Zapier’s Secrets to Product Market Fit with Wade Foster (Co-founder and CEO of Zapier)

When a lot of folks think about like product market fit, they think of it as like an ordering thing. Oh, first we need to make the product and then, oh, the product's great. Now we gotta go figure out like how to go market this thing. Now they're trying to like bolt it on. I actually think product market fit is actually one thing. It's not a sequence of things. You want your product and your go-to-market to be sort of one fluid motion. And the best PLG companies, like, this is how it works. It's like sort of intertwined. I have founders come up to where they're struggling. It's like, I've got a great product. I gotta go figure out how to get distribution for this thing. And I'm kinda like, well, Are you sure? Maybe it's a well-engineered product or a well-designed product. That absolutely could be true. But is it a product that anyone cares about? Because if no one cares about it and you don't have distribution engine that shows that people care about it, think you might want to go back to the drawing board.

Welcome to The Peel, where we explore the world's greatest startup stories. I'm your host, Turner Novak, founder of Banana Capital, the venture capital firm with the funniest podcast. My guest today is Wade Foster. Wade Foster is the co-founder and CEO of Zapier, an automation platform that helps you work faster. Wade and his co-founders Mike and Brian started the company in 2011 and have always done things a little bit differently. They're a tech startup based in Missouri, have run a completely remote team since 2011, and raised a modest $1.3 million seed round after Y Combinator in 2012. Zapier hasn't raised outside capital since then, despite crossing 9 figures in annual revenue a few years ago. They did do a secondary sale to Sequoia in 2021 at a $5 billion valuation. Wade shares his contrarian framework on product market fit, how he deals with imposter syndrome, how he delegates without losing track of the details, shutting down Zapier for a week-long AI hackathon in early 2023, and all the cool AI and large language model powered products Zapier's now building that came out of the hackathon. I love this conversation and I think you will too. Quick shout out to Vlad Magdalin at Webflow for introducing me to Wade. Now let's Let's jump in after a quick word from Secureframe. To automate your work, there's Zapier. For automating compliance, there's Secureframe, giving you more time to focus on growing your business. If you're building a new product, the last thing you're probably thinking about is compliance, security, and privacy, and that's where Secureframe comes in. Longtime listeners are very familiar with Secureframe, the automated compliance platform built by security experts. Thousands of your favorite startups like Angel's Coda and Ramp are already using Secureframe. To get, stay, and automate their compliance with security and privacy frameworks like SOC 2, ISO 27001, HIPAA, GDPR, and more. I'm an investor in Secureframe, and I recommend it to every founder I meet. Join hundreds of other listeners and click the link in the show notes to get started on a demo with Secureframe's in-house team of compliance experts and former auditors. Thank you, Secureframe. And now let's talk to Wade Foster. Wade, how's it going?

It's going good. How are you, Turner?

I'm excited. Excited to talk about Zapier. Can you just really quick explain it to us? What it is, what it does?

Yeah. Zapier is an automation platform. So we have connectors into about 6,000 different apps that you might use at work. So think things like Salesforce and Slack and Gmail and Airtable and MailChimp and you name it, all, all the sort of stuff you would expect, plus a whole bunch of tiny things that you may not know. And, you know, it helps regular people just hook things up. Automate their work, build things that, uh, they maybe couldn't have built without, you know, engineering horsepower in the past.

Can you maybe give us an example of maybe one or two things for someone who still doesn't quite get it?

You know, it goes from, you know, on the simple side, you might have, say, hey, every time I get an email that's got attachments in it, I want to make sure those get like saved to Dropbox or something like that. So you could set a Zap up that does that. To things that are more complicated, say you're running like advertising on Facebook and you have lead forms where you're collecting leads. You know, those, you could set it up so that those leads run through like a lead score where you can say if they have these characteristics, I wanna send 'em, you know, over to my mailing list. But if they have these characteristics, I wanna add 'em to Salesforce and I wanna send a text message to like my top salesperson right away to contact this person. And so you can, you know, it can start from anything that's simple to like, you know, complex, uh, automations that, uh, span like an end-to-end workflow that you might have inside of a business.

So it kind of sounds like anyone who uses a computer to get work done. And use it.

Yeah, it's, it really is targeted to like basically people who, you know, sit at a laptop to do work.

I guess going back to when you kind of started the company, when you came up with the idea for the product, could you do this? Like, did this exist already?

Not really. Uh, you know, at the time, like, the, the idea was more simple than it is today. We'd noticed that, uh, if you would go to the, you know, SaaS products of the day, so this would have been things like Salesforce or Zendesk or Um, yeah, Twilio and Stripe had just launched, but there was also tools like QuickBooks, and it's kind of like that Gen 1 SaaS. And at the time, user forums and public-facing communities were like really popular. So if you remember like Get Satisfaction and UserVoice and things like that, they were super trendy, and it let customers basically in public tell you everything about the product that they liked, disliked, add feature requests, all this sort of stuff. And, uh, you would go to these and inevitably, like, there was always a very popular category, category of post, which was integrations. You know, someone would say, hey, when are you going to build an integration with, I don't know, whatever, PayPal, Salesforce, whatever, right? And, uh, you know, then you would pop in and you go look and see what were customers saying in that thread. And inevitably it'd just be like classic internet plus one, plus one, me too, me too, I can't believe you haven't done this yet, like, you know, give it to me, give it to me, you know, get The thread would get 10, 15 comments deep and lots of upvotes. And then a product manager would come in and the product manager, you know, very politely, as product managers tend to be, would say something like, hey everybody, thank you so much for the feedback. We really value it. We're going to take a look at this and we'll let you know sort of what we decide. And if you've ever worked in a software company, you know that this is code for, you Yeah, probably not gonna happen. And so we looked at that and we're like, all of these forums have this same pattern, this same category. So for whatever reason, either these companies don't want to build these things, they can't build these things, it's not profitable to do these things, integrations are just not getting built. And so Brian, my co-founder, messaged me on iChat and was like, hey, I think we could build something like this. And, um, to me, I got kind of excited because I'm not a very good developer myself. And in my day job, I was working with the Marketo API, which at the time was this old school like WSDL SOAP API. And I was just like beating my head against the wall. Like just, this is no fun. I don't like doing this. I'm having a tough time figuring it out. And so when Brian pitched it, I was like, oh, I would use this right away to sort of take leads out of our CRM and send them into Marketo. So, you know, we, we, uh, that, that's where we started. We built like an initial prototype that was a single trigger, a single action. And so you could say, hey, when this event occurs over here, I want to make sure that, uh, the data from here goes into this other thing. And it was just like a simple, like, point-and-click, drag-and-drop thing that allows you to, you know, any sort of regular Joe to set these things up.

But it wasn't that simple when you first started it, right? Like, it's still kind of janky.

Well, yeah, so the, uh, the first version was bad for sure. I remember the very first customer we got, um, we had sent them— sent him an invite over email, uh, and he was, uh, you know, he's trying to set Zap up and he couldn't do it. And, you know, he shot an email back. I was like, Wade, really? Thanks for— thanks for inviting me. I love what you all have done. You know, being very polite, you know, he's like, I could use a little bit of help. And so we jumped on Skype. It was Skype at the time. And I sat down and just said like, okay, well, like, let me help you set it up. And so we went through setting up the Zap step by step. And as I'm like watching him use the product, I'm just sort of like trying to hide the horror in my face. Because he's struggling and he's struggling not because this is a dumb person. He's struggling because like the stuff we made was just bad. He was setting up a Wufoo. Wufoo was this, uh, form software. It's still reasonably popular. He wanted to take folks from his contact form. So in the UI, you had to select which contact, like which form do you want to use? And when you click that dropdown, instead of showing a list of here are all the forms that are in your Wufoo account. It didn't show the name, it showed the ID numbers of these, these forms. And so he's looking at it and he's like, I don't know which form goes with which. I'm like, oh, here, let me help. Let's go to Wufoo. And in the URL you can actually find the ID and then so you can match it up to that. So like that's the level, like the first version of Zapier was just like pretty janky, honestly. And, uh, anyway, we get, so we go through this whole process, he turns his Zap on and then we go to test it and we, you know, fill out his Wufoo form. And we click submit. He wants the contact folks to move over to his email marketing newsletter. So we go over to the email marketing newsletter. He's using AWeber. We refresh the page to see, hey, is the contact letter in our contact email inside that newsletter? And there it was. And his response was, oh my God, Wade, like, this is awesome. Like, if this— I do this every day, I port this stuff over, it saves me. This will save me, you know, 30 minutes every single day gone from my to-do list for forever. And I remember just like walking away from that thinking, holy cow, Wow, like we made something this guy cares about. If we just actually make the experience of setting up a Zap like not bad, I think this will work.

