All right, everybody. Thank you so much for joining us in the post-lunch session here. We are excited to host Klaviyo. We've got Ed Hallen, Co-founder and current Chief Strategy Officer of Klaviyo. Ed, why don't we start out? I should mention Jackson Ader, Enterprise Software Analyst here at Q Bank Capital Markets, and again, we're hosting you guys at our Technology Leadership Forum. Why don't we start out with the run of show? Ed will introduce himself and the company. We'll do some questions. I will ping you guys. If you have questions on topics that, if there's a follow-up that I should have done and I didn't, please raise your hand. I'll ping you a couple of times so that it's not like we have 10 seconds left. Who has a question? All right.
Great.
Go ahead.
Perfect. Yeah. Klaviyo, we power digital relationships, primarily for B2C businesses. Tangibly, what that means is for the businesses we work with, we are the customer source of record. We're the system that understands, you know, for each customer, everything they've done on the website, everything they've bought, everything they haven't bought, all their support interactions, everything they've returned, their loyalty points. We unify all those points of customer data, and then we use it to power all their customer touchpoints. We started in 2012, started doing that. Before Klaviyo, my co-founder and I, his background had been in building these large-scale business intelligence systems at another software startup.
The primary problem we were focused on solving was unifying kind of all the siloed data systems that, you know, large companies like Starbucks and Walmart, Bank of America's of the world, and took that and said, hey, why can't we democratize this and make the same sorts of technology available to businesses of all sizes? At the time, we started with email. We basically said, okay, hey, if we build this customer source of record, what's the best way to use that to actually drive revenue? Started focused there, leaned in pretty early on to e-commerce as the first beachhead to start with. Fast forward 12 years later, still fundamentally the focus, it's being that customer source of record. Recently launched, we call the B2C CRM, but it kind of really is a way to describe what we've always done, which is powering those relationships.
More and more, we're starting to power those relationships, not just for marketing, but kind of really throughout the entire customer journey and entire customer experience.
That's great. Just given that you were there at the very beginning, I think it's important if we just double click on the secret sauce of the data ingestion and how you guys grew up really as a data asset before it became, like you said, you made a conscious choice. How are we going to monetize this thing that we have? The thing that you have, the data intelligence, why was that part of the secret sauce early on? What was your background with AB that led you to that being what you guys were able to do?
Yeah, so, coming out of college, we both went to work at a company that ultimately sold to MasterCard, that was building large-scale testing software for these large, large companies with huge numbers of consumer relationships, both in retail presences and also online. The core data problem there was, if we, there's a huge volume of kind of highly opinionated data that is fundamentally transaction data. This person did this thing at this point in time and all the data about that. For example, if I walk into a grocery store, it's not just that I walked into a grocery store at a specific date and time, it's that, okay, here are the items I put in my cart. Here's how I checked out. Here's the credit card I used, everything else we know about that customer across all the touchpoints.
It ends up being just a tremendous volume of data. The problem we knew how to solve was, hey, how do you work with these huge volumes of data and how do you store them in a way that they're highly usable, which is very different than a lot of the classic large-scale, at the time, what we called big data problems about, hey, how do we just store the data in a way that you can pull it for data science and any ad hoc way?
In this case, storing the data in a way that we could directly drive immediate actions from the system, whether that was a segmentation that could then push that data into an ad network, like a Meta, to actually target and tailor the ads, or it was triggering an email based on what the customer has and hasn't done at a specific point in time. It was leveraging the data so that it was highly performant, it was very quick, and it was also very easy to get these huge volumes of data in. It was very usable by anyone without the need for an engineer. At the time, if you wanted to do this, and most brands didn't, it was kind of a heavy engineering lift. Even now we see that most marketing systems are built from the messaging side first.
