How a User-Driven Pivot Led to Product-Market Fit

Intro 00:01
Welcome to Startup Success, the podcast for startup founders and investors. Here, you’ll find stories of success from others in the trenches as they work to scale some of the fastest growing startups in the world, stories that will help you in your own journey. Startup Success starts now.

Kate 00:17
Welcome to Startup Success. Today, Merrill Lutsky, who is the co-founder and CEO of Graphite, is here with us, and I’m so excited to speak with him. They are in the AI software development space. They’ve raised $ 81 million. They’re definitely a startup to watch. Merrill, thanks so much for being here.

Merrill Lutsky 00:38
Thanks for having me. Kate.

Kate 00:39
Well, I’m excited to talk to you. Raising $81 million in this market is no easy feat. Congratulations. It says a lot about what you’re doing. Tell us a little bit about your story and what led to the founding of Graphite, because that’s not easy, right? To set off and start your own thing.

Merrill Lutsky 01:00
Yeah, no, I can give you a bit of the background. So I mean, personally, I grew up in the middle of nowhere Minnesota. Got my start in programming, working on robotics for Science Olympiad and like high school competitions. And one of the reasons why I ended up deciding to go to school at Harvard was in the visiting weekend there, they had this program called Hack Harvard, which was a bunch of CS and like entrepreneurial kids on campus, like all talking about startups. And that was just, like, such an engaging community that kind of, like, drew me in there and kind of knew I had to go. It didn’t take me very long to start to, like, get into that world more directly. Yeah, I ended up starting a company that went through YC. I took a year off from school to go and build that. That was with another classmate during, my sophomore to junior year. Was building software for hotels and restaurants to collect feedback from their guests. Ended up having a small exit out of that after raising a seed round, so kind of cut my teeth early on, on building a team, raising money, figuring out how to build a product. Learned a lot from that. A big lesson that I talk about is just really the importance of being passionate about the user that you’re building for and building for a customer that you care about and deeply understand. And I think, for you me and my co-founder at the time, who actually ended up going on to found another startup, and he now works at Anthropic on the team that builds a lot of their coding tools that we end up using. We work with him pretty heavily for building Graphite’s agentic code review companion Diamond. So it’s kind of fun to have that as, like, a full circle moment. But anyways, like, I think we, neither of us, were really passionate about building for hotels, and realized that pretty early on in the journey. And so I think finding a place for, you know, for that company to land, and not being the thing that we were doing for the rest of our 20s felt like the right outcome. But that was also at school when I came back. That was when I met my co-founders, Greg and Tomas, because they were two of the brightest computer science students in their year. They were two years below me, they were asking me a ton of questions about the YC experience, and I really enjoyed spending time with them. We talked a lot about, like, tech strategy. We’d sometimes go read S1 filings of companies that were going public, we really kind of formed this bond of understanding tech strategy and being passionate about the space in general. But it took a while for us to actually come back together and start Graphite. We all went to big companies. So I was at Square and then Oscar Health. Greg was at Google and Airbnb. Thomas was at Facebook here in New York. I talked him into coming to New York instead of going to the Bay. And I think that was a lot of where the genesis of Graphite came from was that we’d all seen what great developer tooling looked like at large companies that had entire teams building their internal platforms. And we got a taste of these workflows and tools that had been kind of purpose-built and crafted for working with 1000s of other engineers in these like, really fast moving projects. And then when we were outside of those walls, the juxtaposition was really apparent to us, that, like, just how good we had it inside of those companies and how much was missing to startups everywhere. And I think you know that was a lot of where the genesis of Graphite comes from. You know, in the beginning, we had that same general thesis, but for a different product. The first thing that we were building was an iOS app, like rollout and rollbacks tool. So letting native iOS app developers roll back a buggy piece of their app instantly on end-user’s devices instead of having to resubmit through the App Store and wait potentially days to get to, like, get that back. So we were building that, we were hiring our first few engineers, got to a few decently sized companies using that tool, and the first it just so happened that the first couple engineers we hired were some of Thomas’s ex Meta co-workers. They had the same experience of kind of coming in real, like really getting frustrated with the state of how code review worked on GitHub, and hacked together what became the first version of Graphite, just to kind of address some of their own frustrations with that and get back their old workflow and some of the tooling that they missed from from being at Meta. So that was sort of where the inspiration for Graphite came from, and how we got started, and from there I can tell the rest of the story, it kind of took on a life of its own, and kind of quickly became the thing we were, we were focused 100% of our time on.