It's kind of ironic that it was a product around automation and saving people time, but it was so difficult to get that first one set up.

There's like this forever debate around like what is the— what is an MVP, you know? And I think in our case, like, yeah, we probably pushed the minimum side pretty, pretty heavily. Um, but the reality is like it was still viable, like it was still working and functional in the way that this person cared about. I think it, it works for us because Zapier is set it and forget it. And so it's not like, you know, if he had to interact with a Zap like every single day and set these things up, he probably would have been like, yeah, this is, this is no good. Like, I'm not going to go through this really janky experience, but he only had to go through it once. And once it was done, it worked for forever. And so for us, like that janky start actually was a perfectly viable starting spot for us. And then, you know, for us it was all about polish, polish, polish from there on out.

The problem was just so urgent that it didn't really matter what the form factor of it was if you could solve the problem.

100%. And for a lot of these folks, the alternatives, they don't have— they didn't have alternatives. You know, the, you know, if you're not an engineer, there's nothing you could do. Like, if you're an engineer, you could go use the API, you could write a script yourself, you could set this up and, you know, be off to the races. But even if you're an engineer, a lot of the times you're not really wanting to go spend time doing this stuff. And so the best that these folks could hope for was to go beg, you know, the engineering team, if their company was lucky enough to have an engineering team. And likely the engineering team was like, eh, get lost. We have other things to work on. Or you just try and figure it out yourself. And like, that's kind of a non-starter for a big chunk of the population, you know, circa 2011.

Yeah, it's almost like it wasn't urgent enough of a problem for some people, but it was super urgent for other people. Like if you actually had to build it, we're like, eh, doesn't matter. But the people who needed to use it, it was like pants on fire.

You know, a lot of companies you're hired in to do like a specific function and the engineering org is often hired in to, to own a certain thing, but it turns out the engineering skillset is quite valuable at every other function, but every other function usually is starved for the, that, that skillset. And the engineering function doesn't actually want to go help with those things because they actually have a thing that is more pressing to go attend to. And so all these other functions could benefit from the skills, but they didn't actually get hired in with those skills. Those functions don't exist. And so all the jobs that require that skillset, that, that team is sort of left wanting. Um, and this is sort of, I think, a big part of why, you know, no-code has been such a big trend over the last like 5 years is because people realize like, hey, There's a lot of folks that if you just gave them the tooling to do this stuff, they would happily go make it happen themselves. But you gotta, you know, get out of Python or JavaScript or what have you. Like you gotta provide it in a, you know, mechanism that's gonna be relatable for them.

And you've mentioned before this concept of, you know, product market fit. How did you know that you had it? Like, was there a way that you think about product market fit specifically with Zapier?

Michael Seibel, who's the, uh, runs the batch over at YC always has this way of saying like, you know, if you, if you have to ask if you have product market fit, you probably don't have it. And in some ways our experience mirrors that because, you know, I'd worked at a startup before Zapier and I was hired in as like the marketing person, the marketing intern or whatever. And, you know, they built this, it was very like well-engineered, well-designed product. And I was there to sort of like go figure out how to get customers for this thing. And, um, candidly, like I really struggled. Uh, in fact, I thought at some point in time, like they were going to fire me because like I just wasn't being successful. I was like, and, but I was learning a lot and trying a lot along the way. You know, I was learning how to do SEO and like PPC and email marketing and content and biz dev and sales, inbound, outbound. Like I was trying all of this stuff. Now perhaps not trying well enough, or at least at the time I thought, hey, I must not be very good at these things because it's not being successful. But you know, I did this for over the course of like 18 months. Like, so I was grinding on this and at some point in time I had convinced myself, perhaps for my own sanity, like I don't think that people actually want what we've made, even though it's well engineered, even though it's well designed. Like, I think we're sort of just like barking up the wrong tree, like wrong customer, wrong feature set, wrong priorities, et cetera, because just nothing's happening here. Whereas with Zapier, the experience was kind of the exact opposite. The product wasn't well engineered. It wasn't well designed, but when we put it in front of folks, when we help them set up their first zap, the reaction was always quite similar. It was like, oh my gosh, this is incredible. This is awesome. And so the better we made, the more we polished the product, the more we, you know, made the engineering more sophisticated, like the better the reaction from customers got. And it just kept growing and growing and growing and growing. And so it's not that Zapier was easy, like it definitely was hard, but it felt like all the problems we were facing were things that were You know, it's like we, we just sort of worked our way through it. Whereas this other one, it just sort of felt like I was trying to push this, you know, ball up a straight vertical hill. It's like, I can't, you know, I'm not Spider-Man. I can't walk up the side of a wall. Like, it's just not possible. And so I do think product market fit feels, you know, at a certain— if you've got it, it sort of feels like the boulder is like rolling down the hill and you're just trying to catch up with it. And it's like, this is working and I'm not even sure why. And so for us, it sort of felt like that. Where it was like, we, we kind of got it and it, it works.

Was there maybe a framework to think through when something is working or isn't working?

We were at the time, like, this is sort of like, uh, around the emergence of things like The Lean Startup and, you know, customer development, and, you know, Eric Ries and Steve Blank were doing his stuff. And so we were like reading all that material and pretty familiar with, you know, the concept of an MVP and like how to do this work. And so you know, we were, we were thinking a lot about like, okay, what is the like thing that a customer actually wants here and how do we try and get as close to that as possible and see what the reaction is? And so, you know, that was a big reason why, like, we were actually okay shipping a product that was, or like giving access to a handful of customers and sort of like a limited window with that we knew like, hey, this is not the end state, this is just the start. And then we were going to take that feedback and go make it better every step of the way. And so we were using those like ideas and concepts like pretty liberally.

Really want to jump into this company that you were at before Zapier, before getting in. But it sounds like that company, Veterans United, kind of had some of those values or, you know, all distilled in the way they operated. Yeah.

So Veterans United was actually in between that first startup and this one, and Veterans United was also like a formative place to be, you know, there. I think the thing that we learned a lot from them, it's owned 50/50 by two brothers. These two brothers have— they'd sold a couple of companies before, also bootstrapped entirely. Veterans United, bootstrapped entirely.

How big is it?

It's like owned by these brothers. They do, um, I think they sponsor like a sports stadium or something like that in Virginia or something like that, maybe, or like a big amphitheater. I don't know. So like, they're, you know, they're doing big company stuff here, right? Like, this is not, you know, sort of a nobody company. I think their number one competitor is USAA, I want to say. So if you've seen like the USAA loans, like, they sell to veterans in the United States. They specialize in a a veteran's home loan is what it's called. They took off during the financial crisis selling mortgages of all that, if you could believe it. And so like, sort of feels like weird. It's like, wait, how do— how are they doing this? And well, it turns out the government backs a VA loan. And so when all these other folks are struggling with mortgages, this is actually a safe haven. It's like, this is the type of mortgage that you want to get. And so they're really specialized in, you know, very good at online marketing. So they know how to like pull in the folks searching for you know, a VA home loan. And then they get really good at operationalizing that stuff on the backend. So at first they were selling these leads, but then, uh, they realized like they, they could actually make a lot more money if they actually underwrote the, the mortgage underneath it. And so, um, that's what they started doing is they figured that was the more valuable bit. So it's just like very good at marketing, very good at operations. And when I joined, you know, this was, this machinery was all built. Like they were already like, you know, I contributed very little. So many other people sort of get the credit for this, but I was employee like 500, uh, and then left 10 months later, uh, you know, as we were starting Zapier and there was 1,000 employees there at that point in time. And so, you know, I think the thing that we learned out of that experience was just, you can, things can grow extraordinarily fast. You can have extraordinary product market fit. Uh, you know, do not discount what great marketing and great distribution can do for you. And this idea that like bootstrapping equals like lifestyle business is just false. Like there's, you know, exceptional businesses built who don't raise money and there's exceptional businesses built who do raise money. And so, you know, I think we just like, we're a little bit, you know, dismissive of like zealots, I guess, at the end of the day. And there definitely is sort of like zealots both in the like bootstrapping camp and in the like, you know, venture community who sort of think like there is only one true way to, to build a company.

