Close sidebar

Clear
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Ep 01
Dylan Fox of AssemblyAI
Coming Soon
NOW LIVE
Ep 02
Daniel Sturman of Roblox
Coming Soon
Coming 10/24
Ep 03
Mike Murchison of Ada
Coming Soon
NOW LIVE
Ep 04
Shensi Ding of Merge
Coming Soon
NOW LIVE
Ep 05
Alexandr Wang of Scale AI
Coming Soon
NOW LIVE
Ep 06
Jack Krawczyk of Bard
Coming Soon
NOW LIVE
Ep 07
Victor Riparbelli of Synthesia
Coming Soon
NOW LIVE
Ep 08
Cai GoGwilt of Ironclad
Coming Soon
NOW LIVE
Ep 09
Daniel Yanisse of Checkr
Coming Soon
NOW LIVE
Ep 010
Glen Wise of Cinder
Coming Soon
NOW LIVE
Ep 011
Kate Parker of Transcend
Coming Soon
NOW LIVE
Ep 012
Rene Haas of Arm
Coming Soon
NOW LIVE
Ep 013
2024 AI Predictions
Coming Soon
NOW LIVE
Learn more about Season 01
Ep 01
Andrew Bialecki of Klaviyo
Coming Soon
Now Live
Ep 02
Vlad Magdalin of Webflow
Coming Soon
NOW LIVE
Ep 03
James Theuerkauf of Syrup
Coming Soon
NOW LIVE
Ep 04
Suresh Vasudevan of Sysdig
Coming Soon
Coming 4/18
Ep 05
Coming Soon
Coming Soon
Ep 06
Coming Soon
Coming Soon
Ep 07
Coming Soon
Coming Soon
Ep 08
Coming Soon
Coming Soon
Ep 09
Coming Soon
Coming Soon
Ep 010
Coming Soon
Coming Soon
Ep 011
Coming Soon
Coming Soon
Ep 012
Coming Soon
Coming Soon
Ep 013
Coming Soon
Coming Soon
Learn more about Season 02
Season 01 • Spotlight on AI
Season 02
Episode 01

Klaviyo’s Andrew Bialecki on proving a tech startup can be built anywhere

A conversation with the CEO and Co-Founder of Klaviyo

When Klaviyo became a public company in 2023, we celebrated a moment over a decade in the making. In this episode of Spotlight On, Andrew Bialecki shares learnings from Klaviyo’s unique growth story, from the diligence that came from bootstrapping to the 'unfair advantage' gained by scaling a startup outside a typical tech hub. 

Klaviyo is an intelligent marketing automation platform. Under Andrew’s leadership, Klaviyo has grown to support over 143,000 brands. The discussion covers how a startup changes during various stages of growth and offers an open look into the evolving nature of the CEO role—from the earliest stages to operating as a public company. 

Conversation highlights:

  • 00:00 – Background and Klaviyo’s founding story
  • 06:03 – Reflecting on the hustle it took to gain early customers
  • 10:27 – Challenges & benefits of bootstrapping, what Andrew would do differently today
  • 18:36 – Launching in Boston and learnings for startups building outside of a tech hub
  • 22:47 – Building enduring partnerships like Klaviyo x Shopify
  • 27:20 – How Klaviyo stays ahead by being a customer of their own products.
  • 32:54 – Benchmarks Andrew looked for during the pre-IPO journey
  • 35:53 – Klaviyo AI, and the impact Andrew sees from the technology in the future
  • 41:00 – Advice for the challenges a founder faces as their roles change and evolve

Featured: Andrew Bialecki, CEO and Co-Founder of Klaviyo and Ping Li, Partner at Accel

Learn more about Accel’s relationship with Klaviyo:

Read More

Ping Li (00:08):

I'm your host, Ping Le. I'm here with Klaviyo CEO and Founder Andrew Bialecki. Andrew, it's great to have you here. Thanks for coming.

Andrew Bialecki (00:17):

Yeah, thanks for having me here.

Ping Li (00:18):

Well, to kick it off, maybe give our listeners a little quick soundbite, brief overview of what Klaviyo does and what the platform is all about.

Andrew Bialecki (00:25):

We started Klaviyo about, gosh, 12 years ago, and we help, gosh, what is it? 140, 150,000 businesses, mostly consumer businesses, retail, e-commerce businesses. We help them build digital relationships with their customers and then deliver awesome digital experiences that they can monetize. So for a lot of our customers, we are 20%, 30%, 50% of how they drive growth and how they drive revenue, and we help them build direct relationships with customers versus having to work through marketplaces or ad platforms.

Ping Li (00:56):

Go back to 2012, you and your co-founder, Ed, were thinking about, what was it your second company, right? The first one didn't go so well. What was the story there?

Andrew Bialecki (01:07):

I started my career after college. I was not a CS grad, but I really wanted to do programming. I was a physics nerd and realized that the pace of that was too slow, so I was going to get into programming instead, and so I went and worked at a couple of startups. So Klaviyo was actually Ed and I’s first company, but along the way I had worked at two startups that had worked. One got sold to MasterCard for a couple hundred million dollars. The other got sold to HubSpot and became the core of HubSpot's marketing platform. So those were really great companies to learn from. Then I worked at one that didn't work out so well and it kind of never got product-market fit. I think we sold maybe a couple thousand dollars of software in a year. So it really never got off the runway.