Kate 05:41
There’s so many things that you’ve said that jumped out at me. First, I love how passionate you all were that you were reading filings and so, you know what drew you to Harvard was the whole startup mentality. I mean, that just shows that you have that drive, that mindset that is so important. But I like how you and your co-founders went to work with some big companies at first. I talk to a lot of founders, and I feel like that can give them a really good foundation. And then you said it kind of took on a wild ride from there. We get how you all came together. I’d love for you to walk us through that, because I read about your pivot in TechCrunch, and it’s one of the most remarkable pivots I’ve ever read about. I mean, it’s so user-driven and such a big change. I know the audience will love hearing about it from you directly.

Merrill Lutsky 06:35
Yeah, happy to talk, to talk you through it. It was really just a fascinating experience for us, and somewhat fascinating and somewhat terrifying at times experience to go through. In retrospect, it all feels like, you know, of course, this was the right decision. But at the time, we really weren’t sure. We were building this iOS rollbacks tool, and we had, we had commitments from some large companies to pay for that, and they were using it in production. It was difficult though, because we weren’t seeing, you know, we were spending a ton of our time on sales and trying to get intros through our VCs to different companies, the mobile teams of different companies, and get them excited. For one, it was very slow on the sales side, like we’d get somebody interested every now and then, but it just wasn’t – It was very hard and very grindy. It wasn’t really a problem that many companies felt like it was like a burning fire for them. And then even when we’d find one, technically, it was a tough product to build because we were, you know, we were building on top of a lot of undocumented Apple APIs. They were changing constantly. Every company uses different libraries and frameworks when they’re building their app that we had to adapt to. And then the worst part of it was that every company only releases their app every two to four weeks. So we’d basically get like, one chance to make this thing work, and if it didn’t work in that one day that the mobile team decided, okay, we’re releasing the app today, then we’d have to wait another two to four week cycle to even – we could try to fix the problem if something went wrong, but then we’d have to wait for that feedback cycle. So a lot of like factors there that were kind of like working against us. And while that was happening in the background, we hired a couple of Tomas’s ex Facebook co-workers, who came in. We’d set them up on GitHub. We teach them how to do code review here, and they had this kind of allergic reaction immediately where one of them said, like, I feel like a caveman using GitHub. That’s the level of allergy that they felt and frustration. So much so that they hacked together what became the first version of the Graphite command line interface, just to use internally and get back this workflow of stacked pull requests that they were used to from Meta. And at the time when we built it it was strictly an internal thing. It was hard coded to our repo. We didn’t really intend for this to be something that would live outside of our internal developer team. And what happened next was pretty remarkable in that at some point, word got out that we built this, and we started getting inbound emails and requests from other ex-Meta engineers that were like, Hey, I heard that you built this thing. I heard you built a stack tool, like, can we use it? Can we get it at our company? Like, can you guys make this available? Can you focus your company on this, please? And for a while we’re like, okay, like this doesn’t make any sense. Like, why are we? Like, no, we’re building, we’re building, like, mobile rollbacks. That’s the whole focus and but this kept happening, and it would just be, like, week after week, we’d just get there’d be more and more noise about it. And once or twice there were even demos where we’d be pitching, we would be speaking to, like a mobile team, and we’d be pitching the iOS rollbacks tool, but an ex-Meta engineer would come on the other side and be like, you know, okay, okay, mobile rollbacks is cool, but, you know, tell me about the stack diffs thing. Like, I heard that you guys have built this tool. Like, I really miss Fabricator from Facebook. Like, will you give me this? Like, this is the thing that I’m really here for. And at some point, it just became kind of too much for us to ignore. But at the same time, we weren’t sure that there was that much of a business there. We weren’t sure there’s not much of a need. And, we had something that was going. We were also, we only had about a year’s worth of runway left at the time, so we weren’t really sure, you know, we were like, Okay, we have something that’s kind of working. Do we keep going that path or, you know, or do we kind of take a flyer on this thing that everyone seems excited about, but, you know, we’re not sure how much of a traction that we’ll be able to actually unlock there? So basically, what we did was we designed a test for it where we said, let’s give this a try. We’re going to work on this full time for one month and if we by the end of that month can get to 20 engineers at companies, like meaningful companies that we respect, using this every day at work, then we’ll consider it good enough to kind of commit all of our resources towards this, assuming that we don’t see anything major happening on on the mobile rollbacks product in that time. And yeah, so we did. We made it so that it wasn’t hard coded to our repo, and we gave it out to a first set of users that had asked us for it, and what we found was that that 20 quickly became 50, and then the next month, it became 100 and even without us really pushing for it, this product just spread like wildfire in some of these organizations. And the thing that struck us was it wasn’t just random Indie devs using this. It was like, you know, engineers at places like Ramp and Snowflake and Datadog and, like, some of these, like really massive companies that we hadn’t even been in the conversation with on our other products. So now all of a sudden, these engineers are using this, and we also, we put them all on a Slack channel for a while, it was just a Slack connect channel, until we ran out of, like, you can only get up like 50 companies on Slack connect, and then you had to go and make a whole community for it, because we couldn’t keep it all in one. We basically, we put everyone in there, we’d get really high quality feedback all the time. The blessing and curse of building for developers is that whenever there’s something wrong, they tell you exactly what is wrong with it, exactly how to fix it, exactly what they want from the product. And, yeah, okay, quickly, just on level of engagement, we also made everybody we weren’t charging for the product, but we had every early user commit to doing a half hour Zoom call with us every week, just to make sure that there is like validation of them taking it seriously. And, you know, everyone would do that, and they would show up, and they’d give us great feedback. And I think seeing that, that level of enthusiasm and engagement and spread made the decision pretty clear for us after, you know, after that first month.

