Tony Holdstock-Brown & Dan Farrelly, Inngest Co-founders, on Embracing Feedback from Early Customers
29:17 Share

There's just such a higher bar for dev tools in terms of uptime, reliability. You just have to be better. We're quite fortunate that all of these technologies have emerged and matured in the last five years, because without that, building something like Ingest that layers a really good platform for building workflows on top of these would be really difficult.

From GGV, this is Founder Real Talk, where we get real about the challenges that founders and startup executives face and how they've grown from tough experiences. I'm your host, Glenn Solomon. Without further ado, here's today's episode. On today's episode of Founder Real Talk, I'm joined by my colleague at GGV, Dan Kahana.

who's joining from Israel. Welcome, Dan. Great to be here, Glenn. We're both really excited to welcome the co-founders of Ingest, a startup that's building a serverless, event-driven workflow platform. Today, we have both founders, Tony Holdstock-Brown, who's the founder and CEO, and Dan Farrelly, who's his co-founder and CTO at Ingest.

Here at GGV Capital, we're also proud to announce that we're leading a $3 million seed investment in Ingest. We believe there's tremendous potential for Tony, Dan, and the Ingest team to help every developer build better apps with simpler, more intuitive tools. And Dan Kahana and I are so excited to have Tony and Dan F. walk us through these early days of building Ingest and share what's on the horizon and keep things very complicated since we have two Dans.

Tony, Dan F., welcome to Founder Real Talk. It's great to have you. Yeah, thanks for having us and excited to be here. So Tony, we thought we'd start with you. Maybe take us back to your early days, growing up in England. How'd you end up in the Bay Area? And did you always know you wanted to be an engineer and wanted to be a founder? Yeah, yeah, yeah. The Bay is kind of serendipitous, actually. I didn't plan to be in the Bay. It just kind of happened. A series of events lined up for me to move out here.

But I always knew I wanted to be an engineer. I learned how to program when I was a little kid and I used to solve my math homework by writing programs. So I wasn't really learning linear algebra. I'd write math programs to solve it for me. And I think that was when I knew I wanted to code for the rest of my life. So I've been doing it for almost as long as I can remember.

Back in the day, I was working on the side when I was a teenager on different side projects. I got a job on the side doing this pretty young. And I ended up building something like Spotify for a record label back in like 2007, 2008. Back when there were like no frameworks for single page apps. And that was super fun. Really enjoyed it. And then fast forward...

Docker was a thing in 2015. My best friend from England had just moved to San Francisco working for KPMG and I'd met Steve Francia, SPF 13, kind of a prolific code developer, an event. And I was talking about how awesome Docker was and that it would be fun to work there. And so he interviewed me with a bunch of other people from Docker, ended up offering me a job at

relocated me to San Francisco on basically the same week that my best friend from England was moving to San Francisco. So he sort of spurred the move to the Bay, but regardless, I was always working in tech. So yeah.

So yeah, kind of serendipitous. It worked out well. I like it here. Your best friend story. Okay. Yeah, yeah, yeah. And Dan, you were most recently the CTO at Buffer, but years before that, you had briefly co-founded a live event and services company based in Brooklyn. You know, what was your first founding experience like? And then tell us a little bit of how you and Tony hooked up. That's a throwback for sure. I much prefer the software business to live music. It's really hard.

But I love the experience. We were a super scrappy combo, my friend Jake and I, and we're just constantly trying to make things happen. And I think one of the things that we learned that was really fun that I took from that was just trying to keep customers happy. One venue, I remember they needed a stage. So I borrowed some tools and I built a stage.

So then that same venue turned around and said, "Hey, well, we need a website for events." And I said, "Okay, yeah, we can do that." And we walked out of the venue and Jake asked me, "Do we know how to do that?" And I said, "Well, we'll find out. We'll figure it out in two weeks."

And so I figured out how to build them an events website. And funny enough, that was the first time that anyone ever paid me to write code. So this is kind of like a fun experience that ended up putting me in a different direction career wise, but still love doing that. It was a lot of fun. And, you know, I think probably two, three, four years after that, that was when Tony and I met. He had just moved, I think, from England to New York and

And him and I were both working out of this independent co-working space in the Lower East Side. And at the time, I did not know. I was still relatively new to the industry. I know no other developers. And I met Tony, who's just right next to me, a friendly guy, chatted with him. And we just kind of hit it off from there, became friends and grabbing beers and whatnot, talking shop about the...

