Not often we have a founder with us, so it's actually really good to kind of start, like, slightly differently because from a lot of investors, they're still, you know, I'm still getting, like, kind of more questions where they're trying to understand the company better.
Mm-hmm.
Like, when you guys kind of came together and started Klaviyo, like, what was the initial thing where you looked at the market and thought, like, "Oh, I can solve this. I can do better here"?
Yeah. So I think there's kind of two things came together. So the two of us had previously worked at a startup that was building large-scale business intelligence and testing software for really the Fortune 500. So, like, the Starbucks, the Walmarts, the folks like that of the world. And the specific problem that we ended up working on, the technical problem, was unifying a lot of their data into this intelligence platform. So and then we'd use that to help them make decisions.
And so, as an example, we'd, you know, somebody like Starbucks, if they were trying to figure out how much they could raise the price of a cup of coffee, you know, in the Starbucks data system, the sales system that knew what each individual consumer had bought was totally siloed from the inventory system that knew if that store had enough beans of all the different types, was totally siloed from the time clock system so they could figure out if it was understaffed in a given day, and was definitely totally siloed from the real estate system, which we could then use to figure out, was it raining outside and when was it raining? And so we unified all that data, made it available at very low latency so we could run these huge-scale calculations.
So that was kind of the background that we came to thinking about a startup with. When we went to found Klaviyo, you know, our approach was we had a belief that if we could find a problem that existed in the world that we could use software to solve and that people would hand us a credit card, to solve their problem, that we could build a business and that if there were enough people who had that problem, we could build a very large business. And so, you know, as we started exploring what those different problem spaces were, we knew we wanted to bootstrap. We wanted to, you know, wait till we'd really found a working business before we went out and thought about how to capitalize it.
Mm-hmm.
So, early on, we took what we'd learned in the original company and said, "Okay, let's build this true database customer data source of record, that understands, you know, as more and more businesses are digital, everything a customer's ever done, when they've been on the website, how long, which pages they looked at, what they bought, what they haven't bought, what made up the basket, you know, any support interactions. Let's just build a pure data source." And at the time, we thought that would apply to, you know, we mostly knew software. So the first market we thought about were software companies. But we quickly, you know, as we built this out and started to find customers, we said, "Okay, this database is most powerful if we have some way to drive action on top of it.
Yeah.
And so we stuck, we thought, you know, email was just kind of getting commoditized, and we said, "Okay, take this, what's really valuable, this data platform, stick the messaging component atop it, and then we can use the data platform to both drive intelligence and value and also measure the impact of what we were doing." And so in that process, we got our first 10, 20 customers who are really all sorts of different businesses.
Mm-hmm.
So, apps, software companies too, we ended up with a decent chunk of e-commerce businesses. And that's when we really realized, "Okay, hey, software in the marketing category has really always been about the app, about the messaging." That was, we thought, the easy part. This database part of truly understanding your customers in a personal way and then being able to trigger messaging to them in, like, a very targeted way that drove results was something we could uniquely do. And that's really, that was kind of the beginning.
Mm-hmm.
And so we had the first 20 customers, half ended up in e-commerce. We leaned into e-commerce really as a way to grow efficiently, without needing much capital, and we were, you know, effectively off to the races.
I mean, congratulations, by the way. It's like, you know, it's an amazing story. If you think about it in that space, there were kind of quite a few guys. There still are quite a few guys, like, you know, doing great in a lot of them. But like, how do you get that differentiation through? Because, you know, like, it's relatively easy to say, like, "Yeah, I can send emails," and it's like, "Yeah, but it's not.
Yeah.
You know, like, you know, it feels like there's, like, a lot of, like, education, "You can do this better," or was it just the demos spoke for themselves and then you're off you go?
Yeah. I mean, so, a couple things made that a lot easier, so first, you know, we very much focused on this as a, you know, see if the customer has a problem. And so rather than going and say, "Hey, are you sending emails?" it was very much, "Look, are you sending this type of email to somebody who's been on your site and then they don't complete a purchase? Like, are you triggering that email?" And if they said no, we could say, "You are throwing away, you know, $100 for every person who does that." That was one, so the fact that we could quantify exactly what the impact was.
Mm-hmm.
