ServiceNow, Inc. (NOW) Earnings Call Transcript & Summary

Thanks, everyone, for joining us. My name is Sterling Auty. I'm a software analyst here at JPMorgan. Very happy to have with us CJ Desai, who's Chief Product and Engineering Officer at ServiceNow. CJ, thanks for joining us. Really appreciate it.

Of course. I'm delighted to be here, Sterling. Thank you.

Listen, maybe just to start us off at a very high level, how would you say that the kind of the core ServiceNow platform has evolved over your tenure?

So I've been here 4.5 years. My good friend, Frank Slootman hired me, and he's doing some wonderful things in the industry. So the biggest thing is, I always had conviction, Sterling, that ServiceNow is a platform company, and innovation in platform benefits all the product lines that we have tremendously. So our dollar invested in platform has multiplier impact when you look at how every product benefits. So what we have done over a number of years specifically is, I will start with a simple thing: integrations. So many of our customers said, hey, ServiceNow needs to integrate with X, Y and Z. Let's just pick Microsoft Teams. Are you going to buy something like some of your competitors have done? Or are you going to build it in the platform? We built it in the platform. We call it IntegrationHub, and we created spokes for many of the Microsoft Teams, Zoom, et cetera, et cetera. So the platform now has core integration capabilities for most widely used applications out there, from the SAPs and the Oracles and the others of the world, including Adobe. So integration is one area. Second is, as you saw, Sterling, beginning 2017, we started acquiring a lot of teams that had really good machine learning engineers. And then we said, okay, your focus is not to have machine learning as a generic technology but machine learning for ITSM use case or for CSM use case. So AI and how much we have evolved, including AIOps, is the second area besides integration. Third, I would say, is around all the things we have done on process optimization. We just came out with ITSM enterprise, CSM enterprise. Again, our goal is not to compete with [ Lotus ]. Our goal is process optimization for ServiceNow processes, and that is now native in our platform. So when you look at integration, AI as well as all these capabilities, including process optimization, and then we did user experience technologies, that's where the platform has changed significantly. I tell my team, if you did an MRI on the platform in 2017 versus now, it looks like a very different platform. So that's how I would say it.

It's funny that you pick that 2017 time frame because the other thing that I've noticed is it feels like the pace of innovation in terms of the number of major deliverables that you've brought forth since that time frame has actually accelerated. So is there something that you specifically did to change that pace of innovation?

Yes. I would say -- and I'm sorry, I have some noise going on outside. But I would say one of the things is, if you know, we started with IT. And under Frank and Scarpelli with the teams that were here, they said, let's go after HR and customer service. And you remember that and we came out with scope applications in 2016 where there were a lot of skeptics on what's going to happen here. Because this is a crowded space, how do you differentiate against HCM and all of that? And Sterling, we shared at our Financial Analyst Day a couple of weeks ago, in Fortune 500, 35% of CHROs now rely on our HR products, our employee experience product. That's a pretty big deal for a product that is only 4 years old. And similarly, for customer service, we have north of 1,500 customers out of 7,000 who are in various stages of rolling out customer service management products. So that was the beginning. And then what has happened is what I saw when I spoke to multiple customers with my team, we saw a pattern. They are saying, okay, your platform releases on every 6 months, but there are so many other use cases we can use ServiceNow for. So we said how do we ramp up the innovation pace? So like IntegrationHub that I talked about, we released in '18. But if you look at 2020, we had 7-plus new products. In 2021, we'll have 5 or 6 new products. So our goal is that our platform provides all the core services for us to create these applications, Sterling. And unit economics are awesome. When you look a ServiceNow's R&D cost as an expense to the revenue, we are still somewhere in the $0.16, $0.17 to $1, despite having this many products because the platform provides all the core services to be able to create new products And that has accelerated significant. AI is built in, user experience built in, integration built in. And you have -- if you have 24 engineers, we can create a 1.0 minimum viable product in 6 months.

That also brings a good point. You talked about that everything is built on that single platform. I want to make sure that I don't confuse that in its original instance, ServiceNow was built on a single-tenant architecture versus a multi-tenant. Does that even matter? And how has that evolved to meet the kind of scale that you deliver today?