about the industry. So yeah, that was how him and I met back then. And then, you know, as you guys have worked on Ingest together, how do you work together? How do you figure out who does what? How have you decided to split up the work? At the start, it's like an all hands on deck type deal. You know, you've got this idea, you conceptualize it and you have an approach, but you're still looking for product market fit. You're still looking for the best way to sort of

help your users understand what you're building. So at the start, we were splitting things up based off of ideas that would come up with weekly. How can we get this done the fastest? Who's going to do what? So it was based off of speed to begin with. I think now things got a little bit more stable. We have a team that we can depend on. We can delegate a lot more. Things like growth, marketing, product, engineering, longer term planning. We know a lot more than what we need to do and we have resources. So we kind of

take ownership in areas that we have a little bit more affinity towards and kind of groove with our skills right now. But, you know, until we are able to hire and delegate those things out, but it just helps us make us even faster. So we're always trying to optimize for speed in that sense. We both sync on product and priorities weekly.

We have ownership over specific areas, some delegated to the team, but the key focus is really how can we build fast and how can we make our customers happy? So splitting based off of natural affinity and where we can work fastest is still kind of a thing too. Tony, on your LinkedIn profile, you have a recommendation from someone quite a while ago that says you're a born developer and the more technical the issue, the more you enjoy it. Do you

He later speculates that you might develop Facebook too in the future, which as far as I know, you're not working on Facebook too yet, but maybe walk us through kind of the problems you and Dan both saw in your previous roles and how you're looking to solve those problems with Ingest. Paul was a great person. Previously, prior to Ingest, I was running engineering at a healthcare company out of San Francisco, building a vertically integrated Invisalign competitor.

We had to build medical record treatment planning patient apps from scratch. Thesis was with engineering, we could make orthodontic treatment much faster, much cheaper and much more accurate. So it really relied on how well we could build our software to operate the clinics.

And in doing so, we had really complex workflows that we had to build for every aspect of the company. It could be something as basic as an appointment, which when you get into things is really nuanced. Patient schedules and appointment, we have to send them intake forms. If they don't complete the intake forms within a specific amount of time, do X, Y, and Z. And if they don't give us authorization to hold their medical record data, we have to do a certain set of remediations.

Really complex for the engineers to build. Event-driven stuff is a natural way to build this because it models what happens in real life. But at the same time, it's really complex for a team to get behind because you have to spend a lot of time learning, a lot of time building out the platform on top of the bare bones infrastructure.

So honestly, we spent a lot of time reinventing the same stuff that every other engineering team has done when they build out queues or event-driven architectures. Really difficult. Ultimately, the code was messy. The developer experience wasn't necessarily the nicest, but it worked. And that was the inception of Ingest, taking a step back and thinking,

how can every developer build something like this, which is necessary everywhere, in a really easy way? Because the healthcare flows, while extreme, they're just a more extreme version of flows that you have in every business. And after speaking with a whole bunch of engineers and all my friends, everyone faced the same issues. That's super interesting. Obviously, software development has changed a lot over the past 5, 10, 15, 20 years. I'm curious, what recent changes and innovations do you think

make this a particularly interesting time for you guys to be building ingest. It's interesting because there's so much new and so much old. So like, actually, if you look at SQS, amazing technical preview in 2004, hasn't really changed too much in 20 years, which is almost crazy. Whereas you've got platforms like Vercel, super new, amazing developer experience. Everyone from bootcamp grads to experienced engineers can deploy easily on Vercel.

So there's this mixture of new technology for normal engineers that want to ship fast, like Vercel. And then there's this legacy stuff for building queues and complex apps, which hasn't really improved in the last 20 years. So there's this gap in which if you're building something more complex than a set of APIs, if you're building these complex workflows that we had at Uniform, that our customers have at Ingest, right?

How can you do that using a modern platform? You kind of have to go back 10, 15 years and use stuff that's been around for quite some time. It's quite difficult. So the rise of serverless, the rise of like event-driven plumbing and the rise of, you know, modern databases like Clickhouse, OLAP stores mean that we can combine several technologies in a really novel way.

and then give users this really, really clean API that takes care of everything you'd normally have to build yourself.

So I think like we're quite fortunate that all of these technologies have emerged and matured in the last five years, because without that building something like ingest that layers a really good platform for building workflows on top of these would be really difficult. As a seed stage startup, you're running really fast. And Dan, you mentioned you've hired some people, but you're still quite a lean team. And so, you know, all the feedback you get from the market's really important.

you now have some customers and you're getting feedback from those early customers. Anything you're picking up that is unique and surprising, things you weren't expecting,

feedback you weren't expecting or customers using the product in ways you weren't expecting? And how's that impacting how you're thinking about building going forward? I think it starts off. And one of the always fun, surprising things, and this might be applicable across different types of products, but it's fun and surprising to see new companies adopting your product even early when it doesn't do all the things that you might need to do. If they find something that they love, they will deal with the friction.