Maybe I'm part of the problem, but there's definitely this kind of like, like VC, uh, industrial content production machine. It tells you, you gotta, you have to raise money from us. Like we are the only way we save your company. Look at how awesome we are. You need, you need us, which, you know, I actually don't fully agree with. I mean, to your point, like if there's so many cases where it's like you really probably shouldn't, or maybe you raise a little bit. I mean, we'll get into your guys' story, but I think it was an interesting approach. Like. You raised a little bit, seemed to be pretty reasonable about it. One thing that feels like enabled that was you don't think Zapier would have worked if you hadn't figured out this like near zero customer acquisition cost strategy. Can you just kind of maybe talk through that?

One of the strengths of Zapier is our SEO engine. And, you know, it's sort of the way it works. The thing that we recognized was that, um, well, uh, we, we sort of borrowed this from this guy, Patrick McKenzie, Patio 11. He's like somewhat popular on the internet, um, Hacker News community, things like this. But, uh, you know, maybe before you knew him as, you know, sort of Patio 11, the money man or whatever, like he, uh, he had this like small, uh, indie site, uh, called Bingo Card Creator. And he wrote a lot about how a bingo card like creator worked. And, um, you might think like, what, what does a like, you know, SaaS company have to learn from bingo card creator? Well, it turns out like quite a bit. What we thought was like a pretty novel thing. And he, he realized that, you know, if he was going to get distribution for his bingo cards, he had to find a way to rank for like various keywords. And, um, you know, when he looked at, like, did some basic keyword research, he started to realize that there's a group of folks who, like, are constantly looking for bingo cards. Um, you know, a lot of them are teachers, they're trying to do lesson plans and things like that. Turns out there's a lot of different subjects in school. And so it was like, hey, do we have a US history bingo card? Do we have a world history bingo card? Can we have a physics bingo card? Can we have an algebra bingo card? Can we have an XYZ bingo card, right? And you could sort of Yeah, you know, sit down and dream up like all the different permutations of a bingo card that you might want to have. And the thing that he figured out was like, how can I make it really easy to build a specific landing page for that exact type of bingo card, a US history bingo card, an algebra bingo card, and how can I make it so that it is as cheap as humanly possible to make a card that targets that exact thing? And so he made a little CMS. And then he hired, I think it was a stay-at-home mom who was looking to make a little bit of extra money while, you know, she was like working with a, you know, an infant in her lap or whatever and just said, hey, I'll pay you for every single page that you can make. And so she would just sit there and just think of— I think she actually was a former teacher. So she would just think through like all the things that she would want at school. I think he— I forget the exact math, but like, call it, you know, it cost him like $1 or $2 to make a landing page. And he knew that for every landing page he made, he would generate 2 or 3 sales. So not much, like, you know, it's not like this is a highly profitable endeavor, but he would sell them for like $10, $20 a pop. So if you think, oh, it costs $2 to make the page and I can sell 2 of these, all of a sudden it's like $2 to make it and I make $40 on the other side. So like the math sort of makes it like, makes sense where you're like, hey, this is a like decent little setup. And So we look at this and go, oh, that mechanism is actually quite interesting because it turns out when you're looking for integrations, it has the same phenomenon where it's like, I want to connect, you know, Salesforce and Marketo, or I want to connect, you know, HubSpot and a Google Sheet or whatever. Right. Lots of different permutations and combinations of these things. You know, there's probably not that many people searching for any particular one. But if you have a lot of these things, there probably is a lot of people searching for the, the category of these types of integrations. And so our entire strategy then became, let's make sure that we have landing pages that are specifically targeting these things. And then we need to keep adding more apps to the platform so that we can have more pages that folks come, come looking for. Uh, and so that ended up creating a very low customer acquisition cost because you would go search, and this is organic search, this is not paid search, mind you, um, for these things, you would find Zapier in the, you know, front page of the listings. You'd click through, you'd sign up for a free trial, and you know, as soon as you'd upgrade to paid, like, you know, we basically get— we're basically profitable from day one because we didn't really pay that much to acquire the customer. All it cost was the amount to build the integration, and that was like a relatively low upfront cost for us. Uh, and so, uh, it definitely enabled a lot of interesting dynamics in the business to, to, to help us scale to where we were.

Your product and distribution were kind of the same, or they just intertwined and worked so well. I don't know if that, if that's fair to say.

I think that's 100% true. I, I think like when a lot of folks think about like product market fit, they think of it as like an ordering thing where it's like, oh, first we need to make the product and we're gonna make this product that works this really well. And then, oh, the product's great, we've got a great product. Now we gotta go figure out like how to go market this thing. And then they're now they're trying to like bolt it on. And it turns out, like, I actually think product-market fit is actually one thing. It's not a sequence of things. You want your product and your go-to-market to be sort of one fluid motion. Um, and the best PLG companies, like, this is how it works. It's like sort of intertwined in the same way. And that's why I'm always like, you know, I have founders come up to where they're struggling. It's like, I've got a great product. Like, the product's not the problem. Like, I gotta go figure out how to get distribution for this thing. And I'm kind of like, well, Are you sure? Like, are you sure you've got a great product? Like, because what I'm hearing is like, I actually kind of don't think you have a great, like, a product. Maybe, maybe it's a well-engineered product or a well-designed product. That, that absolutely could be true. But is it a product that anyone cares about? Because if no one cares about it and you don't have distribution engine that shows that people care about it, I think you might want to go back to the drawing board.

Yeah, yeah, I was going to say, do you just start over? Like, you got to rebuild the product completely different.

Yeah, I mean, sometimes you can sort of like take bits of it, novel pieces of it, and, you know, you can sort of quote unquote pivot into something that, you know, maybe it is a bit of a nugget that somebody cares about. A lot of this is like there's a lot of art in it though. Like you're sort of going through the idea maze. You're trying to like put things in front of customers, figure out what they're reacting to, what they like, what they don't like. I think the more working experience you have, probably the easier this gets because you just have a better sense for like what problems exist and like what those potential solutions can be. And so it does help to like have some exposure to that particular thing. You know, for me with Zapier, it was like a problem I felt like I was working with the Marketo API, knew what was painful about it. And so I sort of felt like, well, I kind of could be like customer number one for this thing. And, and that helped, like that's short circuits what you're building when you sort of know like your problem. And if your problem is representative of a problem that others have, that was pretty helpful.

You said something that I thought was really interesting too. You said. You don't think the exact same growth playbook that you use on Zapier could work today? Can you kind of explain that? Because it's, I mean, I hear a lot of people, they say, oh, you know, this SEO strategy that Zapier used, we're going to use it too.

To be clear, I think you could use this like long tail thing like this. A lot of folks do this, right? This is how Yelp works. This is how Thumbtack works. Like most marketplaces work this way. You know, Airbnb and the like sort of have the same dynamic where they're targeting these long tail keywords. I think if you were to try and do it in any of those spaces I named, like if you were trying to compete with Zapier or compete with Yelp or compete with Airbnb or compete with any of these folks, you're going to have a tough time because those keyword categories are so crowded at this point in time. Whereas, you know, 10 years ago when we were doing this, the keyword categories were not crowded at all. There was nothing there. Like if you go search for the stuff, you would find like these pastebin links and, you know, uh, like really bad API documentation and all that sort of stuff. Versus if you do these searches today, you see Zapier and Zapier lookalikes, um, all over the place that are sort of running this playbook. And so it's just really tough to do it in this particular space. But if you have an idea that is— and you go do searches and you see that like, hey, there's nothing sort of in here, like, I actually do think it could work quite well, um, in, in other categories.

You need to find new and undercompetitive distribution channels. And kind of relating back to what you said earlier about Yeah, combining the product and the distribution, it's, I mean, I think it's nearly impossible. It does not happen very often, but it's almost like you go, you go to the fringes, the edge of the internet, new platforms or channels that are kind of emerging. Um, I don't know what that is today. Maybe it's ChatGPT.