Actually unifying the data typically ends up, you have to unify a bunch of different tools. It's just kind of a Frankenstein where you're tying together your database, your customer data platform to your messaging platform, and each of those steps requires engineering work. Fundamentally, the realization at Klaviyo was to build a vertically integrated stack that includes the data in it in a highly optimized way. It lets marketers drive a ton more value and ultimately kind of anyone who is building that digital or digital customer journey.
Okay. Applied Predictive Technologies, right? For those in the room. Yeah, that was the company you guys kind of grew up with. I remember that. Yeah, from a prior life. Chief strategy officer, I'd like to keep things a little bit high level just for now, given that you set the strategy of the company. You mentioned B2C CRM. Mm-hmm. You are kind of quickly becoming a multi-product company. You've been what I would say is like multi-channel, right? Email, SMS, push notification, but like there are new service products, marketing automation and analytics layer laid on top. Can you just talk about the difference between multi-channel and multi-product and still how it turns into whatever B2C CRM?
Yeah, yeah. I mean, for us, it's really, at its core, just a better way to articulate what we've always done, which is, look, if we, from the customer standpoint, if we each think about how we interact with brands, we don't think about like, oh, this brand I just want to get emails from, this brand over here, like they're the SMS brand. It's that actually we have touchpoints with brands that happen in all different ways. They might happen on the website. I might get new product launches in email. I might get an order confirmation or shipping confirmation via SMS. That might be one way, one sort of notification for the company that makes this shirt. It might be a different one for a different brand to interact with. For consumers, we very much have these journeys that extend across brands, across different channels.
It turns out they also are, and they're not just tied to marketing or service. In any interaction, I'm likely going to buy from that brand again in the future, and if I have an issue with a purchase, I might actually, in that same interaction, think about new products I might recommend while I'm also in the process of, say, returning something I didn't like. The way we think about it really is like the next step in just unifying the entire customer journey versus kind of having a purely disparate, this is the tool for these people that drives marketing. It's no, actually, how do we drive the entire set of relationships across all the digital touchpoints, whatever they are?
Okay, what's still left to be filled in in terms of when you're going this way?
Yeah. I think first, product-wise, the service products, there's kind of three initial SKUs we've launched, all currently in beta, that will continue to, we'll launch in GA towards the end of the quarter. A ton more enhancements that we'll keep launching there. Channel-wise, we are very much channel agnostic. As there are new channels that consumers and brands are adopting, we'll continue to layer channels atop our platform.
What do you, what's it? Channel in this sense is like email, text, SMS, like.
could be messaging, could be something on the website, just any type of touchpoint that we're having. It could be a piece of mail that arrives to your doorstep, kind of like anything that the customer ultimately receives that is a touchpoint with that brand. It could be at the register in a retail store.
Channel has a number of connotations in software. That's just.
Exactly. I think fundamentally it is those sorts of journeys. The core of our business is e-commerce at the moment. Still, about 5% of the business is non-e-commerce. Over time, we fully expect this to really extend into not just brands of all sizes, but also any digital relationships that extend outside pure e-commerce as well.
While recognizing that a lot of these are still in beta and it's not fully fleshed out, is the expectation that you will say, okay, this is a package of products that we are looking to land with you, you know, shirt maker.
Yeah.
Is it going to be, you know, is there going to be one kind of tip of the spear, and then you sell some of the other products over time and kind of stack them on? Basically, is it going to be larger lands, or is it going to be that land and expand model when you think about making money from the multi-SKUs?
Yeah, it's a great question. We're very much in early days and learning. We've, the Customer Hub has been out the longest. The others are newly in beta. Nothing's in GA yet. We very much expect it to be, we've very intentionally seen and decided that customers need to be able to buy pieces of it individually as well. The Customer Hub effectively is a layer that can sit atop web pages and provide that level of personalization. It shows all the discount codes you've ever gotten. It shows your recommended products based on what you have bought in the past or if you haven't bought. If you've recently had a transaction, you can easily say return something. It kind of unifies that exact, personalized website to exactly fit to you in a way that sits atop a brand's website. It's a one-click implementation.