Yes. Sterling, I would say one of the biggest things that I give Fred and the original engineering team credit for, they mainly used open source stack with Java, MySQL in the early days. And the problem they were solving for, because every smart person came and told them, you are wrong, you need to be multi-tenant like company X or company Y or use no SQL or whatever it is, okay? What happens is, if you think the strategy that Fred and the team had is that eventually, we will create multiple products. So say you have a large customer like AT&T. I'm just going to use the name. They are one of our large customers. AT&T will, over time, use multiple products. AT&T has a massive scale. Sterling, what you run into in a multi-tenant architecture is a noisy neighbor problem. If AT&T is your neighbor and say you're Pepsi, you are going to complain about AT&T using up all the resources. ServiceNow having multi-instance architecture is the best technological answer for multiple products being used by a customer where they do not care about neighbor using up the resources for CPU or memory or read replicas, whatever the case might be, number of queries you can do. We don't tell it's like Google. Some of our SaaS peers say, hey, don't do this query because this query is going to take forever. We say, if you want to do this query, go ahead and do this query. You're going to impact your own instance; nobody else's instance. Does that make sense? So I think multi-instance architecture and you know, Sterling, despite having multi-instance architecture, we have world-class gross margins from a SaaS perspective.

Exactly. And that's where -- the next question is, how are you able to deliver the amount of innovation in terms of the number of not upgrades but the idea of CICD in that environment? Why isn't it more costly? How are you able to deliver the kind of gross margins? Because it would seem like you -- it would have a higher cost of management to deploy each new introduction.

So I would say, first of all, having this single platform architecture gives us economies of scale. So say if I'm expanding in Korea and semi-infrastructure providers are the Dells or the whoevers of the world, it is the same data center architecture in Korea as it's in the United States and as it's in Australia. So from a simplicity of the architecture, it is the same architecture everywhere. Now that's one thing. You asked the operational question. We use ServiceNow to automate all the data center operations as well. So that it is truly automated, and our goal is to be as much lights out as possible. So when you have harmony in your architecture and not diversity, it's the same footprint, which allows us to get scale with our suppliers of infrastructure. The software is ServiceNow. It's a standardized open source software, and the team does an incredible job of running that in a 24/7, 365. And even if there is an issue, that a particular, say, a mean time to hardware failure, something that goes wrong for a particular customer, it can be picked up very fast by the other. But having that single platform gives us simplicity from how we deal with our infrastructure providers and how do we operate ServiceNow instances across our farm from San Diego to Sydney. Does that make sense?

It does. It does. You touched upon AI, which has been such a hot topic across a lot of different software companies. How are you integrating AI into the ServiceNow platform?

So our goal always has been -- we were very aware that there will be some start-ups that will say, hey, we can provide AI value on top of ServiceNow data. And great, if they can do that. Our goal was to always build AI in the platform. So for example, when we acquired a company that did chatbots, virtual agents, we said, this is not going to be an off-platform service. This will be built into ServiceNow platform that can be initiated by customer with a few clicks or steps, okay? Similarly, when we acquired a company that did Natural Language Understanding really well, we build that into our ServiceNow platform as an NLU service. We recently acquired Element AI, as you know, and they have some really cool technology, so the team is now figuring out how does that become part of the platform. What I didn't want to do on behalf of my team for our customers is create that complexity on them that, hey, if you want to use AI, it is some big scientist name. It's some additional thing you have to do on the side, and you need some really smart people to figure that. No. Our goal always is to make it easy for our customers to turn on AI services, and it's still a journey, so that they can leverage value and have incident deflection or customer service-based deflection using AI or AI search or AIOps built into the platform.

So actually, I'm glad you mentioned deflection. I think at Knowledge21, your big user conference, I think you had actually demoed in a video, I want to say, for the State of Florida, an idea of deflection. For those that aren't familiar, can you just give an overview? What is it actually doing? And how is it saving them?

Yes. So let's take an example. We can relate to that example. That's why I used that example. Thanks for watching it. Florida Department, I think, if I remember correctly their full name of, Highway Safety and Motor Vehicles. That's the agency. And if you go there today, there is a chat icon there. Typically, high-volume requests are always make up 80% of the requests when people ask questions. How do I -- if it's an organization, how do I reset my password? How do I request a laptop? Or I'm mobile, how do I request a terminal? Whatever it is, okay? What you find is 80% of the requests are high-volume request. And our customers are not able to keep up with all these requests coming into their organization. So one of the things we can do via machine learning is we can figure out what were the high-volume requests, and then those becomes a topic or topic recommendation that you can automate. So for Florida Department of Highway Safety, how do I create a new registration for my new vehicle? Or I'm moving, what does it mean for my license? Or whatever the case might be. So when you go to their chatbot, you will see it comes up with those topics. And most of the topics are used by most of the people most of the time. Similarly, if you use Disney+, you look at what are the requests are. They are all in the similar bucket that the digital subscriber, my 15-year-old, I want to watch Mulan. What does that mean? So most requested topics can be deflected using machine landing. That's it. And we monetize that, Sterling, via ITSM Pro or CSM Pro.