ChatGPT is probably a good one. TikTok, you know, Discord, like there's always something to these like emerging channels, emerging platforms where there is usually like a You, you can sort of get outsized returns because they probably, you know, like the, the big players probably aren't playing enough there because they, they can't make the math work. They sort of look at it and they're like, we don't know how to measure it. We don't know how to do XYZ. So like, we're probably not going to invest there. Whereas a small company can come in and be like, I mean, we can't measure it either, but it like kind of doesn't matter to us because we have no business. So like, if it works at all, it's great for us. So many companies are built off the back of this. Like, you think of how many companies got built off the back of, say, like Facebook early on, like Zynga, like you know, exists because, you know, the Facebook platform totally, uh, was there. Now, after over a while, um, those opportunities start to disappear because the platform matures. And every, you know, Facebook like has an advertising platform, Google has an advertising platform, they know exactly how much to charge you to like extract all the marginal utility out of that. And so the, you know, ROI you get from working on those platforms like starts to stabilize, and you don't get these like outsized returns on the more mature platforms. But, um, yeah, you can at times, like you just sort of, like you said, you have to find the niche, you have to find the nooks and crannies that, you know, just are underexplored, underappreciated. Um, but if you can find them, they're, they're worth their weight in gold for sure.

Yeah, I'm trying to think of an interesting example right now would maybe be, uh, this thing called TikTok Shop. Are you familiar with TikTok Shop?

I'm not.

It's basically shoppable TikTok videos. So, you know, I make a video about this sweater and you like it and you can buy it directly from the video. I have some friends that have small brands and they love it. They swear about like, this is working so well. And then some people that I talk to are like, that'll never work. It's so dumb. It's like no one will, no one would ever use this.

Yeah. It's that information asymmetry where it's like you sort of discover something before anyone else does. And yeah, for a while times are good.

Yeah. I know at one point you went through YC. How did that come about? Uh, Y Combinator, cause you kind of talked about this whole VC versus bootstrap.

Yeah. So we, uh, you know, we'd started Zapier at like a startup weekend and YC had its applications, I think like the next week. And, um, we didn't know anything about like venture community or anything like that. But, uh, the one thing we did know was Paul Graham's essays and Hacker News and So we were like, huh, like, you know, if we could do the YC thing, I think we would do that, you know, especially because we're trying to build this like integration platform, automation platform. And, you know, a big part of that strategy is like hooking into all these different folks. And we're, you know, sort of 3 random, like more or less kids in the middle of Missouri at this point in time. And it sort of felt like, you know, it'd be nice to have a little bit of credibility. And YC felt like a way to get some of that credibility to sort of get a little on the inside. So anyway, we apply for it and we sort of get straight rejected the first time around because, oh, it sort of makes sense. It's like, you know, we had nothing to show. Like, we had barely a prototype. And if you looked at any of our resumes, our track records, we didn't really have much to show there, you know? And so we were like, okay, well, that sort of makes sense. And, you know, didn't deter us. We're like, we'll just keep working on this thing. And so we just worked on it nights and weekends for the, you know, next like 6 months or so. And, you know, we felt like, hey, maybe we'll try the YC thing again, you know, when the next batch comes around, or maybe not. Like, it wasn't— it was just sort of like a little bit of a flyer. Like, if it works, great, but if not, we're going to keep doing this thing anyway. So anyway, like, the next batch does come around. We've made a lot of traction at that point in time. I think we had about 1,000 people on a beta, we had about 10,000 on a waiting list, uh, but we also had a fair number of YC companies that were using Zapier. And so we went to go apply again, we started to ask for, hey, like, can you give us some tips and tricks on like how to apply. Like, is recommendations a thing? Can you like, you know, tell, tell them that you like our product and stuff like that? And so, you know, we, we applied the second time and, you know, I do think like just the fact that we had just demonstrated, you know, from application one, it was like, here's all the things we're gonna go do. And by the time application two came around, it was like, and now he had done them and it was like, and they are working. And so it just sort of made it one of those things where it was like, okay, we don't have a track record, but the second time around we did. It was a short track record, but it was, you know, uh, a pretty legitimate one. Even still, it was like kind of a nerve-wracking, you know, experience going through their whole interview process and whatnot. But, uh, at the other side, they did accept us, and so we ended up going through the summer 2012 batch, which had Coinbase and Instacart and, uh, Benchling and a fair other— like, it was a good— ended up being a pretty good batch.

Yeah, those are some heavy hitters. You're like the maybe 4th, 5th, 6th biggest company out of the batch.

We're kind of lower on the list. Yeah.

Um, so for somebody who's not familiar with YC, Y Combinator, can you like just explain how it works, like what your experience was like going through? Don't have to be super, super long, but yeah.

I mean, so basically YC like funds very, very early stage companies. So think folks that are like idea stage or idea and like basically a prototype or like a prototype plus a little bit of traction. So it's very, very early on and, uh, they fund them in batches. So you get accepted into a batch alongside— I mean, I don't know how many companies it is now. I think it's like north of 100 companies. And you all go through this like 3-month period where you're sort of very involved with Bill. Like you do office hours every week with, you know, one of the partners. Um, you get to see all these other startups working at the same time too. It's like this little pressure cooker where you're just like ambition levels go through the roof and you're just trying to see like how much can you get done at the end of 3 months. And then at the end of the 3 months, they do this demo day where you could kind of get to show off what you do. They invite a bunch of investors in, and, uh, you know, folks can raise a bunch of money. And certainly a lot of YC companies do. But, you know, I think more of that demo day is, is just a forcing function to see like how much can you actually deliver in this like very brief time. Because the reality is starting a company is like so hard. Like you have to— like, if you as a founder do not actually do the thing, Nothing happens, which is so different than from an established company that has a lot of things going on. And so the idea of putting it in a batch and putting yourself around other people create this like friendly competitive environment where you're just kind of want to see like, how much can I actually make happen? And it turns out you can actually compress like a ton of work into relatively short period of time. And that does, I think, help shoot the best ideas through and you kind of get through this like really nascent stage where like the default path is it probably dies. But you can, because you're in this pressure cooker, you can actually push those ideas through and they can sort of like come to life in a way that gives them enough of a runway to, um, start to get a bit of a flywheel going.

You know, I meet a lot of YC founders. I've looked at, specifically looked at the deck, like Coinbase's Demo Day deck. It's like, that is a competitive thing. Like, that's an exciting thing to be building alongside. You're like, wow, look at how well some of these some of my friends are doing that I'm meeting. I mean, it's invaluable.

100%. Yeah, there's a lot of smart people. I mean, I certainly felt imposter syndrome being in that, you know, sort of the Tuesday night dinner rooms where you're around people, you're just like, these people are like incredible. Like, it's insane. I can't believe I'm here.

How did you get through the imposter syndrome as a founder?

You don't, uh, like, I think at the end of the day, like, I think, uh, the reality is like The longer you go, you, you still, like, I still always, you know, you get invited to the next thing and the next thing, and you start to like get, find yourself in like more what feels like more and more exclusive rooms. And every step of the way you just sort of feel like that, I, I can't believe I'm here. But I think what happens, you get, you get better at it because, um, you sort of realize like, oh, I've done hard things before, I've done tricky things before. And so yeah, this one's going to be tricky too, but I've know that I can figure things out. And so you'll get kind of tossed into those things and you're just like, okay, let's go figure it out. Um, I think there's like a certain nativity that sort of like helps you just work your way through some of that stuff. And the reality is like all companies are sort of like this. You're doing— every company is doing something novel, unique that hasn't been done before. And, um, the best ones like just keep finding a way to solve one problem and solve the next one after that. And then the one after that. And so that's just how company building is. And if you do that for 12 years, you start to go like, all right, bring on the next one, right? Like, give me a pandemic, I got it. Give me a banking crisis, I got it. Like, give me the next thing, like, we'll, we'll just keep doing it.

Were those— I mean, just because you bring those up, was— were those interesting for Zapier? Like, any—

I mean, not really. Like, we, uh, we would always been fully distributed, so we didn't really have to deal with the pandemic thing all that much. And then we didn't have any exposure to SVB, so it was not really a huge deal there either. But You know, I saw like, we've had versions of that that are obviously less public and stuff like that. And, you know, you just like set one foot in front of the other and just keep figuring it out.