You just flip it on and it's on immediately. Very easy to adopt, and can very much be adopted in a product-led fashion. Versus a help desk, which, we think there's tremendous power in unifying the two, but there are going to be other incumbent pieces of software. I think we know and expect that, "Hey, pieces of this will get adopted in different mixes by different companies, though the full power is really adopting, ultimately adopting everything over time and land and expand."
Okay. I think one of the things that, I don't know, it might be fully appreciated, it might be underappreciated, but you're one of the few companies in software where it's like solopreneurs, right? Like true self-service, one person, and then it scales up to some of the largest brands that, you know, on earth.
Yeah.
If you think about whether it's the multi-product strategy, is it really, I assume everything at this point is kind of aimed more toward growing that upmarket presence. Is there any differences I should think about, kind of the segment demand for what you guys are rolling out?
Yeah, no, I think the way we thought about it in the beginning, we came from a background of building data systems for incredibly large companies. The upshot is, solving the scalability piece was something we very much knew how to do. The reason we built for SMBs is we were bootstrapping and we believed we could grow very capitally efficiently, starting with people who could adopt very quickly. We started with SMBs in a highly product-led fashion. What we found is the core pitch of, take the database that has all the pure customer understanding, drive, power all the digital relationships on top of it. That same pitch is equally resonant, as long as you have some data and some customer relationships, is equally resonant in the SMB as it is upmarket. In fact, upmarket, you can actually drive a lot more revenue through it.
We've still found it's very much, you know, the sales process is different, the customer support process is different, how we're structured internally is different. It's the same pitch and the same kind of reason for buying. Where we're not focused is on the true day one entrepreneurs who don't have customers yet. Where you don't have data and our data platform therefore is not that valuable. That is not the place where we can drive the most value. Everyone we work with tends to have some sort of incumbent who may have done the messaging on those days until they get their first sale.
I have a follow-up on that, but we're going to switch topics to AI after this follow-up. If anybody has any questions ahead of that, be thinking of them. I'll ask you in a second. You kind of characterize the starting with SMB, not as an accident or like, but kind of your circumstances dictated that choice at the time.
Yeah.
I know that this is probably hard to answer now that you guys are worth billions, but if you could go back in time and switch that to, hey, we have this data asset that works really well for gigantic companies, do you think it would have been, I don't know if it's better, but would you have preferred to start with high-end customers and maybe work your way down rather than what Klaviyo has done, which is start low-end and try and work your way up?
Yeah, we actually tried or we thought about that and we experimented with it. The thing with big companies that we'd seen a lot with our last company was that if you're selling infrastructure to big companies, one, it's a slow sales cycle. It's a long sales process. Two, it's a lot harder to sell infrastructure than it is to sell revenue. The belief was, our belief kind of when you're, as people bootstrapping a software business, the thing we needed was a lot of learning so we could generate as much revenue as possible as quickly as possible and not as quick at jobs.
You're right.
The pace of learning by running all those reps at the SMB with an SMB was tremendously valuable and was incredibly intentional. I just don't think it would have worked without that. We would have had to build a very different business. We would have taken a lot more venture. We likely would have built a business that was way less capital efficient because all of a sudden the incentives would have been different. By focusing that way, we built an incredibly capital efficient business and then we could focus on scaling that into other markets over time. Even today it gives us a lot of power because we can kind of like keep doing the same thing over time.
Sure. Okay. Quickly, any questions on what we've hit so far before we move on to AI?
Sure. As you've moved up market, you talked about customer service, just like obviously customers demand more. The handle of the investments and changes that you guys made in the organization, in order to cater to those larger customers. Maybe just transition to the AI also, like does leveraging AI then enable?
I'll quickly recap just for the final test. The question was around, have you had to make any changes internally about customer service as you've moved up market and does AI make that any easier?
