Great. Thank you everybody. Good morning. Appreciate you joining. With us we have Ed Hallen, the Co-Founder of Klaviyo. Ed, one, thank you for your time today. It's great to have you here. Two, just since I think we've had less exposure to you, maybe less interfacing, would love to maybe understand your background, what your role at Klaviyo is, and then we'll dive into some questions about the company.
Awesome. Ed Hallen co-founded Klaviyo in 2012. For the first about five years of the business, we were co-CEOs. We'd both come out of product engineering at a company that was kind of a key part in the formation of what ultimately led to Klaviyo. I'd been an engineer, been on the product side, had ultimately, at this start-up we worked at, also kind of opened their European presence. Over time I've done a ton of jobs across the company. Most recently was Chief Product Officer, Chief Marketing Officer, currently the Chief Strategy Officer, which really means stretching across the business with probably a spike on the product engineering side.
Got you. Look, I'm happy to have you here because we're going through this transformative era in software and technology more broadly. I'd love to, given the different hats you've worn at the company, how are you thinking from a product and strategy perspective on what the company needs to do to embrace this new AI world that we're in? What are the key focus areas right now for Klaviyo from that perspective?
I think really from our perspective, this age is kind of the most exciting tail end. When we started Klaviyo, before Klaviyo, we'd met at a company called Applied Predictive Technologies, where we were early but junior employees. Fundamentally where we spent our time was building large-scale customer sources of record for people like Walmart, Bank of America, kind of very large organizations where we pulled together all of their data and then the sales, inventory, staffing in the stores, kind of like everything possible into a singular customer view. Then enabled these organizations to run effectively large-scale analytics across it. When we started Klaviyo, the basic idea was like, hey, if you look at all the different software categories that exist, at the time, largely they existed separately from data.
What if we built these customer sources of record for businesses of all sizes and all types? Through a set of kind of twists and turns, ended up saying, okay, we knew we were going to bootstrap, so the best way to go to market was actually somehow take that customer source of record and provide a way to drive outcomes atop it, so we stuck marketing atop that. Then of our first 10, 20 customers, looked at them and said, okay, we know we have to focus. We don't have infinite resources. That's when we said, okay, hey, three of the 10 are e-commerce, leaned in hard to e-commerce.
Through that lens, kind of the way we've always thought about ourselves and the platform is what's differentiated is this data platform underneath and our ability to provide that infrastructure in a way that makes it very usable in a very low latency fashion. The ability to, as our customers take marketing actions, to understand exactly what happened. A number that we call KAV. Over time to add intelligence. The more consumer profiles we have, the more company data, the more outcome variables that we have, the more we can understand, hey, when we do X, Y actually happens. For many years, I think we've been thinking about ways that, okay, it's not just humans providing the intelligence atop our platform, but it's also AI or machine learning or other tools atop that.
In the world of LLMs, it's really just another set of tools atop that infrastructure, both the data infrastructure, but also then the customer sending compliance and other infrastructure. It really couldn't be more exciting. What we've seen is it's yet another way for our customers to drive more dollars of revenue using the infrastructure that we provide.
I think that it's an important point where when we get asked a question about what the future of the application tech stack looks like, that there's a set of companies that are, in a way, the infrastructure of something, even if they're not an infrastructure provider per se. As you think about that and you think about the argument that sometimes gets made that while the value may go into the agent or into the LLM, can you help us understand what the moat is for Klaviyo, where it will continue to be actually the most relevant layer even as agents are built on top of it? I think double-clicking on that would be very helpful.
Yeah, totally. Our end customers are effectively optimizing for their own growth. The reason they rely on us, one, is it brings together all of their different data, everything historically, but also we're pulling in 4 billion data points every single day across nearly every consumer profile in the U.S., as well as a massive number globally. Coupled with that, leveraging the infrastructure to understand, okay, who can we message and when? What happens when we message this person or this type of people in the past? What are all the affinities for different types of messages? Kind of a completely unique set of data that exists atop that, but also then a set of pipes, things like deliverability that we've invested in for a very long time to actually drive outcomes from these customers.
