Okay, good morning, everyone. Welcome. I'm Mark Murphy, a software analyst with JPMorgan. It is a honor and a privilege to be here with Kate Bueker, who as you all know, is the CFO of HubSpot, as well as Dharmesh Shah, who is the Co-Founder and CTO. Well, first off, welcome to the conference and thank you for taking the time.
Thank you.
Thanks for having us.
maybe you could just give us the 30-second version, to introduce yourselves and HubSpot for the benefit of anyone who's not aware out there.
Sure. I'm Dharmesh, the Co-Founder and CTO of HubSpot. HubSpot is a SaaS platform. It's a CRM platform primarily focused on small and medium-sized businesses. The company is about to approach its 17th year anniversary, and I'm having a better time than I've ever had at HubSpot right now.
I'm Kate Bueker. I'm the CFO, and I am approaching my five-year anniversary in June, and I am also having a great time.
Those are big words because you always have a pretty wonderful time, as long as I've known you. Let's talk about the resiliency of HubSpot for just a moment. It's been a tough period of spending on front office software because it had accelerated so much during the pandemic, right? We've had several quarters now where we've been in a tougher environment. It's been a different environment. HubSpot has shown this resiliency.
When we look at it in Q1, the growth rate was 30% constant currency, and that's on about a $2 billion top-line scale. When we look at it in contrast, we're forecasting these growth rates that range from about 5%-15%. That applies to Salesforce, ZoomInfo, Qualtrics, Twilio, et cetera. To what do you attribute this superior resiliency for HubSpot? Do you see an environment here that is translating into faster market share gains?
Yeah, maybe I'll start. You're welcome to add in. I would highlight maybe three things that I think contribute to the resiliency of HubSpot. I think the first is, you know, I talked about the fact that I am approaching my five-year anniversary. HubSpot is a very different company today than it was five years ago when I joined. We have been on this journey from a single application or marketing automation company to a platform company, a CRM platform company that serves the whole front office. You know, that means a different level of necessity for our end customers. Today, 60% of our customers use two or more hubs, and 45% of our customers are on three or more hubs. It means they're running their whole front office on HubSpot.
The second thing that I would attribute to is a little bit of the current environment. We are in a world where total cost of ownership is front and center in decision-making. For small and medium-sized businesses, there are more and more making platform choices, consolidating a disparate set of applications onto a single platform that can help them drive efficiency, and HubSpot is really well-positioned in that, in that order. The last thing is really around time to value. In this environment, people are also looking to de-risk the decisions that they're making. So they're looking for investments that pay off quickly. HubSpot has, frankly, by virtue of the fact that we've grown organically and we've made some very smart choices around how we build product, is one that is very easy to get started with, easy to use, and that makes a difference in this environment.
broader platform versus five years ago, TCO matters more, and then the, the time to value being reprioritized. Yeah. Okay.
That's exactly right.
You tucked in a little comment on this on being an organic platform, and I wanna go right into that. Dharmesh, you've had this messaging that we love. You call it crafted, not cobbled. There's this quote, "The dream is to have a single, cohesive, uncomplicated CRM crafted by people who care." This is something that I really value, that type of architectural cohesion. We've heard you for many years talk about having an Apple-like type of a design ethos built into the product. Can you help explain to us how does HubSpot benefit? How do your customers benefit from having this kind of contrarian approach, where for the most part, you've really actually been avoiding M&A?
Yeah. To start out with, we don't really avoid M&A. We're actually committed to this kind of cohesive platform, and sometimes those two things can be at odds. The right M&A, I think, would make sense for us.
Yes.
Back to the kinda importance of being a platform, you know, having built everything or kinda organically in-house, within HubSpot, let's talk first about the benefit to customers. As a result of it being one cohesive platform, one database, one set of user interfaces, I mean, it makes it much easier for the customer. For us in the market that we're in, that's really not optional. That's high on the list of what customers want is ease of use and simplicity, kinda reduce the complexity, compared to alternatives, in the marketplace. The other one is a little bit more subtle. The value of a platform to HubSpot, outside of, you know, benefiting our customers, is that it makes it simple and easy for our customers, but it makes it simple and easy for HubSpot to innovate.