(01:51): And then along the way I had my own little side hustles and had tried to spin things up and had those kind of flame out. You're talking about by the time we started Klaviyo, I feel like I've learned from the best. So one piece of advice I give founders is just go spend some time studying with other people that are great. I remember when the first startup that I worked at, or the second startup I worked at was in Boston and I'm from Boston and I moved back and I said, I don't even care what this startup does. I just want to work with people that are good founders, like people I can learn from that know how to run a company well. And so I was lucky enough to hook up with some serial entrepreneurs and they really showed me everything. I felt like I knew the engineering side, the product side, the design side, but I didn't know the sales and marketing side. So I learned a bunch of that stuff. Yeah. Anyway, so it was 2012, I mentioned that that third startup I joined was kind. I, at one point, the CEO really great guy came to me and said, Andrew, can I give you an IOU on your next paycheck? And I said, okay, this is probably not long for the world. So I worked there a few more months and then yeah, I told Ed, I said, Hey man, I really want to start something. Let's play around with some ideas.

Ping Li (03:03):

Did you ever consider working at a big company to learn some of the things you did or did you say, Hey, look, I need to do it at a startup now, Google, Facebook, were those too big.

Andrew Bialecki (03:12):

So it's interesting, when I graduated was the mid 2000s, and obviously Google was all the rage and so was Facebook. And I remember I applied to both, and I'd only taken a couple of CS classes, so I was like, I'm probably underqualified, but whatever. Literally never heard back from either of 'em. So I was just like, alright, I guess I'm not meant to work at Big Tech.

Ping Li (03:35):

So only the startups would hire you. Basically.

Andrew Bialecki (03:36):

Only the startups would hire me. And then once I got my taste for that, it's funny, the first company I worked at when I joined was maybe 40 people and it grew to be 200. And I was like, wow, this is a really big company. I don't think I could do this. So it's funny, I remember around when the second company I worked at Boston got acquired by HubSpot. I remember thinking it's like, oh, also it's kind of a big company. It's like a couple hundred people. I don't want to do that. So I kept going earlier and earlier stage. And so yeah, Klaviyo is actually the biggest company I've ever worked at. So every week its like we set a new record? 

Ping Li (04:11):

And so did the startups that you worked at, did they inspire you to start Klaviyo?

Andrew Bialecki (04:17):

If you think about what Klaviyo does at its core, we started out building this new type of database technology. I was just obsessed with this idea of how do I automate myself? How do I scale myself? And you think about that, it's like, well, the core of any human is the brain. It's like, how do I store data? How do I think? And then how do I put it to work? So interestingly, the first company I worked at, this company called Applied Predictive Technologies, still exists, great company just outside of Washington DC in Virginia. And they were doing exactly this. They were basically building data science, building these systems for big companies, Bank of America, Starbucks. So I learned a ton from that. And that was really early. It was actually, it was pre-Hadoop, so it was all just how big of a Dell server can we build and we'll solve SQL Server on it, and then how do we try to figure out how to do machine learning?

(05:07): So it's funny actually, with all the artificial intelligence stuff we were doing that way back in the mid 2000s. It's just the techniques were different. The tooling was a little different. It got me really into data. And then the second company I worked at was this company called Performable. They were big into marketing, Bartech in some sense. When we started Klaviyo, I was really interested in this idea of how do we build these modern data systems that are good at aggregating data and making sense of it. And then when our customers started saying, well, we want to apply that to marketing, I was like, okay, well, I feel like I've got a little bit of background there. We will build that too.

Ping Li (05:45):

Klaviyo obviously has a lot of ecommerce retailers, is that an area that you were like, Hey man, this is kind of cool. I like to shop online a lot.

Andrew Bialecki (05:54):

I remember when my aunt and uncle used to give me Amazon gift certificates in the nineties, but no, I wasn't some big e-commerce fanatic. It's funny. So when we built Klaviyo, the first thing I feel like anytime you start a company, you got to hustle for your first, if it especially a B2B company, you got to hustle for your first 10 customers. So I remember we used to make a list this was probably six, seven years ago where people would ask, where did the first 10 customers come from? And I sat down and we're like, oh, okay. Well, actually, our first customer came from direct messaging somebody on LinkedIn. He was part of a LinkedIn group, and then I met him at a coffee shop somewhere in Boston and just totally hustled into a deal. The second was at a startup dinner I went to. So it was all kind of random, but the initial idea was I was like, well, no, I was kind of building this technology for other startups.

(06:44): I was like, oh, you build an app, people sign up for it. Wouldn't it be great if you could automate getting them set up the onboarding experience? So we did that and what was interesting was they just had no idea. They didn't even know how their own users were using whatever they built. And so it was actually, it was a lot of consulting with them to figure out what are even the right workflows? What do you even want your customers to do? They hadn't really figured out product market fit. I had a friend of mine from high school who had an e-commerce business. He was selling quilts, online quilts of all things. And I remember saying, Hey, so I think this might apply to your business. You’ve got now tens of thousands of people that have bought quilts from you. 

Ping Li (07:32):

And it was all D2C?