Any, any of those less public things you're open to sharing? Just like near, near-death experiences, anything kind of crazy that happened that you've never talked about before?

The bigger mistakes we've made are things of inaction, to be honest. And I think these are things that happen are probably less relevant for a very early-stage founder. But there are things that happen as companies grow that are really important where You know, as you grow, um, hire all these folks and, you know, you hire really smart, talented people and they have a bunch of advice, but alongside that comes in process. It comes in rigorous thinking, comes in a lot of things that are quite helpful. But alongside that, you can get paralysis. You can get— you can always talk yourself out of idea. There's always a reason not to do things. And so that like mindset can actually slow you down from taking chances, from taking risks. And so when I look back and think through some of our biggest mistakes, they're either places where we acted or didn't act, or we acted too slow. And so a big thing that I try and preach internally is how do we decide in situations of uncertainty? How do we take chances when we don't know? And what does it look like to be rigorous, but also decisive? And so those are the things like without getting into specific stories, I do think are really important for companies that are growing and scaling to, you know, sort of keep part of their DNA. Because if you don't do that, you will sort of tap out eventually, because somebody else is going to be coming up right behind you, willing to, to take a big swing, willing to do something insane and different that, like, the math doesn't check out, but it turns out is actually quite interesting.

I had a conversation with Atai Damty at Unit. He has this concept of operating with precision. Which to really distill it down, it's basically take small measured steps in the most highest priority thing of whatever you need to do. It's not quite related in terms of like avoiding inaction, but it, it kind of helps you just making sure what you're doing is important and you, you get it done and you, you're making progress towards something. So example would be, you know, like for your SEO strategy, for the landing pages, you probably could have spent days, weeks, months making the perfect landing page, or you can just make something in 5 minutes, see what happens, test it, see what you need to fix, slowly incrementally improve it. And you know, 2 months later it's perfect. And meanwhile, if you would have waited to launch and got it all perfect, like it's not even there yet. So I don't know, that's a recent lesson I've learned on that.

The one I liked recently is I heard Jensen, the CEO at NVIDIA, talking about how he likes, uh, markets that are, uh, the, the markets zero. Um, and basically the idea is that like he, he wants to be in markets that, uh, the, the biggest markets over the long term are usually the ones that don't exist. And so he wants to be there at the ground level and then he wants to surf the wave, uh, to, to, to the highest heights. And so I, I always thought that was like a good mental model to think through. It's like, oh yeah, the best ideas often are going to be ones that look a little weird and atypical because they're, they're nascent. They're, they're not all the way fully formed yet. And it kind of comes back to our distribution conversation where, you know, the best marketing and strategies are, are in niches or in things that look small, but they've got an exceptionally high growth rate, exceptionally high track record. And if you sort of catch onto those early, you know, you're, you're positioned to, you know, have a decade-plus worth of, you know, gains over the long haul.

Yeah. I mean, even in, in your batch, Coinbase, like at the time, Bitcoin, What is it even now?

Yeah, I mean, I remember like Brian was like handing out Bitcoin during our batch and stuff like that. So, you know, kind of wild. Yeah, it's like if you, you know, come try it out and I'll give you free Bitcoin and whatnot. Um, unfortunately I did not do that.

Oh geez.

But you know, okay.

Well, yeah, I guess kind of going back to that demo day. So did you guys do a— did you present?

Yeah, everyone does like a 3-minute demo at the very end.

How was your presentation? Do you remember?

It's fast. Like, I think the thing that, uh, they really drive into you— PG really drives into you— is that You get 3 minutes and they actually made, they made 'em shorter because they used to be much longer, but they realized that, you know, the longer they were long didn't, didn't actually mean better. And so the thing that they really drive into you is that folks are gonna remember 1, maybe 2 things at best about what you're talking about. So just condense your entire thing down to the 1 or 2 things that absolutely matter most. Just talk about that and don't spend any time on anything else because most founders get mired in all the details and all the complexity and just do a bad job of saying, this is what we're making. This is why it's a problem. This is who has the problem. This is how fast we're growing. And if you sort of like nail those things, that's usually enough for somebody to go, okay, I'm interested. Let's talk more.

Do you remember what that was for Zapier?

I mean, for us it was the fact that like people need integrations and the number of apps that exist is huge and it keeps growing. And if it keeps growing at the rate it's growing, there needs to be a company sort of sitting at the center of this, helping connect everything.

It seems like all the cloud software was sort of hitting scale or had like inflected at just the perfect time.

100%. Like Stripe had launched the year before, Twilio had launched the year before, SendGrid had launched the year before. So like all this idea of like APIs and like connecting to stuff was just starting to be a thing. You know, there was some, you know, like Salesforce had an API for a while, but like beyond them, like really wasn't that much. And if you had to sort of waited 2 years later, it was sort of de facto the way you do things. So it was sort of like right in that moment when this was starting to be a pretty critical problem.

One of the things I've always just admired about what you've done, you've basically zigged when others zagged. One of the things was related to fundraising. You essentially only raised money. Issuing shares one time. You did a, like, a $1.3 million seed round right at YC. Can you just explain kind of what you did, talk us through it, rationale, all that stuff?

That is right. So we, you know, when we started the company, we're in central Missouri, there's no such thing as like investors, uh, in central Missouri. So like, we're just like, well, you just bootstrap. And we, we had the Veterans United experience, so we knew that like, okay, that's how companies get built. You better not spend more money than you make. Like, that's sort of just how you survive as a business. Looks a lot like personal finance at the end of the day, like, you know, uh, and so that's sort of the mental model we came in with. And then obviously getting exposed to YC and Silicon Valley, you start to see, okay, there's this other way of building companies too. You know, as you read about that, you start to see, okay, yeah, there's some like pros to doing that stuff, but there are also like some big cons. Like you hear the horror stories of, you know, VC folks like injecting a certain mindset that actually can really kill these companies. They sort of force growth at all costs. They sort of give you too much money that you end up spending and blowing way before you're ready to do that stuff. And then if you actually looked at some of the biggest companies, you think like the Microsofts of the world, like they didn't actually raise that much money before they went public. Like they raised relatively modest amounts. And I mean, compared to today's companies, they raised basically nothing. And so you start to realize like money is only like the amount of money you raise is very, it's only loosely correlated to long-term outcomes. And, you know, I think for us, we just sort of looked at it and said, okay, we're going to take this money because we felt like we needed it. But we're going to treat it like it's the last money we ever had. And we never thought of it as raising money. We always thought of it as like selling a part of our company. And if you sort of think about it, that's actually what you're doing. And we sort of felt like, well, we don't really want to sell our company. We like our company. And so if we have to sell something, we'd rather sell a lot less than a lot more. And it sort of felt like that was the best way to sort of maintain control and have the ability to sort of shoot for something bigger versus if we sold more of it, we have less control, and that mean that somebody would have more of a chance to mess up the thing that we were trying to build, um, at the end of the day. And, you know, some of that was probably true, some of that was probably a little bit, uh, you know, sort of maybe not quite accurate. But, um, regardless, we sort of said we're gonna take the money, we're gonna treat it like the last that we ever have, and then that's just how we're gonna operate the business. And then along the way, you know, we would every so often, like, not a huge amount of time, but maybe every year or so, like, we would ask ourselves, like, Okay, do we think we need more money? Is this actually like, or is like money slowing us down? And every time we'd have that discussion, you know, we'd be like, nah, I don't think so. Money is not the thing that's holding us back. Like there's other things that we felt were more like bigger bottlenecks in the business. And so we just never did go back to it. And, you know, I probably helped like in the early days, like there wasn't a lot of folks like wildly interested in it, but certainly as we got, you know, more, more customers and like a little bit of a high pro— higher profile, there was a lot more people interested in it. And then we got a lot more curious looks where people are like, why aren't you doing this? This seems dumb that you're not raising more money, but for us, we were sort of like, we don't feel like it's going to help us. So we just steer clear.

I've heard you mention how fundraising works generally. It's basically you are convincing them to buy or they are convincing you to sell. I mean, it sounds like you, you did, you did that pretty well.