As a result of having one unified platform, new shifts happening in the marketplace in terms of, oh, there's this new thing called generative AI that everyone's all excited about. I think we're in a better position to leverage those shifts and innovate faster and iterate faster than we would be if we had a portfolio of products accrued via acquisition over some long period of years, right? It's just harder.
to get everything coordinated and put things in motion. It's one of the things that has me excited. You don't really know these things. You know, I've been in software for 30 years. At the time you're making these long-term decisions to be a platform, I have experienced that it's like, okay, it's just a matter of time. You don't know exactly how those benefits will show up at the time you're making those decisions. Some are obvious, some are less obvious, but we're kind of reaping the rewards of some of the kind of harder work the product team has done over the years.
Okay, You brought up generative AI. I didn't have to do it, but I wanna make sure we're gonna come back to that because that's gonna be topic number one. Before we do that, Kate, I wanted to also sort of acknowledge, I think you had some wisdom in terms of not issuing the multiyear, you know, targets and growth rates when spending was very robust a couple of years ago. I keep saying companies that did that, they're really regretting it now, and they're having to dial that back. We don't know exactly what the longer range plan is for the business. I think what we can see is we can see just a tremendously huge top of funnel.
We can see that happening at the free and starter tiers, and then you've got all this multi-product, and you have what's happening, what's resonating with the enterprise tiers. How do you think about the longer range, the maybe more from a qualitative standpoint or just what you aspire to, even if we're looking way out, you know, something like a 5 to 10-year horizon?
Yeah. It's a good question. We've obviously made a choice to focus on the small and medium-sized businesses, and we did that because it's really a massive market. You know, we're a $2 billion company operating in a market that is, in a couple of years, projected to be $70 billion. We have, you know, we continue to have single-digit market share across all of our markets. We believe that, you know, each of our marketing, sales, and customer service, sort of our core hubs, can each be multi-billion dollar hubs. You know, we think that that SMB market is largely or relatively underserved versus other parts of the market, and we feel really good about our ability to answer the needs of those millions of businesses.
I do recall Brian Halligan actually said this at the analyst day last fall, that, you're super ambitious to build a giant company. You're up to 180,000 customers today. If you're successful 5 to 10 years from now, how many customers do you think you would have?
I mean, I come back to the foundational mission of HubSpot, which is to help millions of organizations grow better. Like, we chose millions purposefully. We believe that we should be a company that can serve millions of customers.
Okay. You know, we've noticed I was thinking back a year ago at this very conference, you had begun to mention that some deal cycles started to extend in Europe. I would say, I spotted it sooner than most companies. HubSpot has a tendency to be, I think, a little more attentive to conditions or kind of better able to monitor it and the willingness to kind of comment on it. When you think back to then a year ago, what was it that allowed you to pick up on the change that was happening in the macro? If I were to ask you just directionally, do you think the macro and business confidence are kind of, y ou know, directionally, which way do you think they're going here in the month of May?
Yeah. Well, I guess I'll start by saying thank you. This is the second comment that you've made, which is complimenting me in front of my boss, so that feedback.
I return the favor.
You know, if I think about HubSpot noticing the trends earlier, I mean, it's not clear that we noticed the trends earlier than anyone, but I do think there's a couple of things that we do in how we approach the business and the sort of understanding of the business that maybe helped us get there, at least with our external comments sooner. One is the company is a very data-oriented company. We consistently look at a large number of metrics internally on a weekly basis, on a monthly basis, to try to give us insights into, you know, what is happening in the business, what are the, you know, sort of potential outcomes that we could see sort of in the next set of quarters.
The other is that, you know, one of our core cultural values is transparency, and that's an approach that we certainly take internally, but it's one that we take externally. It's really that combination that I think put us, you know, earlier than many companies in sort of having the conversation around the, you know, challenges that were emerging over, you know, the last 12 or so months. That said, like, I'm not gonna provide any new update on the sort of external environment. We had earnings just a few weeks ago. What we said on that call and consistently in our call backs is, you know, we continue to operate in a really challenging macro environment. It is pretty consistent with what we have seen since the back half of last year. It hasn't gotten materially worse. It also has not gotten materially better.
When you think about what has happened with tech spending in this cycle, it has been unusual, right? In my experience. Basically, tech spending has had a recession without seeing an economic, you know, recession, at least not yet, knock on wood. Do you think that it could mean that spending on tech, spending on software has to some extent been kind of purged or been optimized, such that, you know, when we get into the next recession that it wouldn't feel maybe quite as bad at this juncture as it otherwise would?