And they will help you solve the problems. And that is just an actual fountain of information for us. If you can make it simpler for them and help them solve complex problems, they're going to love you. And...

That's something that we've learned from tons of customers, gotten amazing feedback and built some really great evangelists within companies. And I think mostly people don't care how you work, especially in these platforms. They might be curious about the engineering internals, but they're

They're trying to solve a problem at the end of the day. So that's always really, really great. It's like if you're solving something for them, they don't necessarily care how the insides of your product work. And I think another thing is just focusing on

They're time to value because if you can help them ship something faster, if you can get them to that aha moment faster, that means that you're going to build those long-term happy customers. And I think some fun things, there's always something wacky that comes through that someone wants to do. And some people are doing some crazy things. We thought of mostly basic workflows and whatnot, but...

There's some people that are coordinating like fleets of VMs on fly.io. And one other security company is also using us to spawn like millions of VMs and fly and run security analysis. So it's funny. We didn't think that people would use the workflows for infrastructure management because

but people are, right? So there's always new ways that people are surprising us. I'd love to understand a bit about, you know, now you guys have raised your seed rounds, what your seed round, what sorts of developers are you guys looking to reach and,

What channels have you found are interesting ways to reach them? So broadly, we speak with engineers that really just want to build and ship things faster with minimal overhead. So technical leaders who want their team to work faster, who want the ability for every engineer to work on workflows instead of

background, like a, like backend engineers. Those are the types of engineers that adopt us right now. We're focusing on the TypeScript ecosystem. JavaScript kind of eats all the things is super common for everyone to pick up. It's got the broadest appeal. More SDKs are coming soon. We have a go SDK and an Elixir SDK that we've got in preview. But realistically, if you're building an app, you've got node, you've got a queue or you need a queue.

Those are the engineers that adopt us and we help them build much faster. In terms of discovery, there's a few different things that are working out for us. Ecosystems and partnerships are really good because talking about serverless, it's been really difficult for you to build anything reliable or anything that's not an API in a serverless fashion without Ingest. You adopt Ingest, you drop in RSDK, you deploy to your existing provider like Percel and

And then you get everything within the sales functions. So that's great as is content marketing and word of mouth is actually surprisingly good for us too. So engineers really, really appreciate Ingest when they get to use us because we give them a really good DX and tools that they would never have the opportunity to use. Once they start using us, they really appreciate us. And we've seen

thankfully and we're really grateful a bunch of people on twitter talk about how they're using ingest talk about the things they've used whether it's ingest in ai or ingest in starter kits or ingest for video sort of workflows a bunch of people talking about us and recommending us to other engineers which is yeah really really really good but um engineers are a tricky bunch we don't like being talked to we like to figure things out ourself and we uh

We like to exercise critical judgment. So for us, it's about writing things that engineers find helpful without necessarily, you know, espousing the ingest way. And I think that people will see the benefits regardless. It's a line to balance for sure. It must be super gratifying to see people adopting and

getting value out of ingest and then tweeting about it or getting word of mouth going. That must be really fun. A side benefit of that would be

I assume, you know, folks are excited about your platform, getting value out of it, might want to come work with you. And, you know, Dan mentioned earlier, you guys are starting to build out the team. What are you looking for when you add people to ingest in the hires you've made and the hires you intend to make? Maybe you could talk a little bit about like the type of person you're trying to hire, you

the roles you want to fill and curious how you're trying to build a team that stays lean and mean, but also can get a lot done. You know, overall, we have a lean towards senior experience folks right now. You know, we need to operate highly efficiently and as quick as possible, which means people need to operate, like be able to run with,

the minimal context or be able to absorb the context themselves, right? They need to put themselves in the shoes of the user. And that comes from also having been a developer who has built systems like that. So we have to continuously try to make complex things simple. So what's the best thing? We try to find developers who have built these systems, who have maybe even led teams that have dealt with these problems so that they can

understand how do I enable more and more developers? So we have ex-senior leadership, ex-managers. We have even some people who came out of our community. They come right out of the user base. They found us on Twitter and they love what they're doing. They've built systems, open source even, that are similar to Ingest years ago.

And they said, you know, this didn't work, but I see this in your system, what I always wanted to build. And those people come in because they have the context. They understand the problem that we're solving. They understand our target customer. They understand how we work. And that allows us to move even faster there and just kind of run at kind of light speed when it comes to development. So ramp them up in much less time than it might take otherwise. That's awesome. I think it's always...