Kate 12:49
That is truly such a wild story. I mean, that is explosive. It really is. Talk about incredible word of mouth. Was it hard at all to walk away from the original concept?

Merrill Lutsky 13:06
Yeah. I mean, there’s certainly, it’s funny, sometimes I’ll have conversations with – one kind of interesting thing that we’ve seen is almost all of the initial customers for our old product have now become customers of Graphite, which is kind of fun to see. Every, every now and then I will be talking to like a mobile engineer or mobile lead, and I’ll tell them about the story, and then they’ll say, like, wait a second, that’s actually something that we just think we could use. It’s hard for a startup to build, and it’s something that, you know, could we have built a business there? Perhaps, but I think, I think we found something that was both like, much more of a burning need, something that we were much more excited about. And I think that has just so much more of a connection to now with all the changes to the software development life cycle, with AI like just connect so much more to that world, and will play such a huge role in defining what software development looks like for the future, which is pretty exciting to get to work on.

Kate 14:07
Yeah, that is super exciting. And what a great community to be a part of. Of, right? Like you said, it’s the way you get feedback, and the, you know, just the kind of stories of what you see your customers doing with your product that must be super inspiring, right? To see some of, yeah, some of these large scale implementations.

Merrill Lutsky 14:28
It’s the large scale ones. It’s also, you know, sometimes there’s, there’s some companies which are just doing something that’s, like, really, really interesting with Graphite, or really meaningful. Like, yeah, there’s a kind of a technological art studio called Breakfast Studios in Brooklyn that makes these, like, interesting, interactive, kinetic sculptures, that’s using Graphite. There’s a company called Proxima Fusion that’s trying, that’s literally trying, to build next generation fusion reactors. And their engineering team is building it with Graphite. And to be able to, like, support teams like that, that are that are doing something. It is also, you know, then certainly, there’s like the Shopifys of the world, you know, companies that are, you know, 1000s of engineers building some of the some of those visited properties on the web, like, it’s a whole range of like, interesting and important companies and individuals that we get to support now.