I think it's a really interesting question. you know, we certainly feel like we've been operating in a recession for quarters, even if it's not technically being characterized as a larger recession. What that translates into from a customer perspective or customer motion is two things. The first is that just our customers have a higher bar for decisions, purchase decisions, and that is taking many forms, but we've talked about things like more and more decision makers getting involved in purchase decisions taking longer. We are, you know, losing to inaction, not losing to competition. The second thing that we're seeing is that customers are very focused on optimizing their existing spend. For us, that takes the form of customers cleaning up unused seats, maybe trimming down contacts that they haven't been marketing to.
you know, some of the new functionality that we have in the enterprise versions of our product with some of the permissioning and governance allow our customers who maybe have been managing their business through multiple portals to do some consolidation of their spend. Then we're seeing customers expand their sort of usage spend at lower rates than we have historically. That said, you know, we continue to see our gross retention really nicely in the high, like, hanging in there in the high 80s, which we think speaks volumes of the value of the platform. You know, we feel good about our ability to continue to deliver net revenue retention at sort of a 100+.
Okay. Which is a pretty remarkable feat, and I wanna come back to that discussion of, on the retention in a couple of moments as well. Dharmesh, let's get to the big topic here around generative AI. I have been telling investors that I thought what HubSpot did with ChatSpot.ai, is really the single best example. Like, this is what a SaaS application software company should do to understand the trend and actually go do something about it and code it up and have a demo. A very compelling, a very well done demo. You know, working live right there on the screen inside of HubSpot.
I believe it's in public alpha, you had reached 40,000 users very quickly, also pretty remarkable. I have been told that you personally really did the heavy lifting on ChatSpot. I told like this, kind of starting last fall, you sort of had this obsession with it and really took control. Can you tell us just the story of... You know, what was going through your mind and how did it come together so quickly?
Yeah. I'll start once again by reiterating that the heavy lifting actually happened years ago as we started building out the platform. That's what made it possible to build something like ChatSpot. Now, I've had a personal kind of passion around AI for a long time now, and I've done variations of what is now ChatSpot as early as five years ago. I was going through some old emails, Brian and I had a chat with Sam Altman, who's the founder, CEO of OpenAI. This was June of 2020, talking about GPT.
Wow.
Our question to him is: Where do you think this is headed, right? This is, you know, 3 years ago, which is an eternity. It's like dog years when it comes to generative AI. His statement was that, you know, it's, "We're making progress. It's going to continue that way, and then there's gonna be an inflection point. I don't know when that's going to happen, but it's going to happen." Because they saw the writing on the wall. Obviously they were building it. I've been following the kind of progression of generative AI for a long time. It kinda really hit me as it did, I think, other folks, just the kind of raw potential of something like GPT-3, which was what caught the imagination.
I'll tell you the quick story. I had built something for myself just to kind of tinker with it, and I'm up till two o'clock in the morning every night coding on it. And we had I had 50 users, you know, kind of friends and family and internal HubSpotters, and our goal was, it's like, we'll launch this because I'm a big believer in iteration, and we'll get it to 500 users, like, someday within the first month or so. Two days after launch, it was 10,000 people had signed up for ChatSpot, and two months later now, we have 45,000+ users. And this is a testament, it's less about kinda ChatSpot, but how much the potential of generative AI has kinda captured the imagination. We have all been exposed to it in varying forms, through things like ChatGPT, and I think it's going to be transformative, for both for HubSpot and the software industry at large.
That's a great story, and it's amazing you were talking to Sam Altman 3 years ago. How do you think generative AI is gonna change the experience of using HubSpot? Do you think it's gonna be more of an amplifier? Because you have a lot of people that are trying to crank out marketing copy, they're trying to crank out blog content. Is this actually gonna be more of a tool where you would be able to go in and generate the kind of internally generating reports and analysis using plain English? Because I think I saw elements of both in that demo.
