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

Okay, everybody. Why don't we get started? I'm Karl Keirstead on the UBS software team. On a serious note, I've been taking an informal poll of the best-dressed man at the conference and CJ, as it turns out, I think wins. I trust everybody would agree. And I clearly need to raise my game for next year's conference. But we're super happy to have you. Thanks Team ServiceNow IR team for having CJ come.

Why don't we kick it off with a bit of a feel for how the world feels out there? You were telling me this morning that you've had a number of CIO conversations of late. I think you were even saying on with 1 or 2 this morning. So this is a real-time view of how the environment feels.

Absolutely. So first of all, it's great to be here. Great location. And I am delighted to be speaking to you about everything from business to ServiceNow, including favorite topic, AI that seems to be on top of mind for many of you.

We'll get there for sure.

So in terms of the business environment, ServiceNow, typically, every year, works for many industries. As in, we sell to many industries and as you know that we are in the enterprise, large enterprise and very large enterprise segment. We do not serve the S of SMB. So all my comments are typically enterprise, large enterprise, very large enterprise including governments. And my thing, Karl, I can tell you is, the demand environment in speaking to the CIOs in 2023 feels very similar to 2022. So similar set of concerns, value-based selling that, hey, if I buy your software, how fast can I get to value, so what's your speed to value? What's the cost? And what is your software in service of? And fortunately, for ServiceNow, I'm very proud of ServiceNow team on how we have done in 2023 despite the environment being similar to 2022. And the CIOs, when we tell them that ServiceNow is about productivity, automation and others, it continues to resonate. They will sometimes ask for hard dollar savings with our products like asset management, sometimes it is about risk with financial services type of clients; with governments, it's about our entire platform and what we can do for digitization in the government sector. But typically, I find including the calls this morning that we are still relevant because, one, we are a platform company and despite the tougher environment compared to, say, 2021, our workflow automation positioning in the name of efficiency and productivity continues to resonate with the CIO.

And CJ, if you're getting any early glimpse into not so much the ServiceNow spend, we'll leave that to Gina to guide to. But broadly, IT spend next year, does it feel like it's at least going to be up? Are we remotely back to the 2021 days or still below that?

So definitely not 2021, I can say that. And in 2021, the approval cycles and all that we see today for any kind of transaction to go through, any purchases to go through, 2023 is definitely very different. Coming back to your question on 2024, from what I can tell, so ServiceNow, as I shared, sells to multiple industries, public sector, which is enough -- big enough vertical for us, global public sector, not just U.S. federal, the U.S. state and local. I was in the U.K., talked to public sector customers last week in the U.K., Australia, Japan and other places. Public sector seems -- yes, the budget, the digitization, serving the citizens will continue to grow. Now you look at all other sectors from banking, retail and others. Just based on the sample size and the kind of customers that I speak to, which tend to be large, single-digit growth in IT budget is what I see. Some could be 3 to 5, some could be 5 to 7.

Well, that sounds like a modest improvement from this year.

Yes. Hopefully. Hopefully.

Okay.

But no CIOs come and told me, because you asked me the CIO question, which is our prime buyer, CJ, I'm negative next year or I'm flat. Most are like, hey, I've single-digit percent. I may set it aside for X, Y, these 3 key initiatives. But most of my conversations tend to be which platforms the CIOs are going to invest in, whether it's us, whether it's ERP platform, productivity platform and so on.

Well, CJ, the fact that you're describing the environment as such and yet ServiceNow is still putting up 20-plus percent growth is extraordinary. A growth that's faster than almost every other, if not every single large-cap software firm. So congrats on that performance.

Thank you. Our team has done a great job.

Yes. So within your business, you, this past quarter at least, had a phenomenal federal government quarter. Let's talk about that for a little bit, 75% net new ACV growth. So essentially, how did you pull that off? Like what's the opportunity? What's the problem you're solving in federal government that makes that such a robust area for ServiceNow because candidly, we're not hearing about the same degree of public sector growth from most other software companies?