Andrew Bialecki (07:33):

It was all D2C. It was moms that were buying quilts for their kids when they graduated high school or college, stuff like this. You could customize 'em little bit. And anyways, so he was like, yeah, sure. Here he was building on Shopify. I was like, yeah, look, they're very developer friendly. Can you just give me an API key and let's rig this thing up? I built a little connector into our database for Shopify data and almost immediately he was like, oh yeah, this makes sense. And then I had the customer of ours that I met through a little Boston dinner group was in the men's tailoring business, so think custom shirts and suits, and it was the same thing. They were like, oh yeah, we have these clients. They come in, we get them measured, we get their measurements, and then every time in your shirt, you just let us know the patterns we want and off we go. But they had the same thing, same use cases, and I was like, okay, wait a minute. There's super product market fit here. So yeah, we kind of fell into the whole resale and eCommerce thing.

Ping Li (08:42):

Wow, quilting. That was the genesis.

Andrew Bialecki (08:48):

Yeah, it was hot. There's a whole market.

Ping Li (08:49):

Is that company still around?

Andrew Bialecki (08:50):

Yeah, yeah, yeah. My friend still runs it. It's called Project Repat. The name doesn't remind you of quilting, but yeah, it's a great business. It was funny, he had started that business maybe six months or a year before we started Klaviyo. And funny story you'll find you really need to get close to your early customers because in the early days when we were doing demos of our software to folks that signed up, you always need an account that has a bunch of stuff in it so you can actually spend a lot of time setting those up and it never has all the use cases. So what we used to do is I had asked my friend, I'd say, Hey, can we just use your account? And so that actually worked really well for a long time and so much so that we hired sales reps and then we had three, we had 10, and they were even using this guy's actual business.

(09:39): So we had rules that if you went in there, if you created something, you had to delete it or clean up back yourself kind of thing. You can stay in this house, but make sure you clean up. Well, of course people would forget to do that. So that lasted probably for four or five years. And then one day he got fed up and so he went into his own Klaviyo accounts and created a bunch of campaigns that said “You should never use Klaviyo,” all this kind of stuff. And you can imagine the sales folks that logged in that morning were horrified doing a demo and we're like, what do I do here? So that is one of the all time most brilliant moves. Anyways, so we tell that story all the time. After that, we built our own little demo environment. He's true, but he was a very good sport.

Ping Li (10:25): And you're still friends.

Andrew Bialecki (10:26): We're still friends.

Ping Li (10:27): That's the most important thing. So look, there's a couple unique things about the Klaviyo beginning that I think I want to touch on. One is you didn't raise a lot of money. Klaviyo is mostly bootstrapped you a little bit of sseed capital, but then you became very profitable very quickly. Was that intentional? Was that it just happened to work out that way? You don't like VCs? I mean, let's be honest here. How did that happen, right? Because this is 2012, there's lots of venture funding around, right?

Andrew Bialecki (10:57):

Yeah. I think for us it was, so it was a bit of both. So one I had in my head that there were a lot of startups then that I felt like they'd raise what felt like a lot of money at the time, like a million dollars, they build something, but it was never really a profitable sustainable business. They get acquired and that was like, success. And we wanted my family, my grandparents own a small business and it is profitable. It doesn't employ that many people employs like 20, 25 people, but it's real and it's durable. And I was like, okay, well that's a real business. We should try to do that. And I also felt that if advice I give anybody that has an engineering background is that's really the hard part of you're building software. The costs aren't all that high and maybe it's a little different if there's some compute costs with some of the ML stuff that's going on now, but by and large, the costs aren't that high.

(11:51): So if you can put in the sweat equity, it actually gives you a lot more control over what you want to do. I talked to other founders in the Boston ecosystem and across the country, and you just hear these stories where it felt like they kind of feel like they got on a treadmill that they didn't know they were getting on when they raised investor capital because all of a sudden it was, okay, well we have to hit these growth targets and we also like to do that. We're going to burn cash and now all of a sudden there's this ticking clock and maybe it's good, it forces you forward. But I feel like it creates some unnatural behaviors for us. When we started, the idea was, I told Ed, I said Ed was finishing up business school at the time, so I said, Hey Ed, you're in the last six months of business school.

(12:39): And I had this vision in my head of business school, people don't do anything. I was like, you don't have classes or anything, so here just come work with me every day and let's see if we work hard for five or six months, can we get Klaviyo off the ground? And Ed had wanted to do a startup too, and our feeling was like, look, if we end up fighting all the time, that's fine. We will quit and we'll go to something else and if we can't figure out product market fit, then we'll go to something else. So we kind of felt like it wasn't so much, we were lucky to have saved up enough cash, enough savings from previous jobs that we were like, yeah, we can go a year and be okay. So we felt like let's just give it a go. And then once you get some customers, we were kind of like, alright, can we get what they used to call Ramen profitable? Can you get a couple thousand bucks a month where it's like the thing is, self sustaining. And so it took us two years to get to, I think it was a quarter of a million dollars in ARR run rate. And at that point we felt like, okay, this is a real thing. And then it felt in some sense everything was flipped where now it felt like maybe we put a lot of pressure on ourselves to grow, but now we could do it on our terms. 

Ping Li (13:50):

You were profitable at 250k. 