Yeah. Those are both, you know, great use cases, but both discrete use cases. The reason part of the reason I'm so excited about generative AI is for us, not just a discrete set of features we can add to our existing applications, we will certainly do that, but how we can kinda weave it into the platform and make it what we call one of our Primary Colors, part of the HubSpot framework. Just kinda walk you through this, just in terms of the experience of HubSpot, and let's take it from kinda. I'll make this short. I talk really fast anyway, so it's like watching YouTube on 1.5x, right? It's. You know, let's say one of our customers, they have a website up there they're posting blog articles on.
Those articles will be assisted by generative AI within HubSpot. A customer shows up at the website, signs up and says like, "I'm interested." Two days later, they get an email from that company. It's like, "Oh, by the way, we understand you're in higher education, and you're interested in this. Here's a really good blog post that you should, you know, also consider." The beauty of it, that blog post was written for that individual person. It was not published six months ago, two months ago. It was published five minutes ago once we had the context for that user. Okay, great. Now the user is like, "Yes, I'm buying." Internally within HubSpot, people are using the product, and they're like, "Oh, I want.
I'm the head of sales, I wanna generate a report I'm walking into a meeting for that shows us the last quarter's revenue broken down by industry and geo. I wanna see the top five industries so just so I can kinda, you know, talk to the team. Back six months ago, that would have been 30-35 clicks. Now it's just a, you express the thing that you have in your head, and I think this is the biggest shift in software, is that you kinda state what you want. The software figures out how to get that answer, that report, that workflow, that automation, that whatever for you. I think that's going to be a radical shift in people's expectations of how software should work.
You're stating it in plain English like a mere mortal.
Yes.
That's what you were demoing.
Not the odd thing, but the other really cool thing is that it doesn't have to be English, right? It can be any language...
Oh, yeah.
As long as you can represent the thing you're looking for, the large language models are roughly ambivalent.
Yeah.
in terms of which language you use to articulate your particular need, and then we can kinda take that and translate that into whatever's necessary in HubSpot.
Yeah.
Yeah.
It'll even understand you at one and a half X on the speech though.
Exactly, yes.
Now the demo showed, it showed 20, what was it? 20 clicks and a prompt.
Yeah.
Right? Again, you're typing in a prompt, and it's saving the user 20 clicks on the screen. Are companies gonna still need an admin? Are they still gonna need as many admins just to be running HubSpot?
I think so. Let's take a step back. The reduction in the number of clicks, which I love, I'm all for efficiency, is really not where the value is. It's not in the reduction of clicks, it's in the reduction of complexity, right? It's in the reduction of the cognitive load that a user has to have when they're starting with HubSpot, right? Remember, once again, we're in the SMB space. If we're selling to a 20-person law firm, let's say, which could be a typical customer, they don't really have like an expert inside of HubSpot. That's one of the reasons we've grown as well as we have, is that it doesn't require the overhead and the cognitive load. This takes it even further, making HubSpot more approachable.
It's not about the 20 clicks that you save, in when you're trying to build a report or do whatever you're trying to do, it's the 20 hours you save having to learn how the HubSpot software works to kinda do the thing in the first place, right? It kinda makes the software much more approachable to a wider pool of people, and hopefully helps us get from the 180,000 customers to in the millions someday, 'cause those companies are out there.
I believe, if we take a look at the substrate layer for ChatSpot, I think what we saw in the demo is using OpenAI's ChatGPT. Makes sense, you were meeting with Sam Altman. It was using DALL-E 2, I believe, which was their image generator. Will it use anything else? For instance, if we look around, like, do you think it could use Stability? Do you think it could use Midjourney? Do you think it could even use other large language models, or are you better off kinda picking a substrate layer and sticking to it?
We love OpenAI and have a deep partnership with them. Having said that, ChatSpot already uses a bunch of large language models for different use cases. One of the things I think that's important to recognize, and this is gonna be, I'll try to keep this short because it's super geeky, is that right now, 95% of the excitement that happens in generative AI is when these large language models produce something for human consumption. It's like, "Oh, write me a blog post about X, Y, Z industry, you know, and answer these questions," whatever it is. Or, "Generate an image for me, generate a video for me." In all of those examples, we're generating something a human is going to consume.
Where it's actually really super interesting, and I think where the 95% of the future value is generating things a computer will end up consuming. Generating code to say, "Oh, I'm trying to solve this problem." It's like having a super fast, super efficient coding intern, right? They're sitting there 24/7 just to write little applications for you to do whatever you wanna do. I think that's where a lot of the, kinda value, future value is going to come from.
