He killed a $100K ARR product & pivoted—then raised $375M. | Viraj Parekh, Co-Founder of Astronomer

They were building a Segment competitor. It was working—customers were paying. But every sales call, prospects kept asking about the backend tech instead of the product. 

So they killed the roadmap and pivoted. It took them 18 months to hit $1M ARR. Then they started growing. And so far, they've raised $350M. 

Viraj walks through exactly how he validated the pivot, landed the first 10 customers, and why being outside Silicon Valley forced him to show more traction than everyone else.

Why You Should Listen

• How to know when your side feature is actually your real product
• The exact question to ask prospects to validate willingness to pay
• Why getting to $1M ARR slowly can set you up to scale faster
• How to compete when you're not based in Silicon Valley
• What talking to your first customer 4x a day for 2 months teaches you

Keywords

startup podcast, startup podcast for founders, open source startup, B2B SaaS growth, pivot strategy, developer tools startup, finding product market fit, early stage fundraising, design partners, commercial open source

00:00:00 Intro

00:01:46 Getting caught at the Coldplay concert

00:14:29 Deciding to Pivot From a Working Product to Something New

00:17:27 Building a Business Around Open Source Technology

00:19:38 Selling Before You Build

00:27:37 Talking to the First Customer Four Times a Day

00:30:51 Landing the First Ten Customers

00:35:10 Fundraising Without Silicon Valley Pedigree

00:38:48 When He Knew He Had Product Market Fit

Retry

Chapters

• 0:00 Intro
• 1:48 Getting caught at the Coldplay Concert
• 14:29 Deciding to Pivot From a Working Product to Something New
• 17:27 Building a Business Around Open Source Technology
• 19:39 Selling Before You Build
• 27:37 Talking to the First Customer Four Times a Day
• 30:51 Landing the First Ten Customers
• 35:10 Fundraising Without Silicon Valley Pedigree
• 38:48 When He Knew He Had Product Market Fit

Transcript

Viraj Parekh (00:00:00) :
Oh, Airflow, that's really cool. Let's talk more about that, you know, I saw Airbnb's using it. I'm seeing all sorts of Silicon Valley unicorns adopt it. All these great blog posts about it. My data engineering team's talking about it and this funny thing started to happen. Where we tell customers what our product was powered by and they'd be more interested in that than the product itself. You could have every internal conversation that you want. You can kind of, you can look at studies about market size and, so on and so forth. But when people who are willing to pay say, I will pay you if you can give me an easy button for this thing. That's pretty much all the evidence you need. I don't know of anyone whose first handful of customers hasn't been like a bit of a knife fight, in the sense that you're just getting on the phone with them every day. Trying to make sure what they're doing works and doesn't work. Our first customer that paid us for managed Airflow was a data scientist. I think I had talked to him, no less than four times a day for the first two months, right? But it really is something that you just have to get maniacal about and make sure that you are doing your best to make so the second customer doesn't run into those same roadblocks.

Previous Guests (00:01:02) :
That's product market fit. Product market fit. Product market fit. I called it the product market fit question. Product market fit. Product market fit. Product market fit. Product market fit. I mean, the name of the show is product market fit.

Pablo Srugo (00:01:15) :
Do you think the product market fit show, has product market fit? Because if you do, then there's something you just have to do. You have to take out your phone. You have to leave the show five stars. It lets us reach more founders and it lets us get better guests, thank you. Dude, welcome to the show.

Viraj Parekh (00:01:31) :
Thanks for having me. Really excited to chat.

Pablo Srugo (00:01:33) :
Man, so I got to ask. First question, what's your favorite band?

Viraj Parekh (00:01:35) :
My favorite band? Oh, that's a good question. Where's that coming from?

Pablo Srugo (00:01:38) :
I just assume. I mean, it's certainly not Coldplay, I would think.

Viraj Parekh (00:01:43) :
It is not Coldplay. It is not Coldplay, I'll tell you that much.

Pablo Srugo (00:01:46) :
You know, all jokes aside. Obviously I think people, a lot of people that are not in your industry would probably know the Astronomer name from the stuff that happened with Coldplay and your former CEO. And I don't care to relive the moment, but I will ask this one question. Which is, what happens on the inside of the company the day after that kind of goes viral?

Viraj Parekh (00:02:03) :
Yeah, no, it's a super fair question. Public scrutiny was definitely not something I had on my bingo card when I was like, hey, let's really go deep into building a data company in B2B SaaS world. So there's definitely a brief distraction, right? There's no denying that. We had reporters hanging outside of our house, outside of our office. I think every single one of our employees were blitzed by social media and people trying to get quotes, and so on. And I think honestly my biggest takeaway in hindsight is just, I'm in awe at how incredibly quickly the company came together, and kept focus on what was important. Number one each other and number two our customers. I think the most exciting part for me from all that is throughout all of that, right? We do not lose a single customer or a single employee. People who sign up to work here, work here because they believe in the product, the mission, and the market. And that stayed true regardless of whatever external noise was there. And our customers trust us. And we've worked very hard to build that customer trust over the years. And they see us as pivotal parts of really their data strategy, which is driving their business. So that's what people focused on, which brings me a ton of joy.