Yes. We think about this in two ways. One is we're continually investing to improve the experience of everyone of what's product-led. Wherever we can provide more, you know, quicker path to sale that doesn't necessarily involve a human, we'll keep doing that. Wherever we can provide better customer experience than involving a human, we'll continue to do that. For a lot of customers, we'll keep investing in there. As more and more larger customers, we stretch out market. We definitely, you know, there's a different, we continue to invest in different sales motions, different people, different types of support, different ways of supporting those customers. We will keep doing that. It's very much an optimization of keep investing in product-led improvements that help all customers. At the same time, that lets us focus more time, energy, and resources on the bigger customers.
As we think about AI, it certainly makes it easier, but we really think about, as we think about managing the business internally, fundamentally, every single leader in every single function, we've pushed to say, you need to be thinking about every way that AI can make your business more efficient, whether you are in finance, whether you are in the people function, whether you are in customer support. We expect, you know, writ large across the business that we'll just see a lot of dividends to like, you should be able to do more with less. Culturally, it's just kind of how we've pushed leaders for a long time, but now there's just a lot of tools. Using those as part of every Klaviyo's job is to figure out like, hey, where can I apply AI to either get better outcomes or to drive more efficiency within the business?
What about AI as a product? How are you productizing it?
For a long time, because we have this central customer source of record, we've used AI as a way, just as features atop that. A great example is using AI to help write people's text messages, help them generate the content for email, help set up the flows or the automated customer journeys that are in our product. That is one area that we'll keep investing in. The way we think about that very much is using Klaviyo is very much about driving what we call Klaviyo attributed value, which is maximizing the revenue you get by using Klaviyo. We'll keep investing in AI both ways. Some of that will just be built into the platform and already is built into the platform. Some of that over time could be additional products sold atop Klaviyo. One example that we're starting to use is this conversational agent that sits atop Klaviyo.
It's kind of like, in its purest form. There are even things like the Customer Hub atop that, there's a bunch of AI embedded to provide the experience within that. All of it, our fundamental belief is that in a world where we have this kind of deep customer understanding of the touchpoints and all the experience within a customer, but then across customers of what's worked and what hasn't in driving outcomes, AI is a huge tailwind because we can use it to both drive more personalized customer journeys. Across customers, previously, yes, we did some of that, but you can all of a sudden have unique customer journeys across thousands of people and create unique content across thousands of people that you could never have done without some of the new AI technologies.
Is AI a threat to your actual product at all?
No, it's really like, as we think about it, a pure enhancement because the value to really drive a ton of value with AI, the value is in having that customer understanding that lives in the data. It's really purely by being able to automate more and do more with AI, you should be able to really drive more revenue from the underlying data and from the underlying customer understanding.
What about internally? We're kind of asking a common question these couple of days. How much do you expect your internal software development to be done by AI in the next two or five years?
Yeah, I mean, I think we, and I think every company like us, is in early stages in learning. I think if you think about engineering and the most common technical problems, the highly differentiated piece of what we do is very unique to Klaviyo, right? The piece that we had a ton of experience with, the piece that we've now spent many, many years building, is in building this customer source of record. AI may be helpful in some of that, but a lot less helpful than it is if we think about building the interface for a new feature. If we think about, hey, we're going to launch a new way, a new customer journey builder, and we want to build a new interface atop it, there's a lot more software engineering there that AI can really help power and help let us do faster.
I think we believe AI will let us do more with less, but we think there's a ton of, we still have just a huge amount of room to run to invest in the things we can do with the data we have, both in individual customers, but also across all customers. It doesn't, you know, it's not, I think we're, it's very much not a threat, very much a tailwind of, hey, it gives us more that we can power atop it and doesn't, for the differentiated part of the business, doesn't replace it.
Okay. Switching to international expansion, that's been a big, you know, chief strategy officer, a big strategic initiative really since the IPO and prior to that. How is the international build-out going to look relative to what the domestic one was in your expectation?