That's correct. And Karl, you're absolutely right. because when we posted the numbers, some folks said, hey, that's a big quarter for U.S. federal, no surprise. And my point was exactly the same. Well, every other software vendor could have said the same thing, but they did not, compared to our results. So first, I'll start with the platform and our products. Our platform was ideally designed for public sector. And for me, the public sector, it's not just federal government, it is also state governments, local governments as in cities across the world. So our platform is very flexible. You cannot say, hey, I can do Salesforce automation for public sector. That doesn't make sense. Public sector doesn't have salespeople, right? They don't even have marketing really. They may attempt to market themselves, but they don't have big marketing department. But what they are focused on is digitizing services, making it available to their end stakeholders, including citizens, and ServiceNow is like perfect for that. So not only just in Q3, we have consistently highlighted since last Q4 that U.S. federal public sector has done really well for us. In addition, though, our state and local business in the United States has also done well, has also done well in Australia and also done well in the U.K. And we are now trying to expand that to Germany, Japan and the large GDPs that exist out there because our platform is designed, we have invested in certification, data residency-related requirements, everything that a government needs for their citizen's data. And one of the things, Karl, that many folks don't know because we are a multi-instance architecture, if a government wants, including federal government, that we want to use ServiceNow on-prem. If you're multi-tenant, you cannot do that. We can literally put ServiceNow in a box. And for one big federal agency in September, not only we have put ServiceNow in a box, we put our gen AI offering, our own LLM in a box and gave it to them. And I just visited them 2 weeks ago in D.C., and they are absolutely ecstatic that they can do that and run it on-prem for their specific government agency. So there are architectural advantages besides our investment in platform and certification.

And CJ, if this is working so well in the U.S. federal government sector, you hit on a little bit in the answer to the question, but is there not a bigger opportunity to execute similarly in the U.K., France, Germany, rest of Europe? And would you characterize that as an important investment area in 2024?

Yes. I would say, not only in 2024, but for foreseeable future, I would be disappointed given the demand for our platform and services, if our public sector -- global public sector business does not grow higher than our revenue rate. So I have strong conviction that our new logos, think about a number of cities in the United States or number of cities in Germany or U.K. or India, number of states at all these places or provinces as in Canada. So now that boils down to how much investment we are willing to put in go-to-market in those. So we have prioritized 5 countries for 2024, where we are going to invest even more in state and local as well as federal around these countries, which are typically top 5 GDPs, except China, we don't do business in China yet. But that's where we are focused on. So we are going to increase our go-to-market investment.

So that's one key growth driver. Let's talk about the AI growth driver. I think everybody would find that exciting. So maybe, CJ, can you give us an update on the ITSM Pro Plus traction? You sort of teased us a few weeks back with some initial, I think, 4 customer wins. But if we kind of extend it through the end of November, how does the first few months of traction feel? And perhaps it will take you a full quarter to be able to share some real metrics. But is there anything you can share with the audience about the early traction?

Yes. So at the Investor Day in May is when we first announced our gen AI strategy, and it is our deep conviction that large language models are not differentiators, okay? LLM is not a differentiator. But when you make LLM specific to use cases that you can run efficiently, that is a differentiator for gen AI. And when you combine that gen AI plus ServiceNow action, so you look up something and then you act on it, that combination is a productivity multiplier, not productivity enhancer, okay, because machine can talk in human language and then you kick off a workflow. So that's the key thesis that we had that we revealed at Investor Day in May, just 6 months ago. On your question, the 4 multimillion dollar deals that we signed, even though we released the product on September 29, which was literally the Friday when our quarter closed, and yes, we worked a few months on those deals, that was very encouraging for us to be able to say that because my team and I, we were there in 2018 when we launched our original ITSM Pro SKU, 2018 September, the exact same thing. And we did not make as much money that time on ITSM Pro as we made on ITSM Pro Plus, so that's number one. We are second month in this quarter. And as I shared at our earnings call, in January, we will give you the data points on how we are seeing, across which personas our gen AI is resonating, what is the price uplift we are really getting, right? We have been transparent about Pro giving us 25% uplift compared to 50% list price uplift that we asked for. Here, we have asked for 60%. My intuition tells me we'll get somewhere in the 25%, 30% range if we can convince and all that. And so far, I am having a lot of conversation -- because you asked this in-quarter question, I'm having a lot of conversations around gen AI and customers are seriously evaluating it, and we will share the details in January earnings call.