Kate 15:23
Yeah,I can see where that would be pretty incredible. I find that inspiring too. Just in talking to the founders that come on this show, what everyone’s doing right to be a part of some of those stories, I’m sure you know, brings a lot to the team. Is it the same founders that started on the original idea that are still leading Graphic?

Merrill Lutsky 15:43
Yeah, yeah. So all three, all three of us from the very beginning, from March 2020, when we incorporated went out to fundraise, like the week before COVID. COVID shut everything down. Until now, it’s, you know, the three of us are still going strong.

Kate 16:00
That’s incredible. Can you share with us a little bit about Graphite today? You just shared some great examples of customers, but just kind of the benefits that you bring around development.

Merrill Lutsky 16:12
Yeah, yeah. So Graphite, we think of as the code review platform built for the age of AI. So we take both like the best in class workflows and tooling that we saw at companies like Meta and Google around code review, so helping engineers create review and merge pull requests faster, and then now combining that with LLMs to help further accelerate that process, scanning, giving engineers feedback on every code change that they’re writing in seconds versus hours or days potentially. The net effect is like companies are shipping 30, 40% faster now with tools like Graphite, with that whole tool chain, than they were before using us.

Kate 16:55
That is incredible. I mean, to move that fast, I had some experience in the software development industry, like 20 years ago, and it was kind of a slow pace, right?

Merrill Lutsky 17:05
That was like the Waterfall era: hand it off, wait for QA. Now this world of, you know, we’ve had this, this shift from that world to the world of, like, you know, continuous integration and deployment. And now we have with, with agentic dev, it’s like even faster, even more like constant 24/7, you can have, you can have, like, agents working on your product and iterating on it.

Kate 17:30
That it’s so incredible and there’s so much positive around it. In terms of AI, though, in general development, do you mostly see positives? Do the positives outweigh the negatives? Are there negatives? Like, how do you feel about it?

Merrill Lutsky 17:47
It’s a great question.I think there’s certainly, there are certainly challenges that it poses, and it’s not, you know, I think, I think it has enormous potential, and a lot of teams are already unlocking a ton of productivity from tools like Cursor and Windsurf and even some of the the more recent agentic dev tools like Devin, Codex just released, Google just released tools. There’s so many of them now that teams can play around with and use. I think the challenge of it, and a lot of what’s a lot of what’s actually driving interest in Graphite, is that now you have these great tools for generating code, and engineers can build features faster than ever. They can put up more PRs, but that’s only half of the process. I think anyone who’s worked at a large company knows that that’s not where the story ends, and that’s, that’s just the first step in it. After that, you have to go through review and testing and deployment, and that outer loop, or like that second half of the development cycle hasn’t, hasn’t really changed that much in the past 10 years, like since GitHub invented the pull request. And that’s really, I think, where there’s so much opportunity for us, and what we where we see building Graphite to be is like that outer loop complement to these agent tech dev tools, where, if you have vibe coding needs vibe review, we like to say, but you kind of need a compliment to to accelerate the rest of the process and make sure you keep up. Because otherwise, and what a lot of teams are seeing now is like they have just a higher volume of code changes that they have to get through than ever before. They also have, in many cases, more bugs, like inefficient code security vulnerabilities potentially. It’s not, you know, I think that it’s not enough to just hand a task to an agent and say, go and do it. I think you really need, like, a team of agents that are focused on different things in order to achieve great results. Like, much like you need, why do you have code review in the first place? Having a check, like having somebody else and other folks on your team that are, you know, that are balancing that out, I think helps achieve much greater code quality and better products. And I think the same is true with agentic Dev. So if anything, I think that’s, yeah, that’s really the where, where people are seeing challenges with it. It’s more on, like just letting it run wild and then having trouble dealing with both, like the volume and the potentially like buggy or insecure, non performant nature of the code that gets generated without the help of tools like Graphite.

