Twilio Inc. (TWLO) Earnings Call Transcript & Summary

All right. Why don't we go ahead and kick things off. Thanks, everyone, for joining. For those of you that don't know me, my name is Arjun Bhatia, and I am the research analyst here at William Blair who covers Twilio. For a full list of disclosures you can go to williamblair.com. And with that, it's a pleasure to introduce Khozema Shipchandler, who is the CEO of Twilio. Khozema thanks for being here.

Yes, thanks for having me.

Maybe a good place to start. I think everyone is generally familiar with Twilio. But I think the business has changed quite a bit over the last several years and especially over the last, I'll say, 2 years. When you took over as CEO about 18 months ago, so you've kind of brought your own changes and initiatives to the company. So I would love to hear your perspective. Obviously, you've been at Twilio for longer, but I would love to hear your perspective on how the company has changed over the last 2 years and what you're looking for going forward now that some of those changes have been implemented and you're starting to see some of these early results come through?

Yes. I mean I think in many respects, other than the core products, like fundamentally like everything has changed in many respects. So if you just kind of go back maybe at a high level, I think the way in which we're running the company is like quite different. Like we're running it with a lot more financial rigor, a lot more operating discipline. I think in terms of the growth bets that we're taking, we're being a lot more focused in terms of, rather than focus at 10 things, perhaps all at the same time, with equal intensity, focus on 3 to 5 that seem to have higher efficacy. Once we've got product market fit, continue to invest more dollars behind those, but invest small, invest early, perhaps, maybe experiment a little bit. And then put more in as we find traction. I think you fundamentally see that reflected in the financial trajectory of the company. So we've taken a business that just a couple of years ago was burning through cash, was losing money. And sitting here today, we're high-teens margins. We're generating a significant amount of cash flow. We're GAAP profitable. Over the last 12 months, we've been able to kind of reaccelerate the growth trajectory of the company. So kind of post COVID, like a lot of that had come down like a lot of other technology companies. And I think for us, it was going to be a self-help story. I think we've done a lot inside the company. I'm sure we'll talk about that later, in terms of reinvigorating growth organically. We did a few acquisitions over the last couple of years. I think we rationalized those. We parted with some smaller assets that were just not a great fit. And so -- and on top of it all, we changed out about half the management team. So yes, a lot's changed, right, in a period of a couple of years. And I'd say finally, maybe the last thing I'd kind of point to is that in achieving many of those things, we also shrunk the workforce size. And that was done kind of painfully in the sense that there were layoffs associated with that. But I do think it was the right call for us, even though I obviously feel bad about what happened to those people. I think that the size and the shape of the firm today is perfectly suited for where we're going. And I've never been more excited about the prospects of the company. And I think today, with some of our newer products, and now the addition of AI to be able to supercharge that, the growth prospects look great.

Yes. It almost -- you talked a little bit about the retrenchment, right, like focusing on a few key bets and one of the -- I don't know if it's a bet or I would call it maybe like a framework or philosophy that you've focused the company on is this aspect of communications, plus data, plus AI. Like those are your three core pillars, from a product perspective, from a platform perspective that you've kind of laid out as your value prop. So maybe practically, I would love to hear what that means for Twilio? What it means for customers? And what it means for like how it enhances the value that they're getting out of Twilio from those three buckets?

