for the sales force to sell. The story is very -- doesn't change. Varonis will find. We'll tell you what data is out there and what's important, all the context, we'll fix. We'll do that automatically. We'll alert when something goes wrong. Now we do that for your e-mail in ways that you couldn't before. Now we do that for AI in ways that you couldn't before. Now we do that for databases in ways that you couldn't before. So for our sales force, it's not like they're changing their motion. It's just where else can we do those 3 things for you.

Got it. And just in terms of kind of this data-driven moat that you guys have, can you just talk about how the increasing amount of data that you guys are going to see and particularly as we start to do AI, how that moat kind of grows versus kind of the competitors in the space?

Yes. A big part of the technical moat we have is that philosophically, we are going into the deepest water, the biggest data stores. And it's not just the amount of data, it's the complexity of data. In a place like Microsoft 365, in order to really understand all of the context of the data that's there, you're not just scanning for what's sensitive. You're also looking at where that data is living. Everybody's OneDrive shares and SharePoint sites and team sites, and you're looking at all the shared links and the relationships between the identities and the directories and the data and the classification. And so that's what we mean by deepwater. We have done that everywhere. And we don't use shortcuts where these other -- the start-ups and the other DSPM tools, they kind of have to because they were never built to scale in that way, which means they're really only giving you basic visibility and they're only doing it in places where they can take shortcuts. Philosophically, for us, we scan everything completely and we go into deepwater everywhere wherever you have data, that technical moat is key in a world where the amount of data isn't just growing, but the complexity. We talk about -- it's not one user, it's one user plus 1,000 agents. It's not one data store. It's a galaxy of cloud data stores all connected together. It's not just one model, it's the 1.5 million models that somebody has access to in Hugging Face. It's -- there is -- the complexity is much higher and it is growing. So our technical moat that is built at a core of unraveling all of that complexity without taking shortcuts is going to serve us exceptionally well in a world where the amount of complexity related to the data is growing faster than it even was before.

Got it. And Guy, you mentioned kind of changes to the compensation structure around kind of changing what you're incentivizing. Are there other ways either through partnerships that should expand or other kind of channel expansion that can kind of help amplify the reach that Varonis has?

Yes, absolutely. I think that partners are very much interested in what we have to offer because of the value that we can provide customers. And we're trying to focus on the right partnerships, not necessarily the number. It's not the quantity, it's the quality. So we're definitely putting a lot of emphasis there. I think the marketplace, both Microsoft and AWS are good partners to ensure and they are interested, obviously, with kind of the consumption part of our selling. So we're definitely focused on that. I will say, though, that we always want to make sure that we take our destiny in our own hands. So it's not like you outsource it and sit and not do much. So we're definitely focused on the outside sales team and the marketing campaigns that we're running together with the focus on breaching out with some of the channel.

And we spend a lot of time talking about kind of top line drivers as you do at these conferences. But just how does that kind of trickle down to the bottom line? And just how do you see the margin trajectory of the business over the next couple of years?

The conversation about margin, I think let's start from 2023. When we rolled out and when we announced the move to SaaS, we laid out a 5-year plan. We put kind of expectations from an operating margin, ARR contribution margin perspective and free cash flow perspective. People were somewhat skeptical in our ability to generate cash throughout the transition, especially in the early years. But we have done a great job of improving our -- not only the margins, but also the free cash flow and generating some significant cash in the early years where you have the most investment that relates to the transition. And we were, I'd say, probably 2 years ahead of plan in getting to the numbers that we have in the model from a bottom line perspective and getting close from a free cash flow perspective. And the only impact that we have on free cash flow in 2026 is that component of customers that aren't converting, mostly that state and government, as I mentioned before, and that's roughly in that $30 million to $50 million bucket that instead of dragging that over multiple years is happening in 1 year of churn. And you don't want to kind of eliminate all your investments because of that one isolated event. So in 2026, there is somewhat of a headwind related to that element only as you continue to invest, but there's no change in the philosophy. And I think we've actually been one of those companies that have always been focused on top line and bottom line and free cash flow generation. And we always believe in making investments that make sense from an ROI perspective, that philosophy hasn't changed. And we've been very focused in kind of getting to that model. And I think that the path for that model, the 2027 numbers is still there. We see how we can get there. And we feel very good with the cost structure. And there's also some benefits to the bottom line that can happen post transition, post that 100% SaaS of ARR at the end of this year that we will benefit from in 2027. So ripping off the band-aid has an isolated impact on the free cash flow. I want to remind everyone that our guidance philosophy hasn't changed either. We have guided for top line, bottom line and free cash flow in the same way that we have guided in the past, which means that we -- that's the starting point, and we hope we can do better than that. I would say the only KPI and the only metric that we have guided in a different philosophy is the conversion number with a bull case and a bear case and the expectation is that we will be somewhere in between. But all the rest of our guidance, including the free cash flow is the starting point for the year with a hope and desire that we can do better than that and progress throughout the year with better numbers.

Got it. And then just the last question. Obviously, it's become more of a discussion point of late around stock-based compensation. Just how are you guys thinking about stock-based compensation? Is there a need to kind of rebalance stock-based compensation, cash compensation? Just what are those discussions you're having?

If you go back several years and analyze stock-based compensation, there are really 3 metrics that we look at. One is obviously the dilution side. The other one is stock-based compensation out of total ARR. And if you look back, that percentage has actually come down dramatically over the last couple of years, and we've made a conscious effort to ensure that we're kind of thinking of it holistically. And the third component is the stock-based compensation in dollar terms. And I think if you look at kind of the elements that we have focused on, we have done a good job. We obviously put a buyback in place of $150 million, and that also offsets some of the dilution. And I think we -- for obvious reason, I think we put that plan in place because we believe that it's the right time to do that. That will offset some of the dilution. But if you look at some of the metrics on stock-based compensation out of total ARR, we have reduced that percentage significantly over the last couple of years.

Got it. All right. Well, Guy and Brian, thanks so much for spending some time with us, [indiscernible].

Thanks very much.

Thank you very much.