But two is we had a philosophy that even if we weren't gonna ultimately have all our growth be product-led, if we built the system such that we could get someone up and running in a couple minutes, we could live on a sales call, set up this email, and they could start to see results for a couple days. And if we made that a free trial, they could immediately start to see value. And so the more we could make that sales conversation about revenue and not about, "Hey, replacing this commoditized system," it was a much easier sale. And we still had the benefit of, "Look, you've got existing budgets set aside for this. If you take that budget and give it to us instead," it made it easy.
So it was a combination of promise of, like, "Look, you can really see this, plus we could measure the impact," and then it was replacing existing sites.
Yeah. Is that where the Klaviyo attributable revenue? I can't pronounce it anymore.
Yeah.
Anymore? Yeah, comes from?
Exactly. So it really came. I mean, part of it just came out of, like, the idea that we were bootstrapping. We couldn't rely on status in any way to sell or anything else. We had to be very good at making the sale. The best way to make an easy sale when, like, otherwise we didn't have any credibility as two guys, you know, basically working out of a basement was, "I will show you the dollars you are making." And so from the really earliest days, it was about this data platform, but then the attributed value was the way of selling that data platform.
Yeah, yeah, yeah, yeah. Okay. And then, like, at some point then the Shopify relationship came in. Like, talk a little bit about that, what it meant and what it means for you guys now.
Yeah, so you know, early on, we realized that, like, a core part of the value was getting the data into that data platform.
Mm-hmm.
And so, you know, early on, we integrated with as many people as possible, and we had to get very good at building those integrations. So we integrated with Magento at the time, BigCommerce, Shopify, all these different platforms. Shopify at the time, you know, we, you know, we were growing fast in the SMB. We focused on the SMB 'cause the sales cycle was very quick. So as a startup, we could learn very fast by having short sales cycles.
Mm-hmm.
As we started to have success with Shopify, we still had not raised money. We knew we had to find a way to grow that was very capital efficient. And because of that, you know, online ads weren't an opportunity. We had to rely on other strategies. And so we settled on, okay, one was gonna be word of mouth. So every customer, we would, you know, go above and beyond in service and get them to recommend people. But the second was a belief that, you know, platforms could be a tremendously valuable way to grow. So we focused on agencies. So, okay, who are the other people who work with our customers, and how can we add value to them? And then we focused on platforms.
With Shopify specifically, you know, what we said is, "Okay, you know, they're never gonna do anything to help us as a tiny startup out. What are the things that they're trying to do that we can really help them with?" At the time, they were just launching Shopify Plus.
Mm-hmm.
And part of the pitch for Shopify Plus from the Shopify side was, "We're gonna open up a bunch of APIs, and you can use them to grow faster." Problem was most Shopify stores didn't have engineers on the team, so there was no way to use these APIs. So we immediately built functionality into these APIs and offered it via our marketing platform. And then what we did is every time we got a Shopify customer, we'd say, "Okay, who is the sales rep who sold you on Shopify? Can you introduce us to them?" And we'd meet with them and explain, like, "Hey, you can better do your job selling Shopify by selling Klaviyo, bringing them up in the sales process 'cause we use the APIs. You'll be able to sell Shopify Plus." And so that was kind of the beginning of this virtuous cycle.
Yeah.
And then, the deeper we got, the word of mouth expanded Shopify. The agency model expanded Shopify. And then, over time, we built a closer and closer relationship between the companies, you know, culminating in the investment a couple years ago.
Yeah, yeah, yeah. That's, yeah, that's a pretty amazing story. Yeah. Well done. And, like, bringing that now to, to today, like, where are we in terms of that kind of market understanding? And, and the reason why I'm coming or the question why I'm coming to this is if you look at the performance of you guys over the last few quarters, you know, it doesn't look like there's a downturn. Like, I mean, well, if you look at everyone else, it's like, you know, coming down, down, down, down, down in terms of growth rate. Is that, like, a function of people understanding your value proposition better and, like, you just kind of seeing it better? Like, you know, IPO gives you more airtime, etc.? Like, how, how do you guys think about that kind of difference in growth that we're seeing out there?