Pablo Srugo (00:03:12) :
That's huge, obviously and I mean, it makes sense if you think about it. It wasn't like a data breach or some employee that went rogue and did something nefarious. It really was a personal thing. It's great that it's played out that way, I guess, is how I would frame it. The one question I do have, maybe just to still touch on it. Is, just take me inside of the discussion of, OK, what are we going to do? Are we going to let this ride? Ignore? Are we going to face it head on? How do you think through what your action is? As much as you probably don't want to find yourself in that situation, you still have to figure out what the response is.

Viraj Parekh (00:03:44) :
Yeah, I mean it goes back to two things, right? The people who choose to work here and the customers that choose to trust you. So all of our discussions inside was number one, what do our employees need to feel supported? To make it so that they feel like they have all the information they need. To make whatever decisions right for themselves and their family. And number two is, what do we need to tell our customers? What do they need from us? Do they need assurance that, hey, we're still functioning as a company. Do we need to make sure that we're not more vulnerable to cyber attacks, because we have more online scrutiny and so on. So it just goes back to those two things. Everything else is just noise. If you focus on the core of the business, the core of the thing that people are looking to you for, and really the core of the thing that outside of any individual moment or any individual person. What do you want your company's brand to be associated with? What do you want people to think of when they think of your company? Just focus on those things and all the right answers reveal themselves. And, it's one thing to think about what happened on that day. But then I think the real proof in the pudding is around what happens after that. When you look at kind of what we did for the rest of that quarter and coming into the next quarter. Not only did we have our biggest quarter in company history after that. We also made major strides from a business perspective in terms of the product and services we're shipping for our company. You know, we saw mass adoption in Airflow 3 in our customer base. Which was a huge milestone for not just our company, but the open source community that we serve. That happened earlier in the year. We launched a new IDE that has an agentic authoring experience and last but not least. We kind of completely revamped and launched 1.0 of our private cloud product. For customers that still have a large on-prem presence and need orchestration inside of their data center. So it's like, hey, on the day you focus on the most important things, but the real proof in the pudding and where you really figure out what your company's made of. And what the soul of your company looks like is what you do in the months after it. And when I look back on this, you know, I just think about. I just feel pride in how the company not only reacted that day, but the weeks after it.

Pablo Srugo (00:05:41) :
Cool, well let's back it up and really kind of dive into the piece that we really want to be talking about here. Which is the zero to one part of the journey. Maybe just the first question, when do you start Astronomer? What year?

Viraj Parekh (00:05:50) :
Yeah, so Astronomer focused on the Airflow company. Started in 2018, and it kind of grew out of previously Astronomer. The company that started before, and we were focused on something else. We were building a clickstream product. So if you know the company Segment.io, we're basically a competitor to them and I can tell the story in a very long way, but just in the interest of time. Because I know you don't have five hours to talk about this.

Pablo Srugo (00:06:15) :
We don't have five hours, but we do an hour and it's all about depth. So don't gloss over it either. The details are really what matters, right? The story of every startup really is high level. We put some now, people loved it, we grew the team, we grew sales and marketing. And here we are, right? And so it's, the details of just how you ended you where you ended up, I do think they are important.

Viraj Parekh (00:06:33) :
Yeah, all right. So if we double click a little bit then, right? So we had this clickstream product.

Pablo Srugo (00:06:37) :
Which is what, by the way? What's a clickstream product? What does that do?

Viraj Parekh (00:06:40) :
Yeah, so if you are a marketer, or if you have a website, or if you have an app or something. You want to be able to track what customers do inside of your app. So every time somebody hits add to cart or remove from cart, or something. You want to be able to track that. So what our clickstream product did is, hey, you add our tracking snippet into your app and then, when a user goes and clicks something. We'll make sure that, that event gets into an analytics destination of your choice.

Pablo Srugo (00:07:06) :
It's like Mixpanel as well, this kind of thing?

Viraj Parekh (00:07:08) :
Mixpanel, Amplitude, or for a lot of companies. They want to own that data directly. So we'll stream it into their data warehouse, Redshift or Snowflake or BigQuery or something along those lines and what had happened was. The first version we built of that product, we built that backend orchestration piece. We used a cloud-native service for that, something called SWF, Simple Workflows and that was good. But what we found was we needed something a lot more dynamic, and we needed something that gave some more optionality. So Airbnb had just open sourced this tool called Apache Airflow, and we were like, hey, this looks pretty cool. Let's give that a shot. So we built Airflow into the back end of our product. In a way, it was never exposed to customers, right? Customers just click buttons around where they want their events routed and in Airflow, a workflow is called a DAG. So on the backend, a DAG would be generated and that would make it so that customers send these events to the right places.