If we recognize that, one of the nice things about, you know, generating code or generating, structured data out of these large language models is that they're measurable. They're what I would call deterministic. For instance, if I generate a blog post through OpenAI, 3.5, 4.0, it's a qualitative judgment as to whether that blog post with 3.5 is better or worse than 4.0, is better than worse than Anthropic something else.
If I generate a body of code or a report or something that's like that, it's like I can tell you whether the report is accurate or not. I can tell you whether the code runs or not and does the thing it's supposed to do. The byproduct of that is we can and do swap between large language models. Every week, a new model will come out, and every week, I will say, "Oh...
Which one's better? Yeah.
I'm gonna run our benchmarks against this new model to see, performance-wise, is it better? Is it better at certain kinds of things? Is it better at this? We have no qualms about using multiple models to get the job done, to be able to kind of make the most of where the industry is headed.
Okay.
Yeah.
Okay. You have optionality.
Yeah.
What would you say to an investor that who may be currently concerned, you know, kinda website builder capabilities, website hosting, blogging infrastructure? That, you know, this has been a discussion that has impacted companies like Wix and Squarespace and GoDaddy. HubSpot has a small slice of its of its business there. A lot of that is kind of attached on to the other hubs that you've got. What would you say to an investor who has a little bit of a concern?
Yeah, I mean, it's. To me, it's not just about the website building thing. Okay, I'll say a couple things. One is that the website building space, and I've said this, I think for years now, even with relation to our own product, is that website building over the mid to long term or the fullness of time doesn't solve the complete problem for a customer. What a company wants is to create a digital experience for their end customers. Now, website is a important piece of that. It's like, "Oh, I wanna put these 15 web pages up, and I'm gonna do a drag and drop interface." To me, that's just table stakes, right?
What we're gonna see now, the disruption that's going to happen, is that even a year or 2 years ago, just doing websites was probably, like, not enough for some.
Yeah
...of customers. Now it's gonna be like, "What do you mean your digital experience thing that's..." which is what it's supposed to be, "is not connected to your CRM or is not helping me use generative AI to generate a webpage on the fly based on what the customer's looking for based on the context?" That's just going to be table stakes. I think based on whether you're an optimist, like me or not, I think it creates an opportunity for all of those companies, including HubSpot, to say, "Hey, here's a way for us to deliver even more value." I don't think you can sit on your laurels as a website-building, website-hosting company, just generally speaking, not about any specific company. I think it's gonna be an interesting time.
Yeah, for sure. fascinating. Kate, last one on this topic of generative AI. Do you have a framework for envisioning how you would try to monetize AI capabilities? I mean, so if, let's say one end of the spectrum could be, we're basically gonna roll it out there for free. It'll enhance the experience, and we will benefit in terms of loyalty and stickiness. There could be something, you know, in the middle where there's a, there's like standalone pricing. You know, it's $5 or $10 a month if you, if you wanna have at it. There could be a new SKU. You know, maybe there'd be something, a new SKU that would be even like above the enterprise tier.
Have you had any time to think through what is optimal for HubSpot?
I mean, we certainly are thinking and talking about that. This is obviously a very early and highly evolving topic. If you look at HubSpot, and you look at our history, I think one thing should be front and center for you, which is we have always made choices that put the customer at the center and sort of prioritize delivering customer value. Our foundational pricing strategy is first deliver value to the customer, then extract value from the customer, and this will be another example where we take that same approach. You know, you heard Dharmesh use the words table stakes.
There is likely to be some portion of generative AI features and functionality that will exist in HubSpot as table stakes and will exist in our product, you know, as it, you know, is currently structured. There's likely other parts that are either more sophisticated or specialized that, you know, maybe they'll be in the higher-level editions of Enterprise and Professional, and maybe they drive value there that we would, you know, realize over time. There might be a separate SKU for specific things or volume-based pricing. I think the message here is like early days, we're gonna wait to make sure we actually understand where the customer value lies, and then we'll make the appropriate pricing and packaging choices.
Just to clarify one word, did you say could be volume-based, or did you say it could be value-based?
It could, I mean, it could be volume-based.
It could be.
I mean, the foundational decision will be: Where do our customers find the most value from this product? And how do we price and package it in a way that, you know, allows, you know, it is aligned with the value we're delivering?