Yeah. I mean, we, we fundamentally think about it in the same way we did initially, and we think it's the same factors, which is if you have the data platform vertically integrated with the app on top with the messaging, and you can show the value you're generating, it makes you very sticky.
Mm-hmm.
And it makes you like a core growth lever that's very important to the businesses you work with. That both helps with new customers because they're, you know, that's a clear difference as they're coming in the door. It also helps keep your existing customers from, you know, spending on, reducing spend and moving to much more commoditized products.
Yeah.
And so, you know, we think that's honestly how all software should work, is that you've got to be able to prove your value. And the more you have that core data built into the same application, it lets you provide a much better path to value.
Yeah. And then, like, now you kind of I mean, you tried it out or you kind of proved the point in the more SMB market. But in a way, like, you should, you know, you're very powerful in terms of what you can. You should be moving up market, which I guess you're doing now. Like, can you speak to that? Like, I mean, how much is there at some point, like, a scaling question that I need to ask you on the data platform or, like, you know, talk maybe about that motion up market?
Yeah. I mean, even in the early years, just by nature of the way, you know, some of e-commerce works, we'd have brands who would start with us who were tiny and then they'd grow to be, you know, multi-billion-dollar brands. So very quickly, we had to focus on the scaling problem because our customers we sold who were small would grow up and be really big customers. Thankfully, that like, that was the exact technical problem we'd spent years solving for other brands. And so that's where we had a ton of expertise.
Now, I think as we think about, you know, kind of in the future, what we found is, like, the, the core differentiation of, like, "Hey, I need to have all my data and customer understanding in the same system as my marketing." That same problem exists in the mid-market and enterprise as it exists in the SMB.
Yeah.
So, you know, I think in the mid-market, you're more likely to maybe have, like, a data warehouse, but still you've got to file a ticket with engineering and get them to hook the systems together. But if you can truly combine those in a single stack, it just is vastly empowering to the marketer to run tests to see the results to.
Mm-hmm.
Make everything move faster.
And then, I mean, the one question that comes up a lot is, like, as you move higher up market, well, one, they're still kind of best of breed. I mean, they do think about data. So then if you talk up there, it's like, "Well, I have Snowflake, and then I have someone else on top of it, the Snowflake," you know? How do you, like, do you need to be, like, a, you know, same as Snowflake or, like, you know, how do you have to think about that kind of as you go kind of higher up in the stack?
Yeah. I mean, I think what we see is, you know, we're typically selling into marketing, and what they typically experience is some sort of Frankenstein stack. So there might be a Snowflake. There might be a CDP that sits on top of it. There might be a marketing application. And the whole thing is expensive and unwieldy to wire up and to make it do what you want, and it's challenging. So they experience a problem. The current stack is, like, not getting the job that they want done. On the flip side, like, if you think about systems like Snowflake, like, there are a lot of they're designed to be generic data stores, but to power, you know, all sorts of different workloads.
Mm-hmm.
Data science to, you know, whatever that is. What you actually need, what the marketer needs, is that customer understanding with a very opinionated data model that's built around a person. So people have facts about them. They do things. They don't do things. You need to understand all those things. And if you want to structure that data in a specific way to make all that available very quickly in segmentation and to drive outcomes, you're gonna trigger an action with that.
Mm-hmm.
Which is a different data problem than you need in a classic data warehouse. You know, every business likely needs both, 'cause it is a highly opinionated structure designed for marketing and driving action versus a, "Hey, I, you know, maybe I'm gonna run an analysis of my inventory and which SKUs and how fast they're moving and which ones the margin's not great on. I need all the data in a structured and a very generic way.
Yeah, yeah, yeah. Okay. But, like, from, like, it's fascinating because, like, I used to, we kind of do a lot around our work around, like, big data.
Mm-hmm.
Like, we have, like, Databricks coming in tomorrow as well and stuff. Like, if you think from your perspective, like, how is your underlying stack you need to constantly think is like, "Am I still on the right thing?" or, you know, "Is there something where I kind of probably might need more skill and my stack needs to change?" How do you think about that? It's like.
Yeah. I mean, that was one of the earliest problems we had to solve, and it was one of the problems we had the most experience with. 'Cause very quickly, just by nature of the business we work with, it was a very large amount of data, and a lot of those customers were growing very fast, just because e-commerce was growing very fast.