Does that then -- is there a way that you're deflecting? Or can you take somebody from the voice side over to that chat side with the same deflection?

That's a tough area for us yet. We have not gone into what the fine lines of the world are doing on the voice side. If somebody just directly makes a call, we typically go via self-service channels of you come on a digital front end or digital front door, and then you try to either search or you try to do a virtual agent or a chatbot conversation. We integrate well with, of course, the fine lines of the world and every Genesys of the world and so on. But we have still gone on the self-service channels such as mobile, web or a chatbot on mobile or web but digital front door.

So once you determine like one of those deflections for high volume, do you then integrate with some of the other chatbot, I think the live person or Genesys or others that are out there? Or do you actually help create that yourselves?

So we currently integrate with all the call center technology providers per se. And yes, the goal is if you cannot answer, how do you deflect to a live call? Absolutely. That's part of the product. We even have support for Amazon Connect, which is currently getting a lot of traction in our installed base, so yes.

So we've talked a little bit about integrations. They seem to play a really important part for customer adoption. Are there any particular integrations that really move the needle? I think you touched upon Microsoft Teams.

So I am excited about Teams, Sterling. We have this classic one, Azure, ADO solutions or Jira spokes, and all of that works very well. But what I saw, Sterling, in the last 2 years, and I give Microsoft team a lot of credit, no pun intended, Microsoft Teams' adoption has gone significantly higher in the largest enterprises. So largest enterprises for communication and collaboration, not engagement level. Communication and collaboration uses Microsoft Teams a lot. So the idea, I think, 2 weeks ago at Knowledge21, Levi's did a case study on how they are doing more deflection using their Microsoft Teams and ServiceNow integration. So we are seeing Microsoft Teams adoption at an employee level, more than developer level, significantly high, and ServiceNow just seamlessly works with Microsoft Teams. So I am really excited that a super relevant conversation on Microsoft Teams, our Azure integration, AWS integrations, GCP integrations, those are resonating really, really well.

Yes. And that's another great segue. So as you think about the AWS, Azure, GCP integrations, what is that giving the customer? What value-add do they take from it?

I would say their processes were sometimes very fine-tuned on the IT side, both service management and operations management side, for their private cloud or their own data centers. And say, if you have a CIO who has migrated 20% of workloads to Azure, they will ask me a simple question, CJ, I'm moving SAP from on-prem to S/4HANA in Azure, can you tell us what you can do? And we have all these best practices on which ServiceNow products can help you enable your migration as you take SAP from on-prem to S/4HANA in Azure. So whatever workload-specific things are happening, sometimes like SAP, like we migrated S/4HANA in Azure, and ServiceNow was used in multiple areas to make that migration help not only accelerate, but once it is in production, you still have incidents coming out of S/4HANA. You still need to monitor what's happening in S/4HANA in Azure. And our Azure integration with ServiceNow really helped us out. So I would say, whether you're moving physical to physical workload, physical to virtual, virtual to container, whatever the journey you are on for whichever public cloud you are trying to migrate to, ServiceNow still becomes -- there was a time, Sterling, as you know, where AWS could have created an incident management solution and go after our ITSM. But they realized that ITSM was the gold standard for incident problem change management, and then they had to work with it because the CIO still relied on ITSM, even if their workload moved to AWS.

Absolutely. We had a question come in from investors said, are you starting to see more competition in your core IT solutions coming from ERP vendors like Salesforce, SAP, Workday, et cetera?

On core IT, definitely not from SAP or Workday. I have seen some things that come about Salesforce and Tanium partnership, great, if they're trying to do something there. Tanium integration and ServiceNow integrations work really, really well. And if they are going after a particular use case, I do not see them in the customer because ServiceNow is the platform. And if they are trying to solve a particular use case, that's great, but we are typically almost always used as a platform.

No, that makes sense. We touched a little bit upon digital transformation, and that usually brings with it kind of the idea of automation. So as a workflow platform, it seems natural that automation would be a critical component. So how is automation being factored into the ServiceNow platform?