I want to unpack a couple of things you said. So one is just about your ability to monetize AI and LLMs relative to other software companies. So I'll put him, he's not here, but on the stage a little bit, but Aneel Bhusri at Workday is in one end of the spectrum arguing essentially that the idea that software application vendors can utilize a third-party LLM embedded in their product and just turn that pricing down. I'm paraphrasing a little bit, but he argues, good luck trying to do that. Everybody is doing that. It's table stakes. It's going to be very hard to monetize. So if that's true, CJ, why do you think ServiceNow will be able to monetize your AI product when other software companies might struggle? What's the key difference that gives you an edge in monetizing?

So I personally feel, Karl, it's an insightful question is it is expensive to do R&D on generative AI, right? These are expensive R&D personnel. We are fortunate we have many of them. But it's expensive to build gen AI products specific to the use cases of ServiceNow. So that's number one. Number two, we also spend lot of energy on creating LLMs that are specific to use cases that can be run efficiently so that there is no gross margin hit, right? So we are very careful on how we orchestrate that aspect of it. And then, I would say, for us, given that we are not a system of record, we are system of action, generative AI plus ServiceNow workflow improves your productivity significantly, and that's why I use the word productivity multiplier. That's what we have to convince our customers that you need to pay extra for it. Because if you did not use gen AI, by default ServiceNow enhances your productivity because you're streamlining the workflow. But now with gen AI, that workflow can be accelerated. As Jensen calls it, accelerated computing that you can get things done faster and sometimes exponentially faster because the machine can understand what you are trying to do. And that is the price to be paid for it. But only pay if you are seeing the productivity gain. And that's why, if you remember in the July earnings call, I said that we would take 10% out of the -- whatever the productivity gains to be had. Now productivity is a complex science, a big research topic. So you have to work with customers, get them comfortable sharing their data. But I am convicted on behalf of ServiceNow that this is truly a productivity multiplier for us in ServiceNow use cases versus a system of record companies, which says, okay, you're storing data, Karl lives here in Southern California, this, that and so on. I don't know when is the productivity enhancement coming.

Okay. That's interesting. So let's talk about maybe another point you made, CJ. Maybe it's broader than ServiceNow. And that's this idea that the LLM is not the differentiator. This has been a key topic for everybody in the audience because we've had the OpenAI drama on our minds for the last 2 weeks. So I'd love you to articulate your views a little bit. Obviously, Microsoft with OpenAI, Oracle with Cohere, AWS and Google with Anthropic have made bets on essentially close source LLMs. So it's interesting that you're conveying that you don't think that maybe is the right strategy. So what else for you guys? What are the models underpinning ServiceNow's AI products? Are they generally open source, homegrown?

Yes. So we absolutely believe that LLMs are not differentiated unless you make it specific to a use case, right? I'll stick to that. But number two, yes, we are using open source LLMs or whichever LLMs that are in service of our use cases. I'll make it very specific. For text to code, I don't need a 175 billion parameter model for ServiceNow text to code. So we looked -- we used Hugging Face open source model called StarCoder. Our engineers train that because we know how to write ServiceNow code. And it's a 10 billion parameter model that can run on NVIDIA A100s in our data center at a much cheaper cost. Awesome. Because that's what we need to -- I really don't care that a large LLM that can help answer your history question, sports trivia question, why do I need all those parameters and natural language, it should be specific to ServiceNow. So wherever the best models are, we will use them from open source community, train it for our use cases, run it efficiently, and when the model is smaller, the latency is lower, so the end experience is also good for end users. Right now, so for one of the Q&A within ServiceNow's context, we are using NVIDIA's NeMo. Jensen said, hey, try this out and my engineering team will work with yours to show you that this NeMo, which is another open source model from NVIDIA is an awesome model. We tried it out. He made it work with us. He literally put his engineering team saying ServiceNow is the priority because he knows that if that works, so many enterprises benefit, right? So he uses us as a channel. It worked, and we are using that for Q&A. I'm currently evaluating Llama 2 from Meta for certain multi-turn Q&A. Fantastic. So I am, on behalf of ServiceNow, just looking at my use cases, how we can enable them cost efficiently with smaller model. I'm not trying to solve this, hey, let's have a big model, and that's our strategy for any use cases.