Andrew Bialecki (13:52):

Yeah. I mean we weren't taking salaries, at some point we took, but at that point we were like, Hey, we've actually got enough cash left over. We start hiring people. I mean for a long time, girlfriend now wife used to sleep with my phone by my bed and it would just be constantly buzzing because the servers would be breaking or something. And she was like, wow, you're going to have another engineer that at least you guys can rotate nights. So anyways, so actually the reason we actually, we ended up raising capital at all, it got to the point where you have a bank account for your company and for us it would go up little by little each day of the month and then you'd pay your server bill and eventually we'd run payroll and it would drop down to almost zero near, then it would tick back up. And so the reason we actually raised a seed round our first round was if I just got tired of looking that saw tooth pattern, we literally have engineers I wanted to hire, I'd be like, alright, so I'm going to make you an offer, but here's the thing, I can't hire you for four weeks because I need a few more customers to come in the door and I come down. We're like, well, that's kind of silly. We should raise enough cushion that we can. 

Ping Li (15:03):

Working capital.

Andrew Bialecki (15:04):

Yeah. Well actually another silly thing, we were bad at raising money. So when we raised that, I remember we closed it and I remember telling the partner, we were working with John saying, alright man, you just wire the cash when you can. And he's like, oh, actually, is it okay if we wire it in three or four months? And normally it's like, Hey, no, no, we need it right now. Now would be a good time to send it for us. We're like, yeah, that's fine. I think we closed it in February. We're like, yeah, sure. I mean July's fine just whenever you get a chance. So yeah, it really was. It worked out for us.

Ping Li (15:39):

And looking back, are there things that you would've done differently if you had more capital? Or do you think it kind of would've changed the trajectory.

Andrew Bialecki (15:48):

Yeah, I think I'd say two things. One, in retrospect, I loved building and coding and I’d just gotten in such a groove with Ed and I probably would've hired more engineers sooner because once we got up to five or 10 folks on the team, I mean man, we really started cooking. But that was a shift for me personally. I'd wake up every morning, I would answer all the support tickets and I'd go code away for the evening, the afternoon into the night. And that was just my routine. And it was actually different once we had employees like, gosh, okay, what products going to work? All of a sudden the job different. But I definitely would've done that faster. On the flip side, I think one of the great things that bootstrapping forced is just, it's like a lot of constraint. There are a lot of wonderful practices that got instilled into Ed, myself and our team early on, things I just mentioned.

(16:50): Literally our workflow was at night we set up a little shared inbox for support. And so we had a couple hundred customers. And so every morning in the first job we had was to answer support tickets until there were no, none left. Well, you really focus on figuring, fixing whatever the problems are. So it forces this great connection with customers. I get the same question all the time, we've got to fix that because I can't do the bigger things. So that was great. So we carried that forward actually for a long time. We'd have everybody in product and engineering, you do a support rotation. That was just part of the job. And so that also bred this, you really knew the product

Ping Li (17:29): And what people were doing with it.

Andrew Bialecki (17:31):

And I think it also, we developed some norms. We have these old pictures of, so we worked with a lot of retail eCommerce. Obviously the Thanksgiving, Black Friday, Cyber Money holiday mattered. And so after two or three years it was like, oh man. Yeah, a lot of people show up, got to scale systems. When we were like 10 engineers, I literally just used to say, okay, well we've got 10 big systems, everybody owns one. So the norm in engineering is like, oh, I have a teamwork on this. There's redundancy and there really wasn't. I mean, if somebody's out sick, okay, well somebody else will cover. But it created this awesome sense of ownership that it's, yeah, I own scaling this thing, I need to think it through. Everybody carried a pager. I mean it was on your phone. And that's carried through to our culture today is just this real sense of small teams can build big systems. They own their own stuff. So anyways, in that constraint, I think you learn a lot, you figure out some good patterns that you can carry forward.

Ping Li (18:36):

Klaviyo was Boston based. Do you think that was one of the reasons why you were more bootstrap than maybe some of the valley companies that have more access to capital?

Andrew Bialecki (18:45):

Yeah, maybe I'll tell you the other, the very pie in the sky version of like, yeah, we bootstraped because we had small business in our family and we're going to be one of those too. The other funny thing was, so Ed went to business school at MIT and they have this great entrepreneurship center. And so I used to take the bus every morning and I would just hang out freeload in this entrepreneurship center and Ed would come over after his classes were done. And so I remember at the end of that, Ed wrapped up and since May and they're like, Hey, so for everybody that's starting a company, we're going to do these little no equity needed $20,000 grants for the summer. So we'll give you space and we'll give you some capital. And $20,000 to us was like, oh my gosh, we got paid for the rest of the year with that kind of cash!

(19:28): Anyways, so I remember we had the best attendance at this little entrepreneurship center and we had actual customers, people paying us money, and I remember we applied, we applied, and they're like, oh yeah, sorry, we don't have a slot for you. So I remember thinking on the venture side, it was like, man, how are we going to raise from anybody else? We should have such unfair advantage! We can't get this 20K check. How could it not be for us? So anyways, we didn't have a lot of confidence that would work out anyways. So to Boston.

Ping Li (20:04):

There weren't that many startups getting built in Boston at the time. I mean there always is compared to the Valley.

Andrew Bialecki (20:08):

Well, yeah, I didn't have the perspective of, so a bunch of my college roommates had gone to do tech in the Valley and were deep into it. Yeah, I mean one of the things, my personality was very heads down building, so I actually felt fortunate that I kind of built a network by working at these other startups beforehand. Otherwise, I wasn't the kind of person that was going to go to a bunch of meetups and hang out so myself, I was like, well, it's actually okay. You could read enough online. This was starting to be the era of lean startup. There's enough good stuff out there you could learn through reading other people's materials. But yeah, there wasn't a huge community. So I think that partially led to the bootstrapping. Yeah, I mean one of the things we're working on now within Boston is I think that differs from the Bay Area is if the community is just not as tight knit, it's more people doing things on their own. And we're working trying to even use Klaviyo like a catalyst this to just get more of those entrepreneurs together so they can share some of these stories of Hey, what's working, what's not, get some mentorship. But yeah, so I think because there are relatively fewer tech companies.