I mean that's what we do, right? Workflows are way of automating tasks. Sterling, you've followed us, you know that that's what we tried to connect systems, departments on how work gets done. And in digital transformation, if I have to simplify, I usually have 2 conversations with the season. Conversation number one, for our existing businesses, we need to digitize more. Where can ServiceNow help? Customer service, employee experience, our IT transformation, great. We are reinventing or creating a new business model can ServiceNow help. And we will say, potentially, with maybe like Disney+, with customer service or something like that. So typically, we find that it's usually -- they are focused on optimizing their current business model or creating a new business model. But one thing is always constant. They are still trying to elevate their technology to either modernize or transform technology, and we are right in the center of that. So there are very few digital transformation conversation with our ServiceNow being part of it. Maybe in manufacturing, we are trying to reinvent supply chain process, and you may not have ServiceNow as a key part of that platform. But sometimes we still work well with SAP, 1 supplier onboarding or something like that and still works.

I guess, also, thinking about task orchestration because I can think of a number of different ways that you can do it, everything from PowerShell scripts to others. I just want to know how that kind of automation would be incorporated into the platform.

Yes. So we are known for -- if you look at the primitive of our platform, it is the task, the notification, the approvals, sequencing of task to get some work done, and that's what we do. And then as we have evolved in areas such as DevOps and others through CLI interface, if you need to get something done for a cloud management or other, we can do that, too. We like discussed, we added integration, so that you can orchestrate task and do something in Microsoft Teams or something. But in addition, we are going to add RPA in beginning of 2022. We acquired company called Intellibot, and RPA will become part of our automation platform. So I consider, on behalf of the team, they have a nice term, we will become the automation hub, whether it starts for orchestration, integration with third party and RPA, so that you can model your business processes and digitize them for your enterprise.

Where does that target of the users? So in other words, I can see hard-core IT workers, say, that's fantastic, can take and run with it. But especially as you get out into HR and some of these other areas, will those workers have the capabilities? Well, in other words, will have not only the power, but the ease of use to be able to allow those workers to use it?

Yes. That is our always design goal, Sterling. So as you know, we have evolved from it to HR and customer service. But the largest enterprises, customer service, help desk, can use these capabilities, not only just workflow but integration as well as machine learning and RPA. That's our goal that we want to provide all the things in the toolkit for them to do what they need to do.

Next question that came in is, given the adoption of collaboration tools like Teams, so in other words, the discussion that we just had just a minute ago on Teams, what will be the impact on ServiceNow from Salesforce buying Slack?

So personally, from an employee, because we are -- even when we are used for ITSM or HR, we are employee-facing technology. And in the employee-facing technology in our customer base, I want to be very specific, in our customer base, we are seeing Teams' adoption and Teams' standardization at an unprecedented pace. That is not to say that there are certain communities in a customer base, let say, uses Slack. Great. We have integration with Slack, and it works too. So Microsoft Teams will be always the first-class citizen for us. We have great partnership. They are great friends, and we'll do that because that is employee-facing communication and collaboration technology, and we will still integrate with the Zooms and Slacks of the world despite whoever is acquiring them.

Next one that came in is, how has your relationship with your existing partners changed as you've expanded into new areas like HR and finance and DevOps?

I would say our relationships are always improving. So one large global SI, I would name them, they started in IT, and they have the biggest practice on the IT side. And now they are evolving into HR as well as on the customer service, but it's slow growing. Whereas somebody like Deloitte is very strong in HR. We have many joint customers on their HR transformation practices, and the team does a great job. So Deloitte has become a premier provider of our HR practice as we have moved forward. Punit and the team do a great job. And EY, KPMG, DXC, IBM, depending on where they are on their journey, depends, but it has not slowed us down. It is just if you have a big customer service practice with a vendor A, can you still have a practice with vendor B? Sure, it does, right? If you look at old Oracle and SAP, Deloitte probably has the biggest Oracle and biggest SAP practice. So we don't believe that somebody else has to lose for us to win with the SI.

Let's switch gears to talk a little bit about observability. What attracted you to light step in terms of the option?

So I got, Sterling, today in a lot of investor calls many, many questions on this topic. So I'm going to -- I have learned now how to give the answers in a precise way. So number one, phenomenal team. It is a phenomenal team. They started at Google. They saw gnarly observability problems for Google at scale. And I will bet on that kind of a team all day long. So that's number one. Number two, they were pioneers in creating the OpenTracing standard, which became OpenTelemetry. So they have a platform that can scale, but it's also cheaper. So that's number two. And number three, they were focused on modern distributed workloads like a Spotify or a Twilio or a GitHub, those are their production customers; or Roblox. When you have those kind of customers in production, you know that they saw some gnarly challenges when it comes to scale. And that's it. So we want to take that technology in ServiceNow's customer base. And four, say chase.com, if there are big observability problems, we want to offer that as an alternative to see if Lori Beer would be interested in it.