What we see effectively our customers doing is they add in things like LLMs, they use some of our new AI products, is effectively they're driving higher actual growth in their own business. It effectively lets them do more, but they're still taking advantage of the same underlying data, and they're taking advantage of the same underlying infrastructure too to provide the sending, but at a much faster pace, and they're doing more in ways they couldn't otherwise. One thing we've seen is we've leaned into our MCP, and we've seen that actually customers who use the MCP heavily drive 16% more dollars of growth out of the product. What we tend to see is it's not replacing usage of our product, it's actually they're running the MCP atop it. They're running our existing tools.
They're also r unning some existing, because some of our other AI products atop it, and kind of collectively, it leads to more value for the customer.
I do want to pull on that thread mentioning the MCP. I know that you rolled out the Claude integration recently, and that allows prompt-driven marketing workflows. What are the puts and takes of opening up Klaviyo's data to external AI platforms, whether that's OpenAI, Anthropic, or others? I don't know if you're working with open source LLM providers as well. How do you balance that as a benefit versus the potential perceived risk around it?
I think first, philosophically, the way we think about it and what we've seen play out is, the more that we enable our customers to use our platform, the more value we see them generate. This has played out in that, seeing people who adopt the MCP actually drive that 16% higher KAV, kind of our Klaviyo attributed value. We've seen that be true, and when we dive into the usage of exactly what people are doing is we tend to see that it's opening up Klaviyo to different personas and different use cases. That might be somebody in a different role on the team doing more analysis atop Klaviyo, where they're pulling in both some of our customers' specific data and then also data across other tools. It might be that they're using it to quickly interact and set something up very quickly.
We also see that it's not changing how much people are logging into the tool. We're also seeing that as we launch new products, so we have a product called Composer in beta, which is effectively a kind of the next-gen evolution of our marketing agent. It's a kind of end-to-end, effectively, intelligent coworker that sits side by side with our users. They can do things like say, "Hey, I want to drive this outcome. What's the campaign I should build?" This will actually go and provide that intelligence, build the campaign, set up all the content.
Afterwards, it can go and say, "Hey, I want to understand, am I sending all the right emails, all the right SMSs, all the right WhatsApp?" Audit everything and figure out where there's additional opportunities, kind of that sort of intelligence, and built atop both our unique data, and everything we have, but also highly leveraging the models. What we see is that opening the platform up is kind of additional usage atop these other things and just additional value. I think the way we think about it is, in the world where whoever's acting atop this customer source of record and atop this infrastructure, enabling them to do more and drive more growth, it's just a different portal and lens into doing that we expect to only grow over time.
You mentioned Composer, and I want to understand more about how, I know you mentioned an example of you can launch a marketing campaign or what type of campaign should I launch, but what are the most common use cases of Composer, right? Essentially it feels like a digital twin for a marketer, right? What are you observing and how are customers feeding back value in terms of the value that they're achieving out of Composer to you?
Yeah. When we tangibly think about how people drive value out of our product, it's identifying a set of customers or a set of behaviors and then reaching out to those customers or changing their experience. Now, that could be outbound messaging, things like email and SMS and WhatsApp. It could be web personalization. It could be more and more people using our service offering and our Customer Agent to actually change how they're interacting. That's generally how they drive value with our product. Now, if we interview all the marketers who use our platform, there tends to be no shortage of ideas, and there tends to be things that they know they probably should be doing, sets of customers they should be treating differently, where they're not.
They're not because historically they just don't have the human time and bandwidth to go and set that up. What we see with Composer is that they can actually do way more things way more quickly and kind of take the, where previously they may have had a larger set of customers who they all treated the same, they can even further subdivide to personalize against those customers and to truly personalize those journeys across it. We see that that's a very common use case where it's effectively giving these human superpowers where they can be two, three times as productive as they would've been. The second very common and kind of very valuable use case we see is effectively understanding, hey, based on a much broader purview of what people are doing, am I doing the right things?
Are there key ideas that I haven't thought of for my brand? Are there things that are sub-optimized because I have set them up one way, but if I kind of take advantage of best practices, they could be set up a different way. We see that it's pretty common to go in and kind of continually tweak what they're doing currently to understand new opportunities or kind of missed opportunities as well.
Right. You mentioned Customer Agent, and I know that's a part of the new service offering. You entered that vertical only last year. How does service fit naturally into the overall core marketing function? Or why does it pair well with it, and how has adoption trended versus maybe your initial expectations?