Ping Li (21:22):

You ever think I was starting elsewhere? Saying, hey, I'll go to New York because New York was getting hot at the time, right? That's not too far. 

Andrew Bialecki (21:27):

Yeah, I mean, so my calculus was, I mean, not really. So one, I was from the Boston area and I knew it's like, okay, we're going to do something East Coast. And I had a lot of friends in New York, but I didn't want to live in Manhattan. And I guess said, I guess I wasn't thinking broadly about Brooklyn, things like that. So I remember when I moved, so I moved after college down to DC and when I said, okay, I'm going to join a startup, it was either in New York or Boston. And at the time I was like, okay, no, it's definitely Boston. But yeah, the other thing too that we wanted to do with Boston that we're really working on is, I mean there's so many great colleges and universities there, and it really bummed me out that I had this group of friends and a bunch went to finance to go to New York, find whatever, do finance thing, and then the other ones went to tech and they're like, okay, well we have to go.

(22:10): And I'm like, well, what's up with this? And literally there weren't enough companies. So in the early days we felt like we had an unfair advantage recruiting because this was the early days of AngelList. And I felt like I could email anybody that was like, I'm into startups on AngelList and I could get them to have coffee and talk about joining forces. So yeah, so that whole idea of - part of the reason for setting up shop in Boston other than the local roots was there's so many good college universities, we just literally want to feed off of be a place that people can go from there.

Ping Li (22:47): And I guess you mentioned earlier, another thing in the early days was you kind of fell into Shopify through your buddy's quilt shop. Right that relationship's obviously been really important, they're shareholder investor over time. How did that come about? Were you kind of like this is it? Or did it kind of evolve over time?

Andrew Bialecki (23:14):

It's interesting, so definitely, yeah, I feel really lucky. I guess it is probably destined because you talk to people, it's like, oh, hey, I'm building an e-commerce business. You're like, okay, cool. What platform are you doing it on? People building on Shopify. And I also think that one of the things I love is they were so developer first. They'd really gone out of their way to say, look, there's core things we're going to build into our platform, but everything else, we're looking for an ecosystem to build. And there's not a lot of companies that have really executed that, but that's a really play. So yeah, it's interesting. In the early days when I was just doing a lot of coding building products, ed spent a lot of time with the Shopify team. And another thing that they did a great job of, which I've now seen other platforms really try to emulate is they put somebody on the developer ecosystem.

(24:10): That sounds obvious, but so many, especially B2B companies don't do that. So they'll build a bunch of APIs and then if you're a developer and you're reaching out, there's just nobody to talk to. So I remember literally there were other e-commerce point of sale platforms we built integrations into because once you do with Shopify, you're like, oh, it's actually kind of good. We do the other ones. And literally with Shopify, if we had a question, we could reach out to somebody and they get back in 12 hours. There were other folks where if you said, Hey, I built this kind of integration into your platform, I mean it could take weeks for somebody to get back to us. I'm like, well, built this thing, don't you? Isn't it valuable? And it just took a long time for you to turn it. So they were really receptive.

(24:55): So in the early days, Ed spent a bunch of time with their ecosystem team, and then probably four or five years in, I got a chance to meet at the time a guy, Lauren, that was running the Shopify Plus the larger, because that's where we had more concentration. And then Toby and Harley, and I think it's one of those things when you partner, the way we've always thought about partnerships was you can't start with the commercials, you have to start with the product. There has to be some pull, there has to be some value. Like, hey, this is in the case of Shopify, the mutual, it was mutually beneficial because like, hey, look, we're going to help businesses on Shopify organize their data and then message customers and that’s going to drive sales. And by the way, that's part of how Shopify monetizes their platform, part of what makes their customers successful.

(25:47): So we look for that where there's this natural, both people aligned to this, we have metaphor of pulling on a rope. You need both people to be pulling, and so you get that product market fit or that partner-to-partner fit, and then after that, then you can kind of figure everything else out. So I think what was awesome was by the time we got going, there was such this product to product fit customers were raving about it. And then personally, I think for both Ed and I, it's fun working with companies that like building products and in this case shop, I was obviously a little bit ahead of us, and so people you can learn from anyway, so we've always sought out partnerships like that.

Ping Li (26:28): Yeah. So fast forward a decade, Klaviyo was on this tear. How have you been able to maintain that customer - how many customers do you have now?

Andrew Bialecki (26:41): Yeah, I think we're like 140, 150,000. Yeah, it's a lot.

Ping Li (26:43) Are you still answering support tickets? 

Andrew Bialecki (26:46): I still do like to sample into support tickets. Yeah. I have a little script that sends me every morning just kind of a random sampling of 10 support conversations, and we'll do something similar in our CRM of just looking through 10 random accounts. The thing that we talk about as a product team is there's this great quote in programming that's like, Hey, all software bugs are shallow. They're easy to find if you put enough eyes on them. I think the same is true. Same for software. For a lot of products, I think what you find is the companies that build the best products, this sounds ridiculously simple, but they use their own products in consumer. That's relatively easy. If you're a social network, you can just be on social network, consume from it, and off you go. I think when you're building software for businesses, a natural thing that happens is people don't end up using it.