Yeah.
And so, from the early days, we had to understand, "Okay, how does that data scale? What makes this possible?" And so we had to think about how we structured and stored the data. And then, you know, the flip side, the other side of it was we had to really make sure that the messaging component was seamless. And while, you know, how do we make sure that, you know, whether that was email or SMS or, you know, in our mind, it was really we were channel agnostic.
Mm-hmm.
We didn't really care what the messaging channel was, but like we had to make sure the message got from point A to point B. And so we had to develop a lot of expertise around that as well.
Okay. And then, if you think about it now, like, leveling it up a little bit, if you think about the growth from here, like, how do you think about the different vectors, in terms of up market, international, more channels, etc.? Like, you know, how do you balance that?
Yeah. Yeah. So, I mean, we very much think of it as there's a set of short-term things where we've seen significant growth that we just, you know, we have to capture the opportunity we've seen. So there continues to be a tremendous amount of headroom in the SMB in international, you know, for, you know, even in the first few years, we found that very quickly we had customers all over the world. But, you know, we've never until recently invested in, one, providing the product in different languages, two, providing everything around the product, so, like, all the sales, support, success, docs, and other languages, and three, providing full product functionality.
So the one area of the product that requires localization is really the provisioning of, you know, providing SMS numbers and, you know, local SMS numbers that are compliant with local laws. And so on all three fronts, we've seen that there's, we've been building the infrastructure and have launched in the first set of countries, but we'll keep, you know, accelerating and, you know, speeding up and launching in more countries more quickly.
Yeah, yeah, yeah.
So, capturing that fundamental up market, we continue to see a ton of growth, something we have to focus on in the very short term. But between the three of those, there's just kind of tremendous short-term opportunities where we're already seeing growth, and it's all about how can we deliver on what we're already seeing.
Yeah. I mean, you mentioned earlier, like, the platforms you were working on with Shopify, but then you have, like, higher up the market, you have, like, Magento, you know, BigCommerce, well, maybe slightly more on the smaller scale. How do you think about, like, partnerships with those guys? And is the Shopify will the Shopify partnership eventually be, like, a, a limiting factor or, like, how do you think about that?
Yeah. I mean, you know, in the early days, we very much said, "Hey, to make this work, we've got to make the cost of these data sources, these platforms to be very cheap." And so, you know, very quickly, we launched on not just Shopify, but, you know, Magento and BigCommerce and then ultimately commercetools and PrestaShop. So from just a technical side, had to make that very cheap and easy. From a partnership side, it's been very important that, yeah, this like, it's not there's nothing in the product or no reason this would just be tied to Shopify. And so how do we form the similar sorts of partnerships with others? So we have a partnership with PrestaShop, where we're the default platform for PrestaShop in given regions.
So, you know, I think what we found, though, is that if we sell a given number of merchants who might be on other platforms, if, you know, on average, merchants are moving to Shopify. By the end of the year, we'll see a number of those merchants move from other platforms to Shopify. So, we tend to follow the gravity of the e-commerce world.
Yeah. Yeah.
As well.
Then, shifting gear on the growth levers, like, so SMS is like, in a way, you know, omnichannel makes a lot of sense.
Yeah.
SMS, you're only starting to kind of roll out now. Like, what do you see from a customer perspective in terms of is it just like, "Well, I always used SMS, this guy, so I'm now just doing new email and SMS there"? Or do they realize, like, it does make more sense? You know, you're coming from a better platform, data platform anyway. Like, so how do you often think about that SMS upsell?
Yeah. I mean, we see it very much varies by brand, and it varies by individual consumer within a brand. And so, you know, if you just think about any one person, more and more people are willing to experience, get messaging from a brand across different channels. So it might be that they want their shipping notifications to always be texts to let you know that product just arrived. It might be that for a new product, you know, you're willing to get SMS for the brands you love the most, but for most brands, you're willing to get many more emails with the new product announcements. And so, that would be. It might be one thing for me and a different thing for you.
Yeah.