Kate 20:15
Wow. You sum that up really well. I think you’re spot on in that analysis. Developers I speak to, that’s exactly what they complain about in this day and age, right? It’s just sometimes out of hand there.

Merrill Lutsky 20:31
Yeah, it can be overwhelming. And I think that even the pace at which these tools are adapting and the amount of code and the pace at which they’re able to work now like we just have to the rest of the process has to evolve too in order to keep up.

Kate 20:45
Yes. So everyone listening, founders, it’s the rest of the process that needs help. Remember that. We’re coming up on time. What a fascinating story. I always end the show with just advice for other founders, listening. Anything you know that you could share just for those that are new, particularly, probably in the earlier stages of the journey.

Merrill Lutsky 21:06
Yeah, I think that the, I mean, one of the biggest things for us has just been like having a really close feedback loop with our users, and really fostering that. And we’ve tried to scale that even now that we’re serving company enterprises with 1000s or hundreds of developers at a time, like really putting time into building those relationships, like having direct Slack channels. Like being able to text some of the end leaders at Shopify, for instance, is, you know, that helps so much in terms of, like, being able to like stay on top of everything, and make sure that we’re building tools to like, help them do their jobs better. So I think, like, really fostering those close user relationships, and investing in that from early on. And, you know, making sure that you’re building, doing what we did in the beginning, of like, getting some commitment of time and really making sure that you’re building something that users are willing to invest in, be it with like time, money or social capital, or something else. Having that, I think that is really important for making sure that you’re building something that they’ll ultimately care about and value. The other piece, I’d say, is like picking an industry that’s constantly growing, picking a customer base that’s growing, you know, where you have certain tailwinds that will just make your market bigger over time. I mean, we, we kind of got lucky in that we were solving a problem that we knew was going to get bigger with code review, even before the AI revolution. And now that’s kind of come along and made that 10 times worse. So finding a tailwind like that, where you know there will always be, more software companies, more software being built, more developers, more compute. Being part of, like, a constantly growing market both increases your increases your TAM as you grow, and also just gives you, makes what you’re doing more important, even if you do nothing else, like just with the passage of time, as these background trends happen, like, your work gets more important and the importance of what you’re doing is amplified.

Kate 23:19
I love it. Excellent advice. And your first piece of advice resonated with me about your feedback loops and being close to your users, because that probably helped when you made that pivot, right? It probably gave you that assurance, right?

Merrill Lutsky 23:35
Yes, it’s fueling like, and product-market-fit is like, one of those things where it’s like, you know you it’s hard to describe, but you know it when you have it, and you can really feel like it’s a different level. It was a totally different level of excitement and enthusiasm for Graphite, versus any of the, any of the things that we built before. And I think that having like, yeah, like you can be close to that feeling of like, know, when you get to that, I think that’s, that’s like, super important to to like, be able to just to be able to like, feel when you have that.

Kate 24:03
I think that sounds spot on. And seriously, some of the most successful founders I know, they echo that same thing that you can like, feel product-market-fit. Yeah, that’s awesome. Thank you so much for being here today, where can I find more information about Graphite?

Merrill Lutsky 24:18
You can check out Graphite dot Dev is for our main pull request platform, and then Diamond dot Graphite dot Dev is in particular, is our AI code review agent. You can use both products. You can buy one at a time. A lot of teams are really excited about just adding Diamond to their GitHub repos, having it start to review their PRs in a few seconds, and you know, just carrying on from there.

Kate 24:45
Awesome. I love it. Thanks so much for being here. Merrill. What a story I’ll be interested to see where you guys head next. Thank you.

Merrill Lutsky 24:51
Thanks so much.

Outro 24:52
You’ve been listening to Startup Success to make sure you don’t miss out on future episodes, subscribe to the show in your favorite podcast player. Like what you hear? Tap the number of stars you think the show deserves in Apple Podcasts. For more tools and resources for your own startup success, check out burklandassociates.com Thank you so much for listening. Until next time.