We really think that the historical separation. It was never optimal. Right? If you think about the digital relationship of a consumer, they don't think about it as like, "Oh, this is a marketing interaction, this is a support interaction." Right? What we've seen is, when we thought about service, it was really how do we manage the entire digital relationship? In any interaction, parts of that will be marketing. I've bought things in the past, I'm thinking about what I might buy in the future. Parts will be service. Hey, I'm returning this item, and I want to either get a credit back or actually buy something as quickly as possible. It fit very neatly into that, and this is effectively kind of what we've seen play out.
It's all based on the value and kind of the right to win, is that everything is still based on this unique customer source of record and this unique source of customer understanding for a given consumer, so we can actually base it off a single unified profile, which is a unique position in our market. In terms of adoption, I think we've been very excited about the progress so far with the adoption we've seen. It has skewed larger, and I think we're seeing more and more over time that we're actually landing more and more people with a unified stack across the complete suite. Super excited about it. There's a clear roadmap of things I think we're really excited about, like agents building agents atop the infrastructure. More and more customization, multi-language that's at the top of it.
A set of things that we'll just keep pushing on to drive increased adoption.
Before I switch gears, I want to ask another technology question, which is, as a technologist, as you think about the background of the company, what does AI inside of Klaviyo itself look like today? How are you guys utilizing it?
Yeah. The way we've thought about it for our customers pretty similarly, it mimics how we think about it for ourselves, which is fundamentally it's a tremendous enabler to let us do more things more quickly and to a higher quality. As we think about it, there are obvious implications in engineering, and so that's been a bunch of time and effort invested and just complete change in how we work, and massive changes in terms of just the amount of code that's coming out of models versus being written b y humans themselves. Massive changes there. In terms of one step away, how our design functions and product functions work, where way more likely go from what may have been slide decks years ago to all of a sudden you're very frequently building prototypes and building those and throwing those away.
We've also just seen that kind of with time and attention, we've really been able to rally the whole company around how do we use AI to drive more efficiency, but better results of everything. It's the people team, it's the finance team, it's across everyone. We've built a kind of center of excellence effectively in the product team that's working across all those functions and actually launching features and functionality with them within every function, within every leader. Also driving a bunch of change management of just how do we actually leverage tools and kind of revisit a lot of the actions we're taking from first principles.
Got you. Let's maybe go in a slightly different direction. There's this technology transformation happening, but there's also been, I'd say business transformation inside of Klaviyo, right? Late last year, Chano came on to the Co-CEO role. I'm curious, what was the objective behind that, and what are the changes that are being implemented as a result of that move?
Yeah, totally. At first, we'd done this before, right? Andrew and I have been Co-CEOs for a long time. We knew that he works extremely effectively in that relationship. In a world where he could spend more and more time leveraging new tools and more time on product and engineering to do more and more. Tremendously exciting thing for Klaviyo, tremendously exciting thing for him. Very excited to kind of drive the pace. Had thought about this arrangement over time. There's a set of things, we'd worked with Chano before on the board, that we knew kind of what that partnership looked like, and we were excited for him to accelerate. If we think about some of Klaviyo's key growth curves, we're early days on enterprise. We're early days on global.
We've seen tremendous amounts of success on both of those, but big opportunities to further ramp. What we've seen is Chano has done exactly that. Kind of changing how we operate, changing those sales go-to-market motions to be in front of more CMOs and CTOs as we think about how we actually sell the product on the enterprise side. Bringing in leaders like our new Chief Revenue Officer, Eric Fearday . Just a lot of discipline. We brought a new APAC leader. We're seeing a significant expansion, have really grown our Singapore Hub. Just really a rigor across how we think about go-to-market that we've seen play out, and I think we're very excited to play out further with the partnership with Chano.
As you think about the leadership changes, does that change the profile of the sales rep that you're targeting or the type of person that you're helping on the front lines from a go-to-market perspective as well? How should we think about that?
Yeah. At first, in terms of profile of sales rep, I think this is a just implication of how we think about every person at Klaviyo over time, which is in the world of tremendous new tools and tremendous new capabilities, we really see it's more and more important, one, that people are highly curious and able to learn fast. We call this high slope, and it's one of our kind of core hiring values. Two is that none of this has changed the value on the sales side of kind of like being great at the human aspects of sales. Like we can pull more and more of the sales ops, we can pull more and more of the admin work off someone. It really lets the best salespeople be even better. Those things have changed significantly over time.