Okay. Just out of curiosity, for your engineers, are they auto-coding with a forked version of StarCoder? Or are they also using GitHub Copilot?

So when it comes to -- we do a lot of coding in Java for our back-end, so we are using GitHub Copilot for a certain part of my engineering organization, and it's working well. We're getting single-digit productivity gain, but that's good enough for us. It's worth the price and we can innovate faster. So that's for my developers. For developers who write ServiceNow-specific code, like our IT teams for our customers, they are using our text to code because they can use now ServiceNow. They can code in ServiceNow faster. And we just released another big innovation on November 16, which was 10 days ago, where we now do text to workflow. So you can just say, hey, I want a workflow that does this, blah, blah, blah, and it will create this beautiful workflow that visually you can see with a code behind it that you can edit.

Got it. CJ, let's talk a little bit about the competition on the AI front. I think everybody in the audience knows of some private AI start-ups. There's a number of them. Moveworks, Aisera are all pitching that they can do ITSM workflow automation using next-gen AI. So how often do you encounter those start-ups? And is the pitch that your AI workflow products are better or that there's just inherent advantage in buying them from your fully integrated workflow automation provider in the first place? So in other words, a platform sale will always beat the point solution then.

Correct. Here is my very simple explanation to customers when they ask me the same question. They do ask me the same question. I say, okay, so let me make sure I understand this. If you're going to make a bet on start-up, whatever that point solution company is, to extract data out of your ServiceNow instance, that data now goes to that company wherever they are running, then they're going to run some AI model on that thing and then they're going to give you a result. And you now just created risk inherently, you don't know because some of these companies will claim we are taking data from 100 customers and training the model. We have always been very trustworthy with our customers, whether they are banks, health care providers, government and so on. Your data is your data. We will run the model on your data. We don't co-mingle data from any companies. And you don't want your data to leave ServiceNow Cloud, which is secured, including the government cloud. And we -- it is our job to make sure that your end users have the best experience because we know ServiceNow data, we know ServiceNow processes. If you want to do and take this risk, go ahead and do that. I don't know why you would do that. And we are seeing that some of the customers, Karl, who went in that direction are coming back now because they realized that, that makes no sense to do that.

That's compelling. Let's talk maybe -- let's move beyond AI to some of the core products. So let's talk about core ITSM, still the majority of your revenues. You've got some great...

It's not majority anymore.

Okay. Interesting. Thank you for that. Great momentum.

Because I'm assuming majority means more than 50%. So it's not.

Okay. But one risk, I think, if you were to try to assemble a list of risks around ServiceNow, it's not a long list, trust me. But there have been worries that the core ITSM is fairly highly penetrated that a lot of enterprises you talked to, including UBS, by the way are major customers. So maybe you could address like how much runway is left on ITSM in the Global 2000?