And so what we see is that it's very variable. And then even in the life cycle of brands, if you're a new brand starting on day one, SMS is probably not economical. It's probably too expensive to invest in too much. So what we see more and more is, you know, brands truly are becoming multi-channel. They're even, you know, moving beyond email and SMS to try out new channels. We're launching WhatsApp in 2025, which we're very excited about. But we expect there to continue to be this, like, movement towards, "Hey, I, as a brand, I just want to give you the right channels for you as a consumer that are most likely to make you happy and ultimately drive purchases." Those might differ across consumers, but I'm not, I can't be focused on any one channel. I've got to truly play in all the channels.
And then, how does the channel help your different markets? So, like, you know, me being German, if I think about WhatsApp in Europe, it's like massive, you know.
Mm-hmm.
Like, is that kind of like you get almost a multiplier on both sides then?
Yeah. I mean, I think the thing with WhatsApp is, you know, there's the use case, you know, that we've all experienced personally of WhatsApp. At the business level, if you're not, you know, kind of like thumbs on phone sending the WhatsApp, you know, Meta runs WhatsApp much closer to SMS. And so the economics are much closer to SMS.
Mm-hmm.
It's a very expensive channel because Meta basically acts like the carrier and takes fees in a way.
Yeah.
They would from a carrier.
Yeah.
So we've seen that a lot of European brands. Some brands have figured it out. Some brands haven't. Even though the WhatsApp is so dominant as a business communication tool, it's still very much in the learning early days, but it's a channel we're excited about.
Yeah, yeah. I kind of mentioned. And then, think about, like, other extensions. If you think about, like, you know, reviews, etc., like, is there are those, you know, do you think they can get kind of bigger, or is there anything out on the horizon that we're kind of missing, like social or whatever, that we should kind of pay more attention to that you kind of can do?
Yeah. I mean, the things that we've we think about the most are there always tools like some of the ones we've launched, the CDP, which for us really means, like, some data tooling and some analytics tooling that sits top our database, reviews, which is kind of an marketing add-on. What we really think about is, you know, if our core differentiation is this data platform and the strength of, you know, the differentiation and the what takes us out of the commoditized world of, like, generic senders is take the data platform, drive outcomes with it, and then measure those outcomes. What are the other software categories that mean that don't have the data that people need?
You know, one that's top of mind is customer support and service, you know, writ large, where almost every service interaction exists separate from the underlying database that understands who this person is, which is why you have to say, "Oh, I bought this product. I, you know, it arrived yesterday, and here's the problem with it." Like, all things that, like, with much tighter integration, you can provide a much better customer experience. And so, as one thing we think about a lot is, if we really power digital relationships, what are the other pieces of those digital relationships that the same data platform can power?
Yeah, yeah, yeah. Is that not where the Data Cloud from Salesforce tries to do it in hindsight?
Yeah. I mean, I think exactly. We see people,
So.
Trying to do it. I think it's just the, we tend to see people come at it they start with the apps, and this is very true in the marketing space.
Yeah.
They start with the apps, and they try to build into the database. It turns out the hard problem is the database, so you have to really come at it from the database side and then build into the apps.
Yeah.
And the other strategy so far hasn't really worked for anybody.
Is that kind of a little bit like, well, we kind of should think about you guys more and just kind of actually, it's kind of the data side moving up market. If you solve the data problem, you actually, you know, whatever you do on top, it's not simple, but it's like it's much easier than trying to solve the data problem in hindsight.
Yeah. I mean, I think that's the insight I think we ultimately kind of stumbled our way into, which is the data is not valuable in for its own sake, but the data like, if you have the right data engine stored in the right way, that's what can power these outcomes. And, like, the apps drive the outcomes, but you've got to have the unified stack to actually drive the outcome. So without the data, you won't be able to measure the outcomes or necessarily drive the same level of outcomes. And so it really had to be unified, and so we had to solve this data problem first. It just turned out that's what we'd spent years working on, beforehand.
Yeah. Then, I'm sorry, in all the sections, I had to ask Gen AI, you know? So I almost left too late. Like, so how do you think about it in for your space? Yeah.
Yeah. So we think about AI in two ways and Gen AI in a very specific way. So, you know, our users generate a ton of content. And early on, we had close partnerships with numerous large AI companies and integrated in our products. And so we've now had years of data of seeing people leveraging LLMs to generate content. And we see that, you know, it's had great adoption. People are, you know, using it to change how they work. The interesting thing we've seen is if I am the person who goes in and, you know, designs an email or designs an SMS, if I'm gonna, you know, spend 30 minutes on an SMS in the old world, I still spend the same amount of time today.