Yes. I mean the value proposition is relatively straightforward through the lens of each of our own consumer experiences. So today, each of us receives text messages, e-mails, sometimes voice calls, from different brands that we care about. It could be retail, it could be financial services, it could be a health care transaction. And I would say, certainly, in my experience, the vast majority of those interactions are incredibly flat. They totally lack context. Even though it could be the 100th time that you're being notified, or engaged with in some way, it feels like the first time because they are not using any of the data that they definitely do have about you. And so enter Twilio, right? Like the idea is that how do you create much more enriching experiences when you end up communicating with one of your consumers. If you can do that, and we have a lot of data that kind of demonstrates this, we just put out our state of customer engagement report yesterday actually on this. The data demonstrates that if you can create more intimate experiences with consumers, vis-a-vis the different communication tools that you're using, you have a better relationship with them, you're able to drive more revenue. And actually, the consumer doesn't mind being communicated with more because at least it's engaging, and it's enriching and, it's done in a way in which it's value adding, right, versus just transactional. So that's kind of the opportunity that we see. And every time that we talk to a customer about this, like their eyes really light up about it. And it has an opportunity, especially now with AI, to be able to save the money on the one hand, but also generate much more revenue on the other side of it. And in many cases, and increasingly, this is the case, for the consumer on the other side of the transaction, like their problem is actually getting solved much better, right? So it's sort of like a 3-way win. Like the business benefits because they're able to save money and drive more revenue. The consumer benefits because their problem is getting solved and they're getting a much more enriching relationship. And then obviously, we make a buck off of that. So I think the thing about some of the things that have also transpired over the last 12 months is like we never really put those products together, right? So we bought segment, the business. But if we really ran it as more of a stand-alone thing, and I think now the way in which we're positioning the company going forward is one complete platform off of which through just simple APIs, you can access the totality of all of it. And if you want to ingest data through the communications that you're using, it's super simple to do, and then you can plug in whoever you use from an AI perspective.

You had some very good, I think, and very cool product demos at Signal a couple of weeks ago, which showed this kind of element of the CDP from segment, getting integrated with the communications. From a functionality perspective, I know you've been working on this integration for a few years. But is that live? Is that -- is that something [indiscernible] today?

It's already today. So the example -- so we showed an example on our technology conference. And we actually -- it was inspired by a customer. You can probably guess which one. But it was inspired by a customer. And we said, let's pick something that's really hard so that we can prove to everybody else that the relatively straightforward ones like, I'd say retail is more straightforward than a mortgage application, or a health care interaction, or a financial services transaction, like prove that it can be done when it's really hard. Prove that it can be done across multiple channels. Prove that context can be used to enrich the interaction. Prove that all of that data is stored and can be used later for the lifetime value that the brand wants to be able to have with that customer that all can be done. And so the product is live today. Customers are using those products today, the product that you're referring to is called ConversationRelay. And it's -- I think the take-off philosophy has been fantastic so far. So I've been on the road nonstop since that conference. Customers are super excited about it. I'd say the uptake rate on AI is cautious depending on the industry that you're in. I think it's really, really fast for like e-commerce and retail, where maybe the stakes are a little bit lower. But this is live product. And I think every one of us will end up using it, whether we realize it or not.

And competitively, right, there are other competitors in the space. But I don't know how many -- what do you see out there in the market in terms of how many other players have been able to stitch together these kind of three core buckets? Like who else has...

I don't think anyone can really stitch together the data capabilities, [indiscernible] communications capabilities with AI in the way that we do. So I think in our space in that way, we are truly alone. I mean, I think the consumer data that we access -- so there are other companies that obviously are in CRM, right, some very large ones, clearly. But to be able to reference the data at the consumer level, combine that with the communications channel, deliver it across a multichannel experience, be completely agnostic about the AI that gets plugged into it, I think Twilio really stands alone in that respect, and that's why we're so excited about it.

Yes. Maybe let's talk about AI for a second because it's obviously an important part of your story. But when I think about the communications stack, and I think about all the offerings that are out there, you at the CPaaS layer have your own AI capabilities. The application vendors that are building on top of you have their own AI capabilities. So the middle layer with UCaaS has their own capability. So if you're a customer, it's sort of -- it can kind of get messy real quick. But what is your perspective on why Twilio is well positioned to deliver your AI capabilities to the customer and how you...

Yes. I think it depends on like what the AI is that you're talking about. If it's an LLM then we're not going to develop our own LLM technology, right? Like we don't have the budget for that. I think there are a number of other companies that do it extraordinarily well. In all of those cases where an LLM has to be used we're going to integrate with somebody else. Now there are many instances in which our customers are already using one of those vendors. And so in those cases, we'll just plug into whatever they're using. There are other cases in which we'll recommend based on different technologies that we're using. So LLM, let's kind of leave that to the side because I think that's a different category. I think with respect to like something like our Verify product, or Fraud Guard, in terms of like protecting the customer from like fraud really escalating on the platform. That's one in which they're going to use ours because it's like natively built into the platform, and it's right there. I mean there's like no -- it's easy button simple in terms of your ability to avail yourself of it. It's just like literally the click of a button on the console. And so you're going to access ours in that way all day long. If you're accessing an application that's ultimately running on our rails, I mean, the reality is there will be many instances in which we don't even realize that because that's actually transacting through an ISV. So I would say in those cases, if a customer -- if an end consumer ends up using some sort of AI technology through an application that happens to be on our infrastructure, that's fine. I'm like I'm not hung up on that at all. I think the ability to connect data plus our communication stack, like that's trickier, right? Like that only can be integrated by us. And again, there'll be applications that run on top of that as well. So I actually don't think it's that confusing ultimately for customers. I think that typically, these customers are very, very sophisticated. If you're a developer, you know what you're coming to Twilio for. If you're an enterprise customer, you know at what level you're wanting to access our technology. And so you know what different integration points are. And so I actually think it's not that confusing.