(27:40): They're like, well, I'm not the customer. So something that we've talked about is we're not going to build the best products unless we are users of our own software. We're in it all day. And there's really two patterns there, either one is you can dog food your own stuff. So you can literally use your own software. So for instance, I mean it shouldn't be surprising. We use Klaviyo to message our customers. We actually use Klaviyo to message employees internally. So all of the same segmentation, analytics, all this kind of stuff, we're using it ourselves and we'll build features to solve for ourselves, things that we know that other people are going to want. So that's one trick we say is you can't just talk to customers, you have to be a customer. And then the other part is you just have to get close with folks.

(28:28): Our engineering team, for a long time, I think right up until Covid still was in the habit of, yeah, a couple days a quarter would spend doing support. And it's a bit of a logistics thing. If you start answering a question, somebody writes back three days later who finishes it up kind thing. But you can figure out some processes around that and you do have to be trained up enough to really know the product. So that's our support team. We call 'em projects. They really do know everything anyways, but we built those norms of like you talk to customers, you get in their shoes, you role play through that. And now actually what we're doing is we're working with bigger customers is we'll just send little teams, small teams of engineers and a PM or a designer and we'll just send 'em onsite. So we've done that where customers asked us maybe to prototype or do some ideation with them. They'll say, great, you five folks, get on a plane, go fly to whatever city that customer's in, set up, shop in a hotel, an Airbnb or whatever, and come back in a week with the prototype and then we'll decide if we're going to move forward with that. 

Ping Li (29:29):

You expect your product engineering team to be in with the customers as much as the customer facing parts of the organization.

Andrew Bialecki (29:35):

Well, another thing that I tell everybody when you start on our product engineering team is that I assume look a little bit my background. I joined these other companies to learn the crafts with some intention of yeah, maybe someday I'll start a company or I'll go be a leader at another company, or maybe I'll grow into that company. So I tell everybody, I'm like, look, I'm going to treat you as if you want to be a future founder. So you've got to own the whole thing. So it's not just your product, you're accountable to the numbers, but also to customers. You need to know those. I think the most insidious thing that happens as companies get larger is this viral loop of spending time with customers, building products, giving it to customers, getting feedback. That loop just gets longer. So in some sense, the perfect thing is what I described in the early days of you answer sport tickets in the morning, you code in the afternoon is perfect because nobody can hide rough spots or bugs from you. I think that's obviously harder now. So one thing we pay attention to is our support team is awesome, but sometimes they'll hide issues from our engineering team because they're like, yeah, we know workarounds for that, so don't worry about it. But really it's the kind of thing that you want engineers to see like, no, no, no, that's not acceptable. I'm going to fix that. We try to kick that loop as tight as possible.

Ping Li (30:54):

Did you ever say, hey, look, things are going great. Let's exit and sell the company and make a lot of money, whatever, right? I'm sure people were interested. You guys were a very good company. Did that ever come up?

Andrew Bialecki (31:08):

No, it wasn't really. I guess probably part of my physics math background is I tend to just thinking larger and larger abstractions. So every time Klaviyo got a little bit bigger, I'd be like, okay, cool, but what would it look like at 10 x or 100 x That scale, right? So it was always the goal, the goalpost get even further out as long as we felt like Klaviyo was, we had a mission and a product roadmap that was exciting, but there was a lot of room to grow. If it ever got slow, then like, oh yeah, I'm out of here. But we felt like, I was like, okay, well, and I felt we have some agency in this, we can decide whether it goes slow or fast, how hard it is. So no, we never really wanted to sell the company. I mean, it's funny, obviously over the years you talk to a couple of folks, and my favorite way to kind of deflect that was often you do this with founders and other companies. And so my favorite line was I was like, oh yeah, it'd be cool. It could be acquired by us. I'd like, oh, it's great. So how come you never did that? They're like, well, I would never do that. I have so much left to build. I said, ah, okay, well then you'll probably understand when I tell you that I'm just, we think the same. I think there's actually a lot of respect for that. There are a lot of great companies out there that we've gotten to know, but no, we always felt like we've got a lot that we wanted to build, and so best to just keep pushing.

Ping Li (32:34): And so what inspired the IPO Last fall? There were not many IPOs going on. It was frankly the first software IPO in I think 12, 18 months or something crazy like that. Yeah, you’re wildly profitable. I mean, it's all public info. What was the driver to do it then and there?

Andrew Bialecki (32:54):

It's interesting when we got to, so I think about the phases of Klaviyo. The first two years were all about existential dread. Are we going to die? Boy, is this thing going to make it? And then once you got there and you're profitable, at the time there was this thing around bootstrap companies of like, ah, is it just a lifestyle business? And I was like, no, it's not. We can have the highest growth rates with the best funded companies. We can have both. I think once in my head, I was always like, once you get past a hundred million in revenue, you could be a public company. You're at that scale. So it was probably when we got to around that 50 or a hundred million dollars point, and we realized that was just the trajectory. And I kept getting all this advice from folks, mentors of mine that were like, I've seen this story before.