Cool to see how when you're building for developers, you can get folks super excited and ultimately bring those folks onto your side as well. We've also only hired engineers, even our designer is an engineer, which is awesome. There's a lot of context like Dan was saying. I think like for us, because we're building an engineering product, focusing on engineers is really, really key.

So across the board, it's been engineers, design and dev role now as well. We always see working with companies that there's challenges you can anticipate and challenges you can't anticipate. And, you know, being a co-founder, being a CEO, being a CTO, these are jobs that at a startup of all every few months. So I'm curious, you know, what challenges have you guys seen that have surprised you since you started Ingest? So there's the obvious one, prioritization. Prioritization is really, really, really hard.

If you are running a company, it feels like everything is screaming for your attention. Really, there's a few things that you should be working on that are top priority and figuring out what's top priority, what's highest leverage is sometimes difficult. You can kind of get lost in the weeds. So that's quite difficult. And on top of that, it's really easy to have a tendency to stick with what you're good at. So a lot of engineers will stick with engineering.

designers might stick with design. Fighting that tendency to always focus on the highest leverage item is really, really important. I think like there's also one thing that I've talked to a few founders about, which is really interesting. And it's especially common from my conversations, at least that I've seen for us engineers who move into leading a company to sort of have this issue with delayed gratification.

So when you're an engineer, you can write code, you can ship it, you can get the PR out and you can build your systems really easily. And that green check mark, that merge button. Awesome. It feels really good. If you're running a company, the gratification is a lot slower. You know, like sales processes might take six weeks. It might take 12 weeks.

And figuring out how you can get gratification so that you can kind of keep that response going, how you can shorten that path is honestly really hard, but really good, especially if you're used to such a fast feedback loop. I think that's actually sort of something that's underappreciated and can lead to, you know, like...

I don't necessarily think it leads anywhere bad, but you have to be aware of it. For me, I think, you know, everything that Tony shared resonates so much. And another thing that I have high respect for is coming from like B2B SaaS, there's just such a higher bar for dev tools in terms of uptime, reliability. You just have to be better. You have to have better uptime. And because these other businesses are building on top heaters. So...

You can't cut corners. You can't change APIs so quickly. You can't ship changes to UI so quickly because they're things that people are doing every day in their workflow and delivering products and they depend on you. So there's just a higher bar. And I think the more and more I work on it, the more and more I have respect for and empathy for our own customers and what they need to achieve. So that's just like something that just keeps building over time for me.

It's great to see you guys pouring your heart and soul into the company. And clearly, this is a mission for both of you. We're really excited to see where it goes.

And speaking of seeing where it goes, tell us a little bit about your visions for Ingest, you know, a year out and maybe even a little bit longer term, three to five years out. One year out, we want to become the de facto solution for running workflows and backgrounds. So we have this TypeScript SDK. We want it to be easy for every engineer to adopt and to integrate and run whatever workflows you need, whether it's on the server, serverless or edge.

With the other SDKs coming, Go, Python, Elixir, we want broad adoption across all engineers so that everyone can build fast, no matter your skill level or what you need to build. In terms of five years out, we want you to be able to easily leverage the events that are running through Ingest. So as soon as you adopt Ingest, your engineers are sending events over to our system.

you've got this stream of high quality data, we want it to be really easy for you to be able to click a button and get things like data warehousing, real-time ETL, because you can guarantee that this event data is high quality. If you've got these events coming from Ingest, it's great to be able to take these and work with them immediately. Or even things like Descope, if you've got your signup events, you can put that in a data warehouse and work with that immediately.

So Dan and I have been very well behaved and haven't yet mentioned Gen AI once until now on this episode. And we're VCs. So, you know, we have to ask, how is Gen AI going to impact your business? Is it a good thing, a bad thing, a risk, an opportunity? How do you see it? Definite opportunity here. We have a number of users already using Gen AI with Ingest to orchestrate all of their jobs. These jobs take time.

You know, they don't finish quickly. There's a lot of data that needs to be processed in a pipeline. Often you need to chunk a lot of your requests. You need to manage state. You need to chain jobs into each other to get quality answers out of this. And a lot of that is really quite difficult and to do reliably, you know?

Luckily, we offer that out of the box. It's what Ingest does at its core. And a lot of these products are able to iterate a lot faster because they don't need to worry about all this stuff. They don't need to manage this. And there's never been a simple API for building distributed systems for managing all this.

And you can use an LLM to scaffold these background jobs and design workflows. So there's a lot that we can do. OpenAI just released their new functions in their API, which is pretty exciting because you can kind of like plug anything you want into there and have it feedback. So it kind of, those are all just little ingest functions or little ingest parts of your workflow. So you can kind of use OpenAI to power your entire workflow.

ingest function and it can run on forever and just continue to evolve as, as it runs.