And what is your perspective just in terms of sort of AI monetization because it's a big question. I think everyone is still trying to figure out what the right model is? Do we monetize? Do we not monetize? So what at Twilio -- what is your sense of...

I think it was extremely, extremely early. I think it's definitely going to happen. I mean, I think that -- let's use an example, right? So that ConversationRelay product. So what -- just to maybe peel it back with what actually happens behind the scenes. So what happens at sort of the bare metal version of it is, is that you've got a voice API, right, that's being activated during the course of an interaction between a consumer and the company, right? You've got a data API. I mean we've put it all into one singular API, but just bear with me. So you've got a data API, and that allows you to be able to access the historical data that you have about that consumer and, for that data to get referenced during the course of that interaction. And then in that particular example, what we did was we used a partner of ours to be able to do the AI recognition of voice to be able to process the voice interaction, stuff like that, obviously, through our voice API. So let's call that a separate bit of technology. And that is a live interaction that is happening with the likes of Rocket Mortgage, with the likes of Cedar Health, with the likes of Domino's. So well-known brands, I would say, that are using those capabilities today, there's a discrete AI portion of it. Now our job is to make that super simple, package it all up in one API so that they don't -- they're not exposed to all of that different technology. But with respect to monetizing AI, I would say it's not monetizing AI discretely, but monetizing that full package of things which happens to capture a significant piece of AI to be able to activate that experience. Does that makes sense?

Yes. So you get a core benefit out of it and you'll see volumes and competitively, you'll be better positioned. Okay. Maybe switching gears a little bit. So Twilio started, your roots are in voice. Pretty quickly, messaging took over, I think, as the main use case for Twilio's products. But more recently, we've kind of seen this growth pick up in voice again, and there's what you call this renaissance and invoice happening. So I'm curious what you're seeing with the voice channel because AI has a pretty big role to play there and what you're doing from a product perspective to maybe facilitate that further?

Yes. So we put out some disclosures during our Investor Day, one of which was -- and it's obviously grown since then. But at the time, 9,000 different AI companies working on Twilio, right? And many, many of those being start-ups. And I would say of the start-ups, the vast majority are voice related. So why is that? I think the reason is, is because today, at least, voice, it's just so natural for us to interact in that way, right? So the AI technology on the other side, if you're calling into a customer service transaction, or you have a question, or you're resolving a health care issue or whatever, the AI is able to pick up language. It's able to distinguish accents. It's able to -- more or less, I would say latency has been figured out depending on the cadence of how you talk. It can handle interrupting. And I think most importantly, it picks up a motive technologies very -- a motive nature of a conversation very well, right? The technology is very good at that. That is really, really hard over text today, okay? So how do you convey a motion? How do you convey that someone is actually super angry? How do you convey that they're incredibly delighted, right? Like all these different things, that's important in the context of a service transaction today. And I think it's also natural for us to engage with the technology in that way. I'll give you an interesting example. So we're doing work with Cedar Health. I referenced them a moment ago. So in health care, the vast majority of customer service calls in that industry, believe it or not, 97% of calls are I don't understand my bill okay? So -- and it's actually -- it's 80% for most industries, but it's phenomenally high for health care. Now in the case of health care, what's further interesting is that consumers are typically quite embarrassed about calling in because they don't understand like medical jargon, and they often feel like there's this information asymmetry between themselves and the other person. And so enter AI now. The data shows, and Cedar kind of prove this out to us, that consumers would rather interact with an AI agent in those cases because there's no notion of information asymmetry. I mean it may exist, but there's not a person anymore, right? It's disclosed upfront. And the beauty of it is, is that the cost goes way down, right, for the health care provider. The consumer prefers it because it's a voice interaction and it handles all those things, those attributes that I talked about a moment ago. Their problem gets resolved in the context of it, and they can talk endlessly right, which an agent -- a live agent just doesn't have time for it, or it's too expensive, right? So in terms of like that 3-way relationship of the customer benefiting because they get lower cost, right? And they can solve the consumer's problem. The consumer benefits because of the embarrassment factor and their problem gets solved. And then again, Twilio makes a buck off of that, right? So in that case, kind of combining your two questions, it's voice driven, but at the same time, we're monetizing AI through the voice API that's driving all of that, picking up all of the different data elements of that interaction, storing it for later interactions, training up that discrete model so that they can be used and can be learned from going forward.