In terms of the profile of who we're selling to, I think the biggest thing is, you know, as we become more of a multi-product company, at every level, there are, you know, potentially more stakeholders as we think about service. Also in terms of in the enterprise, I think we see more and more, it's a different skill sets of reps, different experience, different, you know, people that you're dealing with on the other side of the persona. It's also been building out those teams, and thinking about also the leaders on that side, that are, you know, truly excellent at, building these enterprise relationships.
Understood. Then on the international side, another key focus for the organization and for Chano, I know the growth has been very good there. How do you unlock more success there? Is it about localizing the product in a way? Is it about getting more evangelism with the merchant customers into, in that region? What further unlocks growth outside of the U.S.?
There's a set of pure execution things. Things like we'll enter Japan, things like, you know, adding some of the leaders we've added in APAC, things, you know, like building out the capabilities of the sales force. There's a set of things that we're very focused on on the go-to-market side. There's also a clear set of things on the product roadmap that are kind of pure unlocks to more growth. These are things like the multi-language support, increased multi-language support in the Customer Agent. It's data residency, so kind of like where our customer's data lives. Over the past year, we've made a big push into portfolio and how global brands operate across customer profiles in different languages and different countries.
We'll keep, you know, pushing against those product and engineering areas. But it's kind of part of what has contributed to us being, you know, we've seen tremendous success growing internationally, but I think makes us very excited is that there's a bunch of just untapped opportunities that we've, you know, now either executed against or need to execute against and need to drive the results against, both on product engineering and on go-to-market.
Great. You know, we're late into the conversation, but there's this little company that you guys work with known as Shopify. You know, how do you think about what Shopify is doing in agentic commerce, and how does that inform maybe Klaviyo's own work in that agentic commerce world, and how are you thinking about that?
Yeah, it's exciting. The way we thought about Shopify initially, when we were first getting started, it was, "Hey, we can grow in a very capital efficient way if we find businesses we can partner with and we can drive even more value for them." The setup for us has always been, okay, if you take a customer relationship, so someone who's, you know, has a, as they bring a customer in the door, and then they actually provide the infrastructure to provide that checkout or that transaction. How do we ensure they can get more and more growth and deepen that relationship with that consumer over time? That's kind of largely how the Shopify relationship has worked and why it's been so mutually beneficial historically.
As we think about agentic commerce, I mean, it's a different way that consumers are entering the funnel. We've seen, if we look up across our customers, we surveyed our almost 200,000 customers and saw that the average brand is seeing about a 2,000% increase in the number of customers coming to their store via an agent or via, you know, the models themselves. It's the same ultimate dynamic of once you've started to build that customer relationship, you really wanna turn that into an owned relationship where you're building a long-term relationship with that customer over time and driving future transaction and future engagement with those consumers. That is effectively the same. You know, we see that yes, like maybe a different proportion are coming from other platforms or places they may have historically.
Maybe it's, in some cases, it's just a net increase in consumers. The number of relationships that are ultimately being further built with Klaviyo really is only increasing.
Understood. Maybe a couple of financial questions to close out. One of the questions we got a lot after the recent earnings was the decision to absorb some of the A2P fees. Help us understand what the strategic rationale behind that was and the decision going forward to now pass those back through.
Yeah.
What the shape of that looks like.
Yeah, I mean, the overarching landscape of like, "Hey, these fees are changing over time and will change," is constant and will only continue to be constant. We made a choice, originally kind of one of our core values being very customer first, positioning the pricing model very simply, where we've kind of historically absorbed these fees. That, you know, I think for, over time, we've kind of always kept an eye, you know, with the thought that, hey, you know, some customers, that's just gonna drive more usage. Kind of like as we think about that, price point, some customers consume more. As you raise the price point, customers consume less.
We've always, you know, had an eye on that as something that we might consider changing. I've recently just, you know, have kind of made the decision to start to pass more of those costs through. You know, I think we think about how customers think about that. It's like, I think we've, as more and more customers have more and more experience in these types of messaging, they understand how these work. They understand the fee changes. It's a much more established part of the market. We won't go back and kind of recoup or change anything for how we've done it historically.
Going forward, we'll build that in, and think that kind of like as people are more comfort in the channel, is something that kind of people are very used to and used to, you know, how they'll think about it going forward.
Well, Ed, we'll leave it there. We appreciate you joining us today. Great to dig deeper into Klaviyo. We look forward to talking to you in a few months.
Thanks for having me.