Yeah, I mean, I think there were a couple of key things that were missing. I think it's the things we've talked about for a while now, which is, you know, we knew having the product and language in more countries would matter. We knew that having the messaging channels, like the SMS provided in more and more countries would matter. We knew that having sales and support and people closer to the geographies would matter. We found that it's all ended up being true. We kind of have been driving that playbook over the past few years. Now, we'll keep leaning into the same playbook, but there's a tremendous amount of market internationally that we're very focused on growing into and supporting.
I mean, a big part of your growth in the U.S. was Shopify, right? Especially early on. Is there, can you even replicate the same playbook in those other countries?
Yeah. The thing, you know, when we started, we basically said, hey, we're going to lean into the data sources that give us the data into our platform because we're bootstrapping. We need to get customers in a way that's highly scalable without us spending money. We leaned into the platform partners. Our earliest integrations were Shopify and Magento and Salesforce actually and BigCommerce. What we found is, if we landed 10 customers, they'd be a mix across all the different platforms. Over time, more and more of the center of gravity moved into Shopify. That's remained true over time. Even today, if we land a bunch of new customers across platforms, as Shopify does well and kind of converts people from other platforms to Shopify, our mix ends up being more and more Shopify heavy.
I think what we've seen internationally and expect to see internationally is the same thing plays out. Wherever the center of gravity of the e-commerce market is, we expect that that's what it will really match. For us, they're very much data sources and people we can work with very closely in a very symbiotic way. If we drive more GMV for a merchant, that's very good for Klaviyo. We have more profiles in our system that we're monetizing, but also the e-commerce platform is doing the same thing. They actually make more money off the GMV as well. In some ways, as Shopify internationalizes, that's a tailwind for us. There may be Shopifys of specific geographies, but it's not strictly required that it looks exactly the same way as it did with Shopify.
Okay. I've got a couple more quick hitters. I'm sorry that I'm not going to come back to the audience. Who's in charge of pricing strategy over there? You guys are making an awful lot of changes. When's that going to stop?
Yeah. The way we've always thought about it, first, our guiding metric is, hey, how do we drive what we call Klaviyo attributive value? How do we maximize the value that brands make off Klaviyo?
Okay.
When we started in the very early days, we said, look, we're going to drive a ton more pricing value than the incumbents. We're just going to take the mail chain price, we're going to charge 25% more, and that's our pricing. Honestly, we grew so fast that we didn't touch it for eight, nine years. There was a totally unoptimized setup; actually, the guardrails within the software didn't even enforce lots of the pricing. You could just, if you wanted to be on a different plan than you actually, with the pricing set, you could be on a different plan and still kind of use the full product. Over the past couple of years, I think there's been obviously new products we've launched, but also things we've brought our stated pricing in line with how it was implemented in the product.
A lot of this was just the process in the last few years of aligning on that pricing. Over time, we fundamentally believe pricing has to be directly tied to value. It has to be clear wherever possible, and that's what we think that will look like now and what we'll look like in the future. We're really trying to provide a seamless, smooth customer experience too, so they know what they're signing up for and getting into. That's one reason why we invested hard in bringing the pricing in line with where there are differences in how we're saying it works and how it's stated to work and then the implementation of the product. We've got to have everything be simple in terms of the entire setup. That's led to more of the pricing changes over the past few years.
Okay. We'll get you out of here on this. That your exposure to, you know, not necessarily revenue from China, people that import goods from, you know, from China and then sell them on the internet to people like me. What have you been able to size that exposure?
Yeah, very low, basically. We recently listened to our sales, across all sales calls for mentions of tariffs and found that it was in less than 2% of sales calls. Of our brands, the average order value across all our brands is $50. If you think about most of high growth e-commerce, this ends up being true for the large part because you actually have to pay the customer acquisition cost. You're building a differentiated brand. You end up being the types of products that are oftentimes less exposed to, they're not pure commodities in the same way that others are. We've just seen it having minimal impact across the business.
Okay.
Yeah.
All right. Ed, thank you. Thanks everybody. Thanks for joining us. This was great. Awesome.
Thanks.