I almost think you might have more consumers starting to use voice from a -- if there's a better experience, you're not waiting on hold, you're not waiting for an agent, you're not getting transferred 5 times, if there's an AI agent at the end of it.

Well, I mean food delivery is actually a great example of that too where you call and like the first thing that happens is you hit a bunch of IVR menus, which is frustrating in and of itself, but then you're on hold, as you just said, right? So all of this frustration can kind of be taken out of it. And so it's starting in voice. I do think it will move over to the other channels. I think e-mail -- in my opinion, e-mail will be next because it's kind of more long form. So it's a little bit easier in terms of like the emotive aspects of e-mail for that to get picked up. It is harder, but it's going to happen for sure. And then inevitably, it will move to SMS.

Yes. Okay. Let's turn to the business a little bit because you talked a little bit about the turnaround, right? You have -- and you have ambitions. You laid out as your kind of goal that you're running the business for double-digit growth. We've already seen the re-acceleration play out a little bit. I think your trough was 4% growth. About a year ago, you did 12% last quarter. So when you're thinking of this re-acceleration, maybe help lay out for us what the biggest factors are in driving this? And over the next year, over the next 2 years, what you think those -- if those factors will change at all? Or what do you think are going to be the biggest drivers?

Yes. I mean I think it's going to be a combination of go-to-market and product. I think on the go-to-market side, it's predominantly self-help. So I think the self-serve part of the business, which is really the origin of the business, it's the bread and butter of the business. We've made a lot of investments to make the console experience simpler for customers, to make it much more intuitive to get through it as quickly as possible, to reduce the compliance burden. So self-serve has performed extremely well, I would say, over the last several quarters, and I think that's been a growth driver of ours and a very intentional investment on our part. All the way maybe sort of on the opposite side in a way, I mean, it doesn't have to be, but ISVs, like that's been sort of the other, if you want to take kind of two ends of the spectrum. We've put a lot of investments into making the experience much simpler for ISVs. Now ISVs kind of run the gamut, right? They can be very, very small companies. They can be very, very large companies. But the ISV experience, I think we ran a summit, the night prior -- or the day prior to our Signal conference. Got all of our ISVs together, obviously, got their feedback, are starting to integrate with many of them. So in many cases, the ISVs deliver the application layer, which we do not. And so they want to be able to go to market together. I think that has obviously attractive aspects. I think as I kind of look forward -- so those two have been really, really strong recently. My expectation would be they remain strong for a period of time because of the investments that we're making. I think going forward, partnerships, I'm particularly excited about. I think that's an area in which we've under invested, and certainly under delivered. And I think, obviously, there's leverage there, right, if you can do it well. We've started to work deliberately with a handful of our partners, especially on joint go-to-market capabilities. And so I think that's been attractive. On the product side, we're building the right products for this time, right? So all of these different AI interactions, what's definitely true is that -- they definitely require some sort of communications channel to allow for them to take place. They require rich data to be able to inform the AI. And I think what's actually particularly interesting is the reason that it will not be delivered by an LLM is that beyond expense and stuff like that, which is real. But you're not going to give up -- if you're a company, you're not going to give up that data capability over to an LLM, help them train up their model, disintermediate your customer relationship. Instead, you want to keep that all together. So I think the technology that we offer, it's like really compelling and at the right time. And some of these recent innovations -- like the pitch to customers is like, look, if you don't see the ROI, then don't buy it right? Like we want you to buy stuff that's very, very high ROI for you guys, right? So Verify, in many cases, that's like less than 30 days time to value in terms of the ROI that they see from it. ConversationRelay, I mean as soon as you're activated, you're immediately getting ROI because you're immediately saving money, your consumers are immediately benefiting from it. So I think that's going to be really attractive. And then the last thing I would say on the technology side is, we signed up this relationship with Microsoft recently. They're obviously a very compelling partner from a technology perspective. We're using their AI capabilities to help inform our technology stack, and it's very early days, but we're super excited about it.