So I think there's just a lot of opportunity here. And honestly, in the last few days, Tony and I have been hacking on things just to see like, what can we do with this? There's so much potential. So coming soon to ingest is the ability to invoke open AI as part of my background job. Very cool. All right, guys, we're at that time of the episode where we're going to put you on the hot seat and it's the speed round. So,

Just answer the first thing that comes to your mind and maybe, you know, you can race between each other to see which one of you will answer these questions. I'll start. Is there a book or an article that, you know, you've read where you'd recommend to other founders? There's a book that I really, really love. Two books. First, Principles by Ray Dalio. Not just for building a company, but for life. It's really, really good. And then Traction.

by the founder of DuckDuckGo, Gabriel. Amazing book. So good for, you know, growth, traction. Okay, that's a new one for us. Traction. Okay, we'll have to...

I had two quick hitters. Empowered by Marty Kagan. It's just a really good book about influential, about how do I think about empowering teams across organizations and thinking about product and related to technology. And one of their, I'll throw an article in there. This is from years ago now, but the AWS CTO, Werner Vogels, wrote about modern applications at AWS. And it really was influential for me back then about how I think about like the push towards serverless and managed services. And it influenced just kind of how I thought about all that stuff. So that's just kind of like

more on the nose for what we're building and has resonated with me for years. What quick advice would you guys give to a younger Tony or a younger Dan? Leverage, focus on what matters. You know, I think like a lot of time spent could have been used more efficiently. Maybe I could have like learned what to do a little faster instead of building the healthcare company for five years. Maybe I could have

out there in three, I don't know, and did ingest sooner. Leverage is really key and figuring out how to work on the most important things is not just for a start, but across life important if you've got ambition because time just disappears. For me, it's just connecting with more people in the industry. I think I was pretty heads down on product and leading a technical team for years. And honestly, we just benefit so much from

from chatting with people in the industry, connecting with more people. It accelerates our learning and

you know, I just always, I should have spent more time on this. It's just, it's unbelievable. And that's what helps take our product to the next level. Awesome. All right. One more quick one for both of you. Tony, what do you miss most about England? The sunshine and the world renowned foods. Genuinely, I think family and friends. My brother has a little two-year-old son.

And it's kind of sad watching them grow up from afar, you know, so I go back there when I can, but it's not often enough. And seeing them develop is special. And being away from that is particularly sad sometimes. Right. Well, we'll have to make sure you get some customers over. Yeah. Dan, going back to the music theme.

Best live show you've ever seen and who's on your wish list, who you want to go see next? This is hard. I probably could go through 10 of them, but I'm going to go tight and say the LCD Sound System Reunion show back in, I think, early 2016. I think it was at Webster Hall in New York.

It was just like the band that I always listened to, like define my years living in New York City. And it was just the best thing. I was like right at the gate, right in the front of the show. It was amazing. And Bucket List, oddly enough, I think it might be Bruce Springsteen. Like it sounds so easy and generic, but like he's just a legend. And I mean...

- I don't know. That's my bucket list. - The boss, hard to beat. - Oh, so good. - Guys, you both are Bruce Springsteen-esque to us. You're doing amazing work. You were born to run and we're just glad to be part of the ride. Thanks so much for joining us today. And again, congrats on raising your seed round and launching to the world. We just couldn't be more thrilled to be along for the ride and really looking forward to seeing where it goes.

Yeah. Thank you, Glenn, Dan. Thanks, guys. You've been listening to Founder Real Talk. If you like what you heard, please rate and review us on the Apple Podcast app to help others find this podcast. If you have any questions you'd like us to ask our guests or founders you'd like to hear on this podcast, feel free to email us at founderrealtalkatggvc.com. Our theme song is by Grapes.

GGV Capital is a global venture capital firm that invests in local founders. As a multi-stage, sector-focused firm, GGV focuses on seed to growth across consumer, social and internet, enterprise cloud, and frontier tech. The firm was founded in 2000 and manages $6.2 billion in capital across 13 funds.

Past and present portfolio companies include the likes of Affirm, Airbnb, Alibaba, Didi, Grab, Hello Bike, HashiCorp, Houzz, Keap, Namely, New, Opendoor, Peloton, Poshmark, Slack,

Square, Wish, and many more. The firm has offices in Beijing, San Francisco, Shanghai, and Silicon Valley. Learn more at ggvc.com or follow us on Twitter at at GGV Capital or GGV Capital on WeChat.