The first time around the seed round, it was more of like us trying to convince them to buy for sure. We went through YC's Demo Day and I think of the folks we met through Demo Day, only one person invested. The everyone else that invested were folks that sort of reached out to us. Um, you know, in the 6, 9 months prior, you know, it definitely was a situation where like the people that liked us, they just liked us. Um, and you know, when we reached out, like people didn't get it. Like I didn't tell the story good enough, like insert probably all the reasons people have a tough time raising money, but we were still able to, you know, get what we needed and get back to work.

Um, at the end of the day, do you think you would've preferred to maybe raise a little bit more in that first round or?

No, no, preferred to raise a little less, actually.

Really? Okay. I'm just curious about the product evolution over time. You obviously had a product that worked. What did you do next? Like, what was kind of the next big feature or direction that you guys went?

So the thing we're always trying to do is like make automation work for everyone. And so we're always looking at the product on, on more or less two dimensions. Um, how can we make what we have easier? And how can we sort of open up new dimensions for automation. And so, you know, the first one was, uh, probably the first things we noticed was that, hey, the fact that you could sort of only have a single trigger and a single action is like relatively constraining. And so one of the big early additions to Zapier was multi-step where you could now automate these end-to-end workflows. And so that was a big evolution. There was a lot of things that we did that were smaller and probably less visible to then polish along the way so that, uh, you know, more and more people could self-serve, come in and create those automations. More lately it's been trying to open up the aperture for automation. You know, we've got a tables product, an interfaces product, an AI chat builder product, and these are all new mechanisms for automating your work. In each of these things, we're looking for ways to simplify, make it more accessible, to take jobs that are like somewhat tedious, hard to scale, and make it easier for folks to automate those things and then scale them, uh, alongside of it.

And you did this thing where you did a developer platform. I don't think it's really that controversial. Pretty common. How did you get that to work?

Well, uh, yeah, we did this very early, so this would have been in 2012. We had at the time about, I think, 50 apps on the platform and a relatively small number of customers. And early on in YC, they sort of were pushing us to say, what, you know, what are you going to show off on Demo Day? What do you want to say about Zapier on Demo Day? And we, we knew at that time, like, The thing for us was like, we just need to hook into more stuff. Cuz if we hook into more things, it's gonna open up more capabilities for customers. It's gonna mean more landing pages, it's gonna mean more, more growth. So it was like, we just need more apps. And so the big debate was, okay, we can get, just go the 3 of us, we can just go brute force this. We're just gonna go integrate with as many things as we possibly can. And we felt like that actually would be the short term, the faster thing to do. Or we can place this bet on this developer platform and see if third parties would want to be willing to build apps onto Zapier too, which felt like a bigger risk because a lot of times people don't— people build developer platforms, but not a lot of people build on them, uh, after you release it. And so it felt like, well, we could go do this and if it works out great, it's amazing. Like you get all this, like this beautiful ecosystem where tons of developers are coming and building all this magical stuff that you didn't anticipate on top of it. That dream could come true, or you could release this thing and people are like, yeah, we don't care. And so we were having that debate, like, which, which path do we go down? And Brian was like, okay, let's, let's try the developer platform route because, uh, we'd gotten this email from Aaron, the CEO of Box on like a 3:00 AM on a like Saturday morning or something like that. And he was like, why isn't Box on Zapier? And the answer was like, well, it should be. We just, we're 3 people and we just haven't gotten to it yet. So. Hang tight. But it was like, well, if he's willing to send a 3 AM email, maybe he's got a developer over there that would build an integration. And Brian felt like he could build V1 of this, and I think he thought he could do it in like a week, 10, 10 days, something like that. And so we're like, well, if you can do it that fast, then what the heck, we might as well try. Because if we get sort of this like magical ending, they'll be— it'll have been worth it. And so, you know, that's what we did. And, uh, from there we were starting to get some inbound interest. You know, folks like Aaron were saying like, hey, why isn't this on this? And so we started to take that inbound interest and say, hey, we've got this developer platform. Uh, we haven't launched it yet. Why don't you build on this and we'll help you? And so, you know, we had about a dozen folks, um, that we worked with and including some like relatively popular folks, um, who were a part of that. So like HubSpot was a part of that. ActiveCampaign was a part of that. There's a really popular WordPress plugin, Gravity Forms, that was a part of that early, like 12 companies or whatever. We launched, we had like these names plus like, you know, probably another dozen associated with it and said, great, we've got this developer platform, come integrate stuff. And, um, TL;DR, it worked. Like, you know, it was, and it wasn't necessarily this thing where like, it was like, oh, now everyone's building on top of it left and right, but it just sort of was this like steady up into the right thing where the flywheel was just working. We're getting more customers, we're getting more apps, that's attracting more apps, which is attracting more customers, which is attracting more apps. And the flywheel just started to grow faster and faster and spin faster and faster as we went. And so, and you know, I think then our job became just polishing the flywheel. So you can think of the three of us and the staff we're hiring is sitting there figuring out anytime there's friction in this process, how do we make it go faster? And so we're trying to get in there, help customers, help developers, trying grease the flywheel. I think a lot of times folks think like, oh, you build it and this like magical thing just works. It's like, well, no, you can see how it could work. Like you sort of have it going, but it's going really slow and you're just saying like, okay, well, how can I get in there? Like just push it and you just keep pushing it a lot. And the more you push, the faster it goes and the faster it goes, like the faster it spins on itself. Uh, and so that's what it sort of felt like with that first version of the developer platform.

And it's interesting when you think about distribution and product kind of tied together, each new integration adds more distribution in terms of searches and the company probably is promoting you and totally, it's a mini little network effect there. So then after, uh, multi-step, you added collaborations. What is that? Is that like two people can log in at the same time and build things together essentially?

You know, the first version of Zapier was like single player. It's like you can automate stuff for yourself. And if you wanted to do it inside of a company, you know, that, that's a problem. Well, it was just hard. And so we wanted to make it easier for you to work with your teammates. And if you decided to leave your company, the company shouldn't sort of be caught high and dry, um, with no way to sort of like support these apps. And so So yeah, it was just, you know, a way for there to just be more like, I guess, oversight, more controls, more collaboration across these things that are being automated. Because, you know, as we grew, it was less about me automating Wade's stuff and more about me automating my company's stuff. And the company should have some ability to make sure that this stuff keeps working even when employees are coming and going.

Another big thing you did was just making it visual. I don't think I've ever used the non-visual editor. What was it like before? They kind of had like the WYSIWYG click and point, all that stuff.

I mean, it was always relatively visual. I, you know, I think it, it was very linear, the thing that sort of the very first versions were. It was just like this, this, this, this, this. Now you can sort of see it in more of like a canvas view. Um, folks like to call it more visual, but, uh, uh, I think it helps especially when we launch paths. Like paths was a big thing that like in that visual, in that linear model was like tough to see all the different branches. But when you have paths, like you actually want to be able to sort of like zoom in and zoom out and see like how these workflows like fork and move into different areas over time.

So is paths almost like an if-then?

Yeah, exactly. It's like, you know, it's like if the leads do this, send them to the CRM. If the leads do that, send them to the, you know, email newsletter. If they're like, do this, like, let's call them right now because they're worth their weight in gold.

And then most recently you guys have been doing a lot just around AI. I mean, a lot of people have been doing things around AI, so it's not like— it's not like— yeah. So what have you been doing? Like, can you just talk us through?

Well, I think there's, you know, a couple things. I think LLMs, um, the way my co-founder Mike talks about it, which I, I will lovingly steal, is that, uh, it's a brand new sort of, uh, way of solving problems. And so one of the things that I think is smart for any company to do is to go look at all the tough problems they faced over, say, the last, I don't know, 5 years, 10 years, whatever time frame you want to say, and, and look at those problems and say, you know what, could an LLM sort of help me with this? And so for Zapier, that's a big thing that we've been looking at and saying like, okay, where are these problems that have been like really tough for us, sticklers, things that, you know, we just felt like it's as good as it's going to get. Like where you sort of just, you know, we've tried every angle, we're out of ideas and say, well, actually an LLM is a new idea. And so we've been looking at like those types of areas and, you know, we've been looking at it from a couple of dimensions. One is just. Internal operations. There's a lot inside of the internal operations of the company where we think LLM can make things more efficient or can bring new ideas to the table. But the second thing is, where can we actually expose this to our customers and give them these capabilities as well too? And turns out there's quite a bit of areas. And so, you know, we've got, you know, a handful of places where, uh, Zapier has gotten a lot better if you're a user because we have LLM's power and key features. So for example, if you sign up for Zapier today, I did actually.