I want to come back to the Microsoft partnership. Maybe before we go there. Can you -- so you talked about a few products there, ConversationRelay, Verify. You have Conversation Intelligence, also that's coming to market. When you're thinking now of growth, it seems like cross-sell is going to be a bigger part of the equation, right? You want customers to out multiple products whereas in the past, I think Twilio has been more about core communications. So when you're thinking about the growth algorithm, which products are you most excited about among some of the newer launches from a cross-sell perspective?

I think -- I mean we talked a lot about voice, right? So voice is kind of a no-brainer in some respects. So let me just back up one step. So the reason we're excited about cross-sell is we've never really tapped that as an opportunity to grow the business, okay? So sitting here today, like still, I think, about 63% of our customers still only use one of our products. Okay. And yet, the multiproduct customers that we have drive the vast majority of the revenue of the business. Okay? So if you're using SMS, there is like no reason you should not be an email customer, right? I mean the reality is that many of these companies, if you're a startup, or if you're an SMB mom-and-pop style business, like you're not using anything, right? So this is all like a greenfield opportunity. So it's not like we have to go and displace someone, right? So I think that's very attractive. I think the combination of voice with SMS is actually very powerful. So I give you an example. Voice, obviously, putting aside some of the AI interactions for a second. The channel is less utilized, or has been up until this AI moment, because of the fact of robo calling, right? Like a lot of us get phone calls for numbers that we don't recognize, like it's frustrating. So what a lot of customers are doing is they'll -- it's a very like expert interaction. They have to be able to talk to someone before they provide the service delivery. So basically, what they'll do is they'll send a text message to a customer saying, hey, effectively, I'm going to call you the next 5 minutes. Okay? So I need to pick the phone. We're providing the service this weekend. Like look for solar installation, okay? You have to be able to talk to someone before climbing under their house, obviously, right? So that's pretty effective, extremely effective, right? So customers definitely do pick up the phone after an SMS gets send to telling them that they need to. And then you get to kind of these AI interactions, right? So if you're a service delivery provider of any kind -- let's just take retail as a simple example, right? If you a retailer and your primary mode of communications is through e-mail, which it often is today, and it's typically in the form of like order confirmation, receipts, coupons, stuff like that, like there's a real opportunity to enhance the experience with your consumers by pushing promotions, for example, to your customers through SMS. Or in many cases, and we're seeing this like with our partnership with Chelsea. If you're geo-tagged, you also have the opportunity to identify like when somebody walks into a stadium, and like deliver the fan experience in that way, right? So there's a lot of interesting things that you could do across these multiple channels, and I think we're starting to see that take off.

And from a go-to-market perspective, how are you kind of incentivizing the sales team to drive this cross-sell?

Yes. I mean, basically, we're biasing the comp plan in such a way that the best sellers, the ones that are going to end up sitting at some beach at Sales Club next year, or the ones who are doing cross-sell the best. And it's got materially higher incentives to push it. And I think even more interesting is going to be, as we push further towards solutions that combine these various products together and make it even simpler to avail yourself of multiple products at the same time, that's going to make it even faster.

Yes. Okay. Can we come back to Microsoft now because I'm curious what that partnership entails exactly? It sounds like there's certainly a product in AI collaboration. How about on the go-to-market side because you had you had Satya and yourself make an announcement virtually at Signal about what this partnership means. It seems like a pretty big evolution, or a big partnership for the business, at least. I'm curious what the details of the partnership will entail?