So I don't -- I think it was 2018 or '19 with our previous CFO, Mike Scarpelli and I, we said, we are not going to pay a lot of attention to Global 2000 as a number, as in 2000 number because that's very limiting, okay? So our threshold, Karl, for -- our threshold is ServiceNow is an ideal platform if you are more than $100 million in revenue, okay, and you have 1,000 employees or above, then our platform is a true platform that you can leverage for multiple use cases. We are right now somewhere between, if I have to just generalize, 25% penetrated with ITSM for specifically that threshold. Now there are a variety of numbers. Our Investor Relations team will say, hey, it's even less than that. But just in general ballpark, I think it's 25,000 to 30,000 companies who fit that profile that are not in China, because we don't do business in China, and we are in 8,000 plus. So I feel that there is still a long runway for new logos on ITSM. In existing ITSM accounts, we have 3 things going on for us. That's why ITSM continues to be a growth business. Number one, we are still only 40% on Pro. So we still have 60% left to go on Pro. Pro Plus, we just released, maybe we create incentive on combining Pro and Pro Plus in 2024, which we'll do. So technically, it makes sense for our customers to go to Pro Plus. So that's the second aspect. And third aspect is these existing companies as they create more digital services, they increase their IT staff, they will need more ServiceNow. So we do some natural expansion, which we shared that our seat count has gone up 10% within Pro customers on renewal. So I think about the big picture, as in there are still tens of thousands of companies that don't use ServiceNow ITSM, big, big white space in global public sector, parts of Europe, many parts of Asia, Latin America, the commercial section of United States. So I am still -- there is lot of headroom. So that's why I don't look at Global 2000.

Makes sense. Okay. Insightful answer. I'm going to ask about another product segment, which doesn't get as much love and attention as ITSM and AI. And that's the ITOM space. You do disclose it in your queue. It's 12% of sub-rev and it keeps chugging along at 20%, 25% growth despite the fact that broadly, infrastructure management firms, and there's a number of stand-alones have gone through a big growth rate deceleration, your business has not. What's happening? This could be a long conversation, but maybe summarize what's happening in that ITOM space to enable -- to sustain that kind of growth?

Yes. So there was -- back in 2017, 2018, when public cloud started accelerating as in people moving workloads to Azure, AWS, GCP and others, there was a concern that ITOM will not be relevant at that point in time. This is 5 years ago.

Yes, I remember that.

And I give credit to our engineering teams and product teams. We basically said no matter who the CIO is, whether you are moving from on-prem to one public cloud, on-prem to two, on-prem to three, however, many, we will make sure that ITOM helps you with visibility across any cloud estate. So we built integration with AWS, Azure, GCP, even Oracle Cloud, even IBM Cloud, so that it doesn't matter who you as Mr. or Ms. Customer, you will have visibility into your multi-cloud estate. That allowed us to grow because a CIO will say, "My God, CJ, I moved 20 applications to Azure, maybe 50 to AWS and experimenting with GCP on X, Y and Z. But gee, it's nice to know that I know exactly where my assets are." So that's the number one driver for ITOM. Number two, once you have your applications and assets in all this multi-cloud, then you want to know what is the health of those assets, right? Are they doing well? Are they going down? How many Kubernetes cluster you are using in Azure or AWS? We provide all of that through event correlation, aka AIOps, which has helped us grow also on ITOM. So combination of these 2, visibility and what we call health, has helped us grow that business. And I feel very comfortable that, that businesses continue to grow for 20% plus for next 3 plus years.

It's been a [champ]. We only have 2 minutes, but there's one other big product segment I wanted you to talk a little bit about, and that's, you recently announced that the Creator Workflow segment has passed $1 billion. It's extraordinary. So if you could put your finger on like 1 or 2 drivers of that segment, CJ, what are they?

So the no-code and people who position, companies who position themselves as no-code, there is nothing like no-code. In enterprises, you will have to code. So when you want to digitize processes, right, whether you look at U.S. government, public sector or an enterprise customer, we provide not only comfort and a lot of tools so that you can digitize any process that you want to map with integration that you will still need to do because they're fully API-driven. And then the third thing is a lot of investors, including sell-side ask me this question, are you going to sell this to line of business, our Creator Workflow product. And I said, no, we will only sell it to IT because IT does not want 1,000 flowers blooming when they have data issues with apps being created in, say, marketing department. So we put a lot of governance that anybody should be able to create an app but it is still blessed by IT in the ServiceNow platform. And that simple decision and focus on IT as a buyer for processes to be digitized took off the business.

Yes. Awesome. I learned a few things. Thank you, CJ. I hope you did too. Really appreciate ServiceNow helping to make this event a big success. Thank you.

Thank you, Karl, for inviting. We appreciate.

Yes. Thank you.