(33:42): You've just got enough velocity, it's going to get there. And yeah, a little bit for the local ecosystem, but I have this little chart that I show folks at our all hands sometimes, which is like, okay, where are all of the a hundred billion tech companies located in the United States of the ones that are headquartered founder in the us, where are they? And the reality is they're all in the Bay Area in Seattle, and there was one company that was on the East Coast, I tell you what is it? It's IBM. They're like, oh my gosh, wait, it's been so long. We need some new blood here. So whether it was Boston or otherwise, one of the things we wanted to do is just prove that it's like, yeah, hey look, you can build really big companies elsewhere. And so we felt like part of the way to doing that was going public.

(34:34): And so honestly, the timing last year, it got to the point where everybody, I kind of ask people what's it like being public? And most folks were telling me they're like, yeah, I mean there's a little bit of overhead, but it's not really all that different. I'd say actually the fun part for me is, I played a lot of sports growing up. So I actually think that the earnings part is actually kind of fun because it's a little competitive. It's a little like, okay, there's a scoreboard and you got a report and all this kind of stuff. 

Ping Li (35:02): You’re the only person I've ever heard say, call earnings fun.

Andrew Bialecki (35:05):

I think you just have to, there's probably an element to we just tend to be so long-term focused. You don't try not to overemphasize month-to-month, quarter-quarter things. But anyway, so for us it was like, okay, well, we have this saying internally that when you get confronted with a problem, you know you're going to have to solve at some point you should just eat your vegetables. We talk about that a lot internally. So man, here's this system. Oh boy, we're going to have to re-engineer it to scale. We could do it this year. We do it next year. It's like, okay, well, you might as well just eat your, let's just get it done now. Right? It's going to be cheaper now. So I think similar for being public company ready, a lot of it's like the financial reporting, this kind of stuff. So yeah, I mean, I felt like it's like, Hey, let's get that done. Let's eat our vegetables on this, and then we put that in the rear view mirror and keep going.

Ping Li (35:53): So I wouldn't be doing my job as a venture capitalist if I don't talk about AI. So I guess you announced a bunch of stuff with Klaviyo AI. I mean, AI is in everything. So what does it mean to you? What does it mean to Klaviyo and what do you think it actually means to your category, right? The true impact? 

Andrew Bialecki (36:13):

Go back to the first company I worked at where we were doing AI, but we called it data mining. So it just goes through these different names. By the way, AI is way sexier than data mining, much better branding. Anyways, it was actually from there. I always found it fascinating, this idea of computers being able to outthink humans. So there's the old Steve Jobs analogy that, hey, a computer is like a bicycle for the brain. And something we talk about is if you look at a lot of software, it's mostly productivity. Human knows what it wants to do, but the software just helps you get it done faster. There's actually very few software applications that are honestly like the human says, yeah, this is the outcome I want. Now you just go figure out how to do it. That's something Ed, my co-founder talks about a lot is we want to build software that's outcome-oriented, not just productivity.

(37:08): So for us, we've been working on this, we've a little bit to avoid the hype, we created a team internally probably four or five years ago called our data science team, and we deliberately did not call it machine learning and AI, I now refer to that just to, but to declutter it, but we literally called it data science. I was like, we don't want, this is not hype, right? This is, we're going to make this real. It's going to be substantive. So for us, our belief was let's just take what we do with marketing that in the future, it won't be that you buy marketing software that then a human has to click buttons and program and set up. It should just be, it's like, oh, the marketer says, Hey, look, my brand is kind of like this and these are my goals and these are some guardrails, and then here you software go figure out the rest of it.

(37:54): I think that's very possible. And I've been excited about the idea of software applications that you basically just kind of give a high level goal and it figures out all the rest of the details. And obviously in the first couple of years, it'll be the human more in the loop of, Hey, I'll double check all your work, but we have this analogy that if you take the Maps app on your phone, eventually the directions got good enough that you sort of don't question 'em anymore. You just blindly follow. Oh yeah, okay. It probably figured out that there is an accident there and I should go around it. And so I think the same thing happens here. Humans want to double check and all this kind of stuff, but eventually it becomes more autonomous. So that's actually our whole point of view is could you build something that, let's take the marketing domain.

(38:41): Obviously, there's a reward function, there's an outcome variable. So in marketing it might be conversions or revenue or something like that. And so we say, great. So if you have that as kind of the dependent variable, the reward function, can you train up a model that will just figure out what kind of marketing to do and to whom and to just maximize that. So we have a team that's literally working on that. And part of what we launched is our toolkit around a variety of features that will start to do that for you. And we've got a lot more coming there so that we think about putting the marketer or the user in the kind of director's chair. They're directing things, looking at things, but they're not actually, they don't have to be the one pointing and tilting the camera. So as it comes to AI, I think that's fascinating, this idea of applications that will just run themselves, you can give it higher order.

Ping Li (39:24): Does that mean there's no role for the human in loop anymore?

Andrew Bialecki (39:27):

I mean practically we don't think the humans go away or I don't think they do. I mean, let's just take marketing. I mean, there's still so much somebody starts has to decide. So what does my brand stand for? What products am I going to offer? What are my objectives? I do think a lot of all technology, the nice part is if you can eliminate a lot of the rote work that frees things up for the, actually, actually we have these little brainstorm sessions internally of I was thinking with new technology or new products, what are the second, third order effects? So there's this big question we'll have these discussions are like, so what does happen? Where would you put your time? And I'll talk with customers part of this, where would you put your time if you didn't have to do the following tasks? And it's fun because they'll kind of go away for a weekend or whatever and come back and then we'll have a little discussion about it. The interesting part is you'd think some of 'em would be nervous, like, oh man, I don't know what I do. The reality is actually they have tons of ideas of marketing strategies, ways, experience. They want to go with their, they're like, I just never had time to do that, and now this frees me up to do it. So it's exciting.