Yes. I mean it is big from our perspective. I think the beauty of the partnership is, from a technology perspective, there is extremely little overlap in terms of the different products that are offered. And so it's highly complementary, which I think has benefits down the road depending on how far and wide we take this relationship, number one. Number two, their technology is very good, right? I mean they have obviously tremendous enterprise. We did kind of a bake-off in terms of like different AI capabilities that we wanted to have included as part of our kind of native tech stack. And we found the ones offered through Azure to be extremely compelling. We've used Azure in part as part of our technology stack for a while anyway. So it's not like it was like brand new and the integrations were going to be relatively straightforward. I think because they're so used to working with enterprise companies, they've made it remarkably easy for us to be able to integrate into their workloads. So I think that's been great. And then where things really take off is that when you actually put -- when all the kind of people that are signing things like walk away and the product teams are actually sitting together and working on stuff. Like our ability to like launch a handful of products within a few weeks, both at our conference, as well as at their conference, like that kind of proves that these things are going to work from a technology perspective. And we're starting there. I think it certainly opens the door to do go-to-market things down the road, but we're starting on the technology side.

Okay. Very interesting. All right. In the last few minutes that we have remaining, we've touched a lot in this conversation on the product innovation the drivers of growth reacceleration. I think one of the other interesting parts of the Twilio story for investors, in particular, is the margin expansion. And you've made a lot of progress over the last 2 years. You've kind of talked a little bit about the operational changes that you've made. You're at 16% operating margin, and I think your target is get to 21%, 22% in the next few years here. What are the biggest drivers of margin expansion from here? And kind of what are you changing in the business from an operational perspective to be able to drive that?

Yes. There's sort of a short, medium and kind of longer-term dynamic to it. So look, in the short term, the reality is, is that I think we're still like largely untapped when it comes to using AI inside the business, using automation capabilities inside the business. And quite frankly, there's more that we can do and are working on with respect to the right geographic mix of talent in terms of the business. I think the beauty of where we are today, I kind of briefly touched on this in the opening remarks is that we like the size of our workforce and don't feel like we have to add head count in spite of some of the growth characteristics that we're seeing. In spite of some of the product investments that we'd like to make, we can largely self-fund all that, right? So think about the business being kind of relatively flattish in terms of the head count. I think over time, using automation and AI, we should be able to drive some OpEx leverage just off of that before I even get to the revenue line. And I think that's really attractive. And I think in -- it's going to take a little bit of time, but I think the things that we're doing right now are using AI in terms of customer support, using AI with respect to our digital sales reps, diversifying the geographic footprint of the business, especially through backfills as attrition happens. And so all of that's underway. And I think that kind of drives short-term margin expansion, shall we say. I think in the kind of short to medium term, if you will, I think we're obviously targeting like revenue growth that's at higher rates than as part of our actual financial framework. And so you've seen our ability to do better over the last couple of quarters, and I think that's obviously driven some upside, right, to guidance and stuff like that. And so I think if we can deliver at sort of our aspirations, that also drives some additional margin expansion capability. Certainly drives additional free cash flow, which I think investors are particularly interested in. And so that's attractive for us, right? Like we're not intending to go through like some massive investment ramp. Instead we think a lot of that ends up dropping to the bottom line. And frankly, that allows us to kind of explore some things from a capital allocation perspective. We've been obviously very focused on buyback recently. And then I would say in kind of the medium to long term, a lot of these newer products will drive upside to your gross margins, right? So Verify is a really good and simple example where it's effectively an add-on to -- it immediately adds a software-like margin on top of SMS, which is kind of in a middle margin product, at least for Twilio, and that's attractive, right? So the more that we can -- before you even get to like the wholesale cross-sell, like if you can just do these add-ons, right, like that's really attractive. I think voice, given that we're seeing what we are right now in the market, I think it's more attractive in terms of its margin characteristics. So I think that helps over the medium to long term. And then I think ultimately, like this one unified API of ConversationRelay being able to use voice, data, AI, all in one, that has attractive margin characteristics. And so I think those dynamics tend to boost gross margins over time. It's harder to do just because the weight of the business, as you pointed out in the very beginning, is so messaging focused. So it will take time, but we're very excited about it.

All right. Very interesting time in the story. Khozema, thank you so much. I appreciate the time. Thanks, everyone, for joining.

Great.