Pablo Srugo (00:07:59) :
And why build another segment or mixpanel or amplitude? What drove you to do that?

Viraj Parekh (00:08:04) :
Yeah, so our mission was around to get people the data that they needed and when talking to the first couple of customers that we had. What they told us was, hey, they need data from their apps in order to do analytics from them. So it was really about taking your early customer conversations and when you're in that stage of company, almost like pre-product. You just want to build something that you have conviction on will solve problems that you're hearing.

Pablo Srugo (00:08:27) :
But those weren't being addressed by the existing tools in the market?

Viraj Parekh (00:08:30) :
You know, the whole analytics, the marketing, the clickstream space is pretty crowded and there's a ton of great tools out there. So we had found a particular niche based on customers that fit a certain profile. So for us, the profile of customers that we targeted, were those with very seasonal businesses. So if you think of a wedding planner business, right? A wedding planning website has a ton of traffic during peak wedding season, but not so much in December, right? So what we had found was that niche of customers. They needed a solution that better fit their needs a little bit and that's what we kind of designed our clickstream product around.

Pablo Srugo (00:09:02) :
What was, yeah, I'm curious. I mean, so I get the seasonality piece but what about seasonality makes some of those other products not be as good of a fit as they are for, let's say, more standard businesses?

Viraj Parekh (00:09:10) :
It was really on pricing, right? So a lot of the other products would charge you by your average users, and when your average users all happen within three months. It makes your December bill or your January bill doesn't really fit the value that you're getting. So the way that they actually kind of justified price that those companies were getting didn't make sense and we found that. We could provide a service that was more aligned to how those companies operate.

Pablo Srugo (00:09:37) :
Okay, got it. To be clear, this is the product before Astronomer?

Viraj Parekh (00:09:40) :
This is the product before Astro, which is our flagship product today. That's built around Apache Airflow.

Pablo Srugo (00:09:45) :
And this is, we're talking what? 2018, at this point?

Viraj Parekh (00:09:47) :
So this is about late 2017, going to early '18. The dates are a little fuzzy in my mind. It feels like a bajillion years ago, for lack of better words.

Pablo Srugo (00:09:58) :
But yeah, maybe tell me more about that time then. So when you put this product out, what happens?

Viraj Parekh (00:10:03) :
Yeah, so we decided we want to really focus on this problem of getting people the data that they need, and this clickstream product is our first step at that. And we built Airflow in the backend of it. And when you're in that stage of startup, and you're going out there talking about your product. And the problems it solves. What we found was happening, right? And I'm gonna vastly simplify this, was I would say, hey, you should use Astronomer for your clickstream. And they'd be like, why should we do that? And we'd say all these things around, hey, we can give you price for value and all this other sort of stuff. And then I'd say, and we're powered by cool open source technology like Apache Airflow. And what we heard back from the other side was basically, oh, Airflow, that's really cool. Let's talk more about that. I want to do more with that. I saw Airbnb's using it. I'm seeing all sorts of Silicon Valley unicorns adopt it. All these great blog posts about it. My data engineering team's talking about it. Let's talk about Airflow, and this funny thing started to happen. Where we tell customers what our product was powered by and they'd be more interested in that, than the product itself.

Pablo Srugo (00:11:07) :
And what was it about Airflow? What was so interesting about it?

Viraj Parekh (00:11:09) :
Yeah, I think it was really taking advantage of a couple of key things happening in the data space overall, right? Companies were really accelerating what they were trying to do in the cloud. So things like Redshift were getting popular. Snowflake was starting to make noise in the market and so on. So there were all these cloud data warehouses popping up that really gave businesses a ton of value. But the difficult part was getting data to them and then, transforming it and organizing it in a way that it's usable. And previous tools didn't really follow a code centric approach. So there's all these new technologies popping up at the time, right? Around, hey, get your data to the cloud, move with it quickly. Airbnb was writing all sorts of viral blog posts on things that they're doing with data science, data engineering and, so on and so forth. So Airflow is really designed for that, right? Giving developers control over what they can do with their data, with their data workflows in a source controlled back way. Where you could kind of apply software engineering principles to your data workflows and that was the trend that was kind of happening in data space at the time. And Airflow was a manifestation of that, right? It was hooking right onto that. So it fit what companies were trying to do, because it fit what companies with successful data teams were doing.

Pablo Srugo (00:12:19) :
What's the before and after with Airflow? What couldn't you do before that you can, business value you can get now?

Viraj Parekh (00:12:24) :
Yeah, so let me paint a picture for this for you. Suppose you are a data analyst and your job is to get insights out of data. And you wake up one morning and you're like, hey, wait a second, yesterday's data isn't loaded in here. Or hey, it says that we had negative seventy