Ah, nice.

So if you want to create a Zap now, so instead of having to go through and set up a Zap, you know, step by step, you actually can just chat with Zapier and say, hey, here's what I'd like to automate. Or I don't know what to automate, but I'm a podcaster who likes to tell jokes. Like, what kind of things should I automate? And it'll be like, ah, here's some ideas for podcasters. And then you can go, oh, actually, I would like to do that. Maybe I— can you set that Zap up for me? And it will attempt to set up that Zap. And it does a reasonably good job. The conversion rates are actually better than the old manual configuration route. And so we're finding that, you know, these are ways that we can make automation just much easier for customers. And so I think that's the power here is how can we give these powers of LLMs to automate things and to recognize that LLMs are a way of automating stuff that is new. It's a new form factor. It's similar when Zapier first came along, you could do it without code. Well, now you can do it with less configuration. And so that's why we feel like AI LLMs opens up a whole new set of capabilities. The last thing on LLMs that I do think is also really important is most of where Zapier has spent its time is automating things that, um, were very structured. You know, take this bit of data from your CRM, put it into this other application. However, a whole bunch of the world's data and information is actually in unstructured data, and it's relatively tough for us to work with an API around unstructured data. However, an LLM actually is quite powerful at working with unstructured data. Turns out you can feed it paragraphs, words, things like that, and say, extract something from here, or summarize it, or add more detail to it, or categorize it. LLMs have all these really interesting capabilities that can help you take things from unstructured to structured and back, which opens up a whole new world of automation capabilities that didn't exist, you know, or existed in limited, um, capabilities, uh, until they sort of have been popularized, uh, in the last year or so.

Was there a light bulb moment in terms of internally you were like, we need to take AI seriously, or we need to start building things?

Or we'd started to like play around with it probably, I want to say like 12, 18 months my co-founder Mike, uh, was like, hey, I, you know, he was, he was in a management role and he's like, you know what, I don't think the best thing I can do for Zapier is to stay in this role. I think I need to go be an IC and go figure out like what are we going to do around AI and LLMs. This was pre-ChatGPT and so like the world sort of hadn't been, you know, lit on fire yet. But him and Mike went in and, and we're just like experimenting, you know, playing around, trying to figure out like what's going on with this stuff. And then, you know, ChatGPT came out and I think us, like the rest of the world, start to go, okay, like, what are we, what are we doing about this? And so, uh, we had some smaller teams that were starting to work on this stuff in a more organic, bottoms-up way. And, um, you know, we were starting to get some pretty interesting things going. But, uh, you know, when, uh, GPT-4 launches— this, uh, I guess this happened in March, I want to say, of this year— and we start to go like, wow, the capabilities are advancing really fast. This sort of like organic bottoms-up approach, sort of like fueled by, you know, Brian and Mike, my co-founder, sort of love of what this is, is, is actually not fast enough. Like, we need to sort of shake the company, for lack of a better word. And so we said, hey, we're gonna, we're gonna call an AI hackathon, and we're gonna shut the company down for a week and say everyone inside the company is going to go use AI tooling however you want to do it. You can hack on it, you can use it as an end user, you're going to operationalize it. And the idea was just to get people exposed to what's going on in this world. That was a big pivotal moment because all of a sudden that's where all these interesting ideas for how can we take the product started to come to life. Instead of being just sort of like, oh, an interesting idea, you started to see these things show up on the roadmap, figure out how to get a V1, figure out how to get it out in front of customers. And it was because tops down we sort of said, hey, like, I know y'all got roadmaps. I all know you got all this stuff that's going on and all that stuff's important. But you know what, for a week we're gonna set that aside and we're gonna go figure out what the heck is going on with it. And you know, it's kind of an important thing because, you know, when you've got a business, it's so easy to just be like, you know, put blinders up and be sort of a one-track mind. But this is the type of thing that is a paradigm shift and it does require everyone sort of to pick their heads up and just play and learn. And you need that space to, to understand, okay, what are these things good at? What are they bad at? Where, Where is the future heading? What do you think is going to be possible, you know, 6 months from now, 12 months from now? What do you think is actually going to be not possible? It's going to be really a tough thing, like LLMs just won't solve these types of problems. Uh, and so those are the types of things that we wanted more than just a handful of us in the company to be paying attention to. We wanted everyone in the company to say, this is actually just the new way of working, um, that's going to be here.

Yeah, that's a, a bold bet. I mean, I think the last public number, $140 million revenue a couple years ago, so I'm assuming you're bigger now, but that's pretty large scale to shut the company down for a week and spend a hackathon. How do you do that? Like, if I'm another founder, what do I do?

I mean, I just said we're gonna do it. If you're the founder, you can just do it. Just say, that's— this is it, right? And if you believe it to be important, like, you go do it. And I do think the fact that, like, you know, in 12 years of doing it, we've never done that before. So it just sort of, you know, if you're— if you're doing this type of thing, like, left and right, you know, you think you're probably overusing this like superpower you had. But, uh, you know, for us, the fact that we just didn't do it that often was one of those things where people are like, oh, okay, this is serious, like we're gonna go figure it out. And yeah, there were some things that you sort of had to like design around. It's like, okay, you know, we can't stop customer service, so we do need to sort of think through like how some of those like things need to keep operating. But most of the company is, you know, you can delay what's happening for a week and It's fine. And especially if you felt like, and, and we did feel like the, the roadmaps were starting to be a little bit irrelevant, or maybe not irrelevant, we're just like, I mean, relevance too strong. It's just sort of like, there's just something much better. And so like, why don't we go start putting the things that we think are actually much better onto the roadmap?

I think that's true. I mean, you just think of the way that AI is going to change how we interact with computers. It's probably one of the biggest changes. In a while? I don't know, like since the cloud or mobile?

The way I think about it is like, yeah, I was talking to my, my mom who's been a pharmacist for, well, a long time, and you know, she, uh, you know, she, she started working in pharmacies before they had computers. Everything was analog. And, uh, there was a moment when these computers started showing up and it was like, hey, this is how we're gonna do things. And you know, in that transition period, you know, in the short term it's frustrating because she's like, I'm fast, I know how this analog system works, I'm good at this, and this computer thing is slow. Like, I don't know, like, it's, it's slowing me down, I don't know how to use it, like, it's different, it's weird, like, I don't get what's going on here. And so it sort of feels like that's what's happening a little bit with LLMs and AI inside of a lot of these companies, is there's just a shift in the way folks are, are working. And I do think that the companies that say, you know what, we're gonna stop doing analog and we're gonna shift to using computers and we're gonna take the short-term productivity hit to go learn how to do these things are gonna be the ones that benefit much faster from this paradigm shift versus the ones that are like, you know what? Like we'll figure it out along the way and like just sort of like coast our way through it. Like I, I genuinely think I would worry about those companies because I think AI is such a big modality shift that those businesses are probably at risk if that's the approach they're taking.

It kind of reminds me of when you talk about your mom's experience as a pharmacist, when Walmart was kind of expanding and growing, I think it was like controversial. Uh, Sam Walton was like, we need to put a computer in every store. And I believe they used satellites to beam like daily reports back to HQ so he could see how all the stores were performing, which today you're like, duh, why would you not have that? But back in, I forget, maybe it's probably the '70s, '80s. I'm not sure, but whenever they did this. They had an insane advantage over everyone else. As you kind of move forward, some of these things become like, you know, no-brainers, duh. But in— at that time, it was probably a huge decision. It was super expensive. I think they spent a million dollars to literally launch a satellite or something. I forget the context exactly, but it was like a pretty big deal at the time.

Now, Sam Walton, another Mizzou guy— Missouri guy. Um, I totally forgot about that.

Yeah. Um, Well, so did any new products come from the hackathon?

Uh, stay tuned. Okay. A lot of new features though. Like, you know, if you, if you log in and use Zapier today, like you can see, uh, AI capabilities that are enhancing existing stuff all over the place today. It's injected in, in a bunch of areas. Um, but net new products, stay tuned. I mean, probably the best ones to talk about are like the ChatGPT plugin. OpenAI launched GPTs and we have, uh, uh, an AI Actions that you can use alongside of GPTs. Um, so we have a lot of these capabilities that you can use, uh, inside of these. And that's probably the most notable one that, that, that is public right now.