Always interested in increased observability, that's for sure. But I think where investors are trying to wrap their heads around is how far into each of the APM versus infrastructure versus maybe logging -- those have been the 3 pillars of observability from what investors understand. How deep do you see ServiceNow going?

Yes. And Sterling, all those 3 things you said spot-on. Everybody has started somewhere, right? And they are all our ecosystem partners. And for our ITOM product line, we work with all of them, Dynatrace, a phenomenal partner to work with and so on. So -- but everybody has their routes. Like we had with ITSM, Datadog has on infra; AppD and Dynatrace on -- has on app monitoring; Splunk has on logs, and they are good friends of ours. What these guys have done, though, Sterling, is that they took a disruptive approach for modern applications where all these things come together, not only tracing but tracing and metrics. Rather than a siloed approach, and I go to one Chrome tab or Chrome browser for metrics, then I go to another one for traces, then I go to another one for logs. They have all in one place. So you really can figure out how your application is performing and behaving when something changes and figure out what's the root cause of that change and fix it rather than me having swivel sharing across different technologies. And that's why they can scale to a Spotify level or a Roblox level gives me so much optimism. And like, Sterling, you were saying before the meeting started, people are very skeptical when we entered customer service market. And these guys do it. It's 10 years too late. And why are they doing this? And I'm sure some of that skepticism was fair. Our goal is always, if we can solve a customer problem with our amazing go-to-market team, we will always at least put it out there that we can solve this problem really well.

Got it. Next question that came in is, can you talk about the security business? And are you actually at a point where it can be used to find vulnerabilities?

Yes. So we used to have security as a pillar. And then at some point, Sterling, we realized, I want to say, 2018, that 2/3 of security teams are reporting to IT. So we started rather than having that as a pillar, just made it part of our IT core portfolio, our security product. So I have security incident response as one product line, and the other product line is vulnerability response. And these 2 product lines are growing nicely. Security business, as you know, it's a crowded space. Everybody has to say something or the other that they do in security. Our goal has been always to do incident response and vulnerability response for the largest IT organizations at banks, health care, government agencies and so on. And we are doing a nice job in that. We are not into like threat detection or threat prevention, but we are helping organization dealing with large inflow of incidents or large flow of vulnerabilities and which assets they should care about. And we will continue to do that, innovate in that add, add machine learning capabilities over time and work within the ecosystem really well. What we want to do for security incident response and vulnerability response, what we did for ITSM. So really focused on that aspect, yes.

Yes, it makes sense because of the workflow, but you also have the unique asset of your CMDB as well, which is such a key component what a lot of these security products want to integrate into to check compliance and policy.

That's 100% correct. And that's why when we say, okay, we do vulnerability response, the CISO will send you and 50 other people. But because of CMDB, now which assets you are going to patch first because you have physics limitation. You can patch everything. You cannot patch every vulnerability you have, whether you are a bank, health care or intelligence agency. So with CMDB, you can prioritize which assets you care about the most, and those vulnerabilities need to be passed most, and where does that orchestration happens between security and IT tubes. Because security will highlight, but it's typically the IT team that has to carry out the task, typically.

How does the CMDB extend into the cloud? Because we see a lot of talk about CPS and cloud postures, security management. It seemed like CMDB would be a natural checkpoint to help with the configuration management within the cloud as well.

Yes. So we have evolved our CMDB to have full support on our discovery product side for micro services, Kubernete clusters, whatever the case might be on the modern architecture, functionless support, this and that. And customers are still trying to figure out for all those assets how do they feed data into CMDB and keep it current because some of these assets come up and go down really fast. So that's something we are working on, on a constant basis, but we are now seeing that some customers who migrated to public cloud are actually putting those kind of services into our CMDB.

Makes sense. And last question that came in, do you see yourself in observability being more of a niche vendor, being -- just serving a key piece like you've done in security? Or do you have bigger ambitions? So in other words, they're asking, does a company like Dynatrace actually become a head-on competitor rather than a partner?

I would say, for now, we feel that there is enough TAM for modern distributed workloads. And we believe that, that problem we solve really, really well. That -- and Dynatrace, as I said before, is a good friend, good partner and will continue to do so. But we are very focused on this container-based micro services-based architecture that are like the Spotifys of the world and others in enterprise segment. And that's the problem we want to solve. Will we do something more in the future? Sure. Potentially, we can. But let's grow this out really well in our customer base first.

Yes. One step at a time. Excellent. All right, CJ, with that, I think we've come up against the end of that. I really appreciate you taking the time to join us today. Thank you so much.

Thank you, Sterling, and thank you for inviting us.