Ping Li (40:31):

What is something that kept you up at night from the very beginning that you looking back at, you're like was a wasted cycles thinking energy, mental cycles, thinking about that. You're like, it just didn't matter, right? But you were obsessed when you were 15, 20, 30 people and you're worried about, besides running out of money, right? That's an obvious one. Were there things that you kind of now looking back, you're like, huh, you’d tell other founders, don't worry about it. That's not a big deal.

Andrew Bialecki (41:00):

I think the biggest thing for me was getting used to the rate of change and how often the way you spend time needs to change. I feel like I used to spend a lot of time justifying to myself and then sometimes to others that it's like, no, no, no. The way that we're working, I'll give you an example of when it was just the two of us and I was doing all the program engineering and then the difference of, okay, now there's a small team, and gosh, the job changes. And then when we were say 50 or a hundred people and the job changes again, I feel like I spent a lot of time justifying to myself why I should keep working the way I am. And I've heard other folks make this analogy of like you kind of got to break out of that box and see the bigger thing. And I always felt like there was a big unlock when I was like finally gave in and was like, you know what? The way that I'm working now, the way I'm spending time, it may have made me successful, but it's not going to anymore. And that was always really hard. It might mean the skills may be totally different or you're very comfortable. And those were always, I wasted the most time doing that of where I get to, I really like doing it. And they're like, yeah, I know, but that's not your job. So I think anyways, just stepping back to think about that and just reassessing, I try to reassess every three months. Okay, is the way that I'm spending my time, one, is it exciting? Am I motivated by it? And then two, is it effective? And if not, why not there something I'm afraid of or some skill I've got to adapt. And I think if you just kind of step back, that helps a lot.

Ping Li (42:31): You would never start a company by yourself. You thought having a co-founder was a big reason for Klaviyo's success.

Andrew Bialecki (42:40):

Okay, one, should you have a co-founder? My advice is definitely yes. And then two is how to find one. So it's interesting, before we started Klaviyo, I tried really hard. I was like, okay, I need some unfair advantage. And so for me it was like I was just going to be, I can build anything and I've got some UX design sensibilities. I know what good products are. I've been studying and I know how to build them. I felt relatively less sure about myself when it came to the sales and marketing side. So that was something I just would study like crazy. That's part of what I tried to learn from these other companies. I was like, okay, but I need at least one unfair skill. I think with co-founders, you want somebody that one compliments that. So Ed and I had the kind of normal division of labor in the early days of I do the building, I would do the selling. But then also, I mean, you're invariably going to run into these decisions where you're like, well, you don't know what the right answer is. I remember there was a point, it was probably end of our second year, and I mentioned we were doing about a quarter of a million dollars in revenue. All of that revenue, 20% came from one customer. I mean, it is not a huge, somebody paying us 5,000 bucks a month, but for us it was a lot.

(43:53): And it was getting towards the end of the year. It was October, November, and we've been working hard, but they said, Hey guys, we're leaving. And so I was like, oh my gosh. I was like 20% of our revenue at once. I said, there's a little bit of like, man, you got to take a step back. What are we doing? How do we recover, right? I came from a companies before where it's like we say zero churn is the right amount of churn. And I was like 20% like crap, maybe we don't have it. But I think that's one where you just need someone to bounce things off of. And I remember Ed telling me, he's like, no, no, no. He's like, we've got this. And funny, actually, that customer ended up coming back a year later. So that worked out. But I think that the mental side is, I think a lot of us, you can work really hard, but it's the mental side and you need somebody that's like, you feel like you can be pretty vulnerable with that matters.

(44:48): And then in terms of how to find them, I mean, I lucked into it, ed. I worked at that first company down in Virginia, and it was great because you saw a lot of the traits that I think we had similar notions around frugality, hustle, drive. Ed was the guy that helped stand up for that smaller company members. It was only 40 people helped stand up their London office. He was that kind of take risks kind of person. And I had the chance to work with him on some projects while we were there. So I think finding somebody that, and then he and I would kind of play with some ideas over beers or whatever at MIT and something you can kind of argue with. And Ed and I are both very big debaters, so we feel like you could argue with somebody for two hours and you still liked them, was really important. And

Ping Li (45:37): Having seen you two in action. You definitely argue a lot, but you guys always move forward in the right way. So it's clearly been a big part of your success. Well, Andrew, look, I want to thank you for joining us here on Spotlight On, but more importantly, I want to thank you for letting Accel and me be part of the Klaviyo journey. I know there's still a lot more to go, but it's been really fun. So thank you.

Andrew Bialecki (46:00): Well thanks so far, so good. Lots to do.

Meet your host

Partner
Focus
Cloud/SaaS, Enterprise IT, Security

Ping Li

Ping Li joined Accel in 2004 and focuses on investing in business software applications and cloud native technology platforms. Learn more about his expertise.
Read more on Accel.com
Season 1 of the Spotlight On podcast, by Accel
Cloud/SaaS

Cloud/SaaS

Commerce

Commerce

Artificial Intelligence

Artificial Intelligence