What is that? Cause I haven't got a chance to try that yet. How does it work?

You know, you're interacting with ChatGPT in a particular way. And then if you want to retrieve information from any of the other applications you use, or if you want to perform actions in any of the other applications you use, you can, you know, talk to ChatGPT and then Zapier can help assist in those actions. So you might say. Hey, can you tell me what my calendar looks like today? And so then ChatGPT will say, yes, I can. And then Zapier will make that request and go fetch your calendar and then just get displayed. And then if you want to say, okay, actually, can you add this event at this time? Or could you reschedule this one? Zapier will then call out to the calendar and say, okay, update this event on your behalf. So it's an, it's a natural language API is basically what it is. The way it takes that natural language and then converts it again into an API request that like runs in a very deterministic way.

Do you foresee this workflow, this kind of automation with GPTs, chat-based interfaces, AI-native interfaces? Do you have any guesses or predictions on how that's going to kind of evolve over time?

I think we're going to just see a lot of innovation happening in, um, the application layer over the course of the next year. Um, you know, chat has been— as I think everybody's like gotten really excited about chat in particular, and I definitely think you're going to see a lot of products have a chat copilot type experience for every single application that you might use on the internet. You'll probably have a chat thing that's sort of helping you interact with it. But I do think that a place that is underexplored right now is non-chat interfaces. Like, chat is not, I think, the ultimate interface for every experience. There's probably some experiences that are quite incredible for, for chat. And, you know, ChatGPT is obviously one that I think we've all probably used. Found the, the benefits of it. But there's other experiences where you might say, hey, that this isn't actually the be-all and end-all. And that's the area that I think is somewhat underexplored right now. But I do think over the course of the next 12, 18 months, we're going to see a lot more founders just approaching it from totally unique and different perspectives. And, um, you know, it's, it's going to feel like what happened, I think, on the other side of the mobile era. Like, the first mobile apps sort of felt like remnants of the desktop. But, you know, as more people got to play with mobile, we, we invented new form factors. Like, the idea of, like, swiping left and right is, like, totally a mobile-first thing. And so I do think we're going to have the equivalent versions of stuff like that, um, you know, as more founders have a chance to build products on top of LLMs.

Is there anything you're kind of thinking about right now where you're like, for whatever reason, we're doing things different than what everyone else is?

Probably the, the way to answer how we go about this is one, we just like to pay a lot of attention to what are our customers' problems. Um, you know, I think there's that old saying where it's like, you know, if your customers are telling you have a problem, like they're probably right. They're telling you what the solution is, they're probably wrong. So we try and pay a lot of attention to, you know, what are the problems that they're telling us? And you know, that tells us quite a bit. We do spend a lot of time talking to, you know, certainly peer companies and folks in our space, trying to understand like how are they approaching these, these problems. We also like to get outside of that, like go look at other industries, look at other companies, look at history and see like how did those people approach similar situations, similar problems? What's the same about their situation? What's different about their situation? You know, all advice is actually just like a set of experience contextualized. And so you really have to, if you really wanna understand the advice, you have to understand What the situation of that company was. And then if your situation doesn't look like their situation, that advice probably isn't as good or as relevant in some ways. And so we spend a lot of time just going through the idea maze on this and then out the other side is that's where innovation is. That's where problem solving is. And you're just trying to say like, okay, this is what we think is interesting. This is what we think is novel. And that's probably why it looks like we're zagging is because we're just trying to do something that we think is. Helping our customers that they haven't in a way that they haven't been helped before versus just trying to do something that like somebody's already done before. They've already done it. It's like, why, why should we waste our time? They got something good. Let's go do something that we think moves the ball down the field a little bit.

Another question I really wanted to get into. You have talked about hiring a CEO coach. You said that was really helpful. The thing that I thought was really interesting that you said was you've really struggled with delegation and that was something you got better at through the CEO coach. I am also the same way. So I'm curious what you've explored, what you found works in terms of delegation.

I won't pretend that I'm an expert at this. I love being in the details. I love how understanding how things work. I think it's partly what makes me good at aspects of my job. But I do think the thing that has worked for me is to, to really spend a lot of time with the person upfront and say, hey, we are going to do a lot of time upfront. You are— I've hired you because I trust you. You are an expert. I believe in your capabilities and skills, um, so I want you here. I'm here because I've spent 10 years, 12 years building this thing, and I have a set of context and experience and ideas that are grounded in all this time I've spent with our customers. I need your expertise and experience. You need my context and understanding of customers. So we're going to spend a lot of time up front making sure that we're on the same page with those, those things, and then there's a set of shared principles by which we're optimizing for. We're optimizing for A, we're optimizing for B, we're optimizing for C. And I find that when we agree on those shared principles, often then when folks are coming to me with a potential solution or potential idea or potential way of going about it, oftentimes I'm like, awesome, sounds cool. Like that's pretty much how I would've thought about it. But if we don't agree on those principles, I'll find myself looking at it going, I don't get it. Like, I don't think this makes any sense. Something feels off. Like it just doesn't feel right. And then you're trying to go inside your head and say like, well, why doesn't this feel right? What's missing about this? And oftentimes you're trying to step back and zoom out and say like, oh, there's actually some fundamental disagreement that we have about how to go about this thing or what we're trying to do here. And it's useful to go inspect that and then codify that and say, hey, this is This is actually a core part of how we make decisions or core part about how we approach this problem. And then the more you're codifying those things, the more it helps others make decisions, you know, when you're not in the room. And so that's the thing that has worked best for me. You know, I'm still very active. I'm still very involved in a lot of these types of decisions, but a lot of times it's trying to get to the heart of like, what are the principles? What are the things we care about? And then once I feel like we're there. Then it's off to the races. Go, go, go, go, go. I feel really comfortable about delegating from that point on. But if we're not in agreement on those things, that's where I found it's really tough to delegate because someone goes off and just does a thing and you're like, you know, it's, you know, sometimes you're like, I need you to go da da da da do XYZ. And I was like, I thought you were going to go east. You went west. Now we're in a problem because like we're in two opposite directions. Um, so we got to go figure that out.

So it's basically hire someone that is very good at what you need. Spend a ton of time getting on the same page and just trust them. That's sort of where it boils down to.

But like, yeah, I think the thing that I struggled with is like you'd hire this great person and you just sort of assume like, oh, it's just going to work from day one. And you're like, I think the person assumes that too. It's like they thought, they assumed like, I'll just figure it out on day one. But that spend a lot of time together part, that actually matters. Um, and then figuring out how to codify this, write this down for everyone else in the organization to see as well too. And then you build up this like shared, you know, a lot of people call it culture, values, principles. You build up this like shared way of doing things that really helps the organization be faster on decision-making, be more aligned in what they're doing as you go. And that's sort of how you can make a 700-person organization row, row a boat sort of in the same direction.

There's this Instagram page called Traded VC. It's like this meme account basically on Instagram. It's probably one of my favorite pages on Instagram. So they have a question. The Traded VC community wants to know, do you have just a single piece of advice for founders?

I think there's probably like two things that has served Zapier really well. Like one is just doing it. Like just, there's a little bit of like Nike in here where it's like, just, just do it. Like we, you know, just get, just try, go figure it out, roll up your sleeves, make that thing happen. And probably the second thing is it's so hard to get people's attention. It's so hard to stand out these days. So be different. Like, different is good. Like, be you. It sort of works, you know, especially if you're different in the ways that the customer likes. So find a way to like do something different and novel. That generally is a pretty good formula.

Be different and just do it. Just be different.

I mean, there's— and then you got to learn from it too. Like, you know, if it's not working, you got to learn. So not everything works.

Yeah. Well, this is awesome. Thanks for taking the time to chat. This is a lot of fun. And I hope you had fun. If you don't wanna miss future episodes, subscribe to the newsletter, The Split, in the show notes. If you wanna support the show, follow, like, retweet, and send this to a friend who's still manually moving emails from their contract form into their CRM. Thanks again to Wade Foster for coming on and to Secureframe for supporting the show. I hope you learned something new. See you next time.