Okay. Let me ask you one more. I'm actually getting questions. on the iPad, which is usually does not happen. I'll check from the audience as well. I do wanna ask you about payments. I try as an analyst not to get too overexcited about super nascent products along those lines. You know, we didn't do that on Coupa Pay. We didn't do that on Twilio Flex. In this area, and I understand it's very early, we get that, but our partner feedback is actually very good on the topic of HubSpot payments. Is there a way you can just kinda tell us what is happening there? Are there increasing capture rates of GMV? Are customers booking deals, you know, inside the CRM using payments?
Any kind of advice on how to think about that trending out the next 3 to 5 years?
Yes. Payments is, I think as we have said, tried to say consistently and often, a long-term bet for the company. We are purposely trying to be strategically patient in how we build and scale that business. That said, we do see some nice consistent growth in active merchants, which is the North Star that we are currently following as we continue also to see GMV grow over time. That said, last year we made a choice to make payments a, what Dharmesh refers to as Primary Color, a foundational element that lives in the platform layer that then can be deeply embedded in all of the hubs.
That was a strategic choice that certainly slowed us down in the near term, but we felt was the right choice to allow us to go further with payments over the longer term. This year we're focused on continuing to deliver the features and functionality that our customers have been consistently asking for in that native payments, application to invest in some flexibility around back-end payment processors and to make sure that we are thinking about that like end-to-end customer experience as they discover and enroll in payments to eliminate as much of the friction as possible.
That's a good nod to the future, I think just having elevated into one of the Primary Colors. We're down to about 3.5 minutes. Why don't we see if anyone wants to raise their hand in the audience, and we can get you a microphone. Right here.
Hiya. Thanks a lot. First of all, can you just talk a little bit about how you think generative AI might impact the cost structure of the business, and then more broadly, just through that value chain, where the value ultimately accrues? Thanks.
Yeah. I'll start since I've been watching the numbers and as the usage grows. As it turns out, consumption of these large language models, you know, from the foundational model providers is relatively low as to almost be kinda inconsequential. And the other thing we're seeing, and I think, this is not counterintuitive at all, is we're gonna see dramatic reductions as... Like every week there's a new open source large language model comes out. The bigger players are all getting in, all kind of playing catch up, which I think good for customers like us. Our assumption will be this will play out a little bit like AWS did. That there... We'll see infrastructure providers. Yes, they will capture some value, but there will be enough price competition, enough.
We can maintain optionality enough to the degree that we're not dependent on any one vendor. We don't think the COGS will actually play that much into the overall strategy, especially if you compare the COGS to the actual value we're creating for our customers. We think that's gonna be a non-issue over at least mid to long term.
Yeah. I mean, the only thing. That is certainly our baseline assumption. We are paying a lot of attention to it. As we think about the pricing and packaging choices over time, that will be one of the elements we consider.
We have one, kinda right back in the back row by the column there.
Morning, thank you very much for the time. I'm interested in your view on the opportunity in the commerce space. If we compare the B2C world with what a Shopify and others can do with the B2B world, it feels like there is a pretty yawning gap between the two, and both what you can do there and whether generative AI plays any role there. Thank you.
Yeah. One of the reasons we're actually in payments is that we think it's a critical step towards this larger vision of commerce for B2B. If you look at the evolution of e-commerce, which, you know Shopify exemplifies, they have done a fantastic job improving the end user experience for buying things. If you're looking to buy a widget, and you come there, you put it on your shopping cart, and you check out, that needs to exist in our minds for B2B companies. That expectation is gonna continue to rise for the end customer. We wanna do is to take a considered purchase, a B2B purchase that says, "Oh, you're not gonna sell something right online for $19.99.
It's gonna be a multi-month, multiple people involved, multiple step process, which HubSpot has been kinda working in for our entire history. Can we make that more digital? Can we make that more automated? I think, yes, generative AI will play a role in smoothing out that experience, but there's a fair amount of foundational work to do. That's what has us personally excited. It's not payments. Not to say payments is not a great vector, but it's the digital commerce experience improvement that I think where the big opportunity is for B2B.
Okay. With that, we are right on time. Kate, Dharmesh, I cannot thank you enough for taking time out of your busy schedules to be here. Thank you.
Thank you for having us.
Thank you.