It's just that my workflow is a ton of iteration working side by side with AI and then ending up with a much better message once I've used AI to generate a whole bunch of brainstorming. And so we've seen that be really powerful. The way we really think about AI is generative, you know, will certainly enhance the way our users work. The other piece, though, is because we have the unified data platform, how can we best understand what action is best for every given consumer? Is it three emails, three SMS, which times a day under, you know, what actions? How do we, you know, drive more revenue and use AI to understand exactly what action to take for each person and make all that automated basically in a way that, you know, it becomes more and more of a revenue engine for our customers?
Yeah. So how is that like, do you need an LLM for that, or, like, how do you forget about it?
Some of it's LLM. Some of it's much more classic machine learning.
Yeah.
It's really.
It's more like machine learning, yeah.
Yeah. Okay. It's like the combination of all of it. So the LLMs make it much cheaper over time to, like, create marketing content, and test marketing content. But the kind of fundamental underlying, like, I need to derive intelligence from, like, a large quantity of data, what we would have, like, years ago called big data, but, like, largely, you know, you can use some of the AI techniques is going to be very important to the.
Yeah, yeah, yeah. I mean, and then, last couple of minutes, I wanted to shift a little bit on more, the outcome in terms of, like, profitability, etc. Like, is like, you could argue, like, look, you started, like, you know, not taking a lot of money, like, kind of out of your garage, and then from that, you basically have, like, a thinking of, like, "Yeah, I'm not going to, you know, spend money for the sake of spending," and then just kind of drive, you know, hire sales guys, and etc. and that drives kind of the good profitability level for you guys, or is it also the data platform that gives you more efficiency? Like, how do you think about that dynamic?
Yeah. So I mean, the data platform provides a lot of pricing power. It makes our product stickier. Like, it provides a lot. But I think just our because our core DNA was, focus, you know, tightly focused on efficiency, it meant that we built a lot of go-to-market motions, that let us grow very efficiently. And we largely, you know, spent none of the money we raised throughout, you know, for a very long time. And even when we did, we spent kind of, like, small tens of millions on a couple of marketing experience experiments, but that's it. And so just that fundamental DNA has been very powerful.
I think now, as we have a business that can generate lots of cash, it's how do we take that cash and get better and better at understanding which investments and experiments we want to run and where we can lean in and accelerate. But I don't think there's no fundamentally changing our DNA of we're going to approach everything with efficiency and be very disciplined about, you know, where we are spending beyond what we're making.
I mean, and if you think about it, but if you generate kind of this healthy amount of cash, like, how do you think about capital structure and, like, I don't want to end you on the spot, you know, like, well, like, in the end, but how do you think about capital structure, M&A, like, etc.?
Yeah. I mean, it's not something we've done in the past. It's something that we've, you know, just given the fundamentals of the company can make a ton of sense, given, you know, what we see as the, you know, size of the opportunity. It's certainly something I think we'll consider over time. We think about it as, you know, like, our we are a fundamentally product-led, product-driven company, but, like, where we can move substantially faster, whether that's small tuck-ins, whether that's new market entries, whether that's, like, much larger opportunities. Certainly something that we'll want to consider over time, given, you know, the fact that there is we're in the position to do it.
I mean, do you think from your perspective, is it because, like, your core asset is, in a way, underlying data platform, so you can't really buy another so it would be more on top or kind of expertise and stuff?
That's yeah, that's basically right. So I think with everything, as we think about the product roadmap and where we're going, it's okay, right? There's what we're very good at, so it's less likely to be buying something, you know, a new data platform or anything like that because we can kind of fundamentally, that's where our advantage is. But it's, it's what fits around that, whether that's new markets, whether that's new people, whether that's new functionality that could be an app to top that.
Yeah. Perfect. That's a good closing statement as well. I'll let you go. But perfect. Thank you. That was really, I really enjoyed it, and it was actually really good to go back to the basics a little bit. Hopefully, it was really helpful for the audience as well.
Thanks for having us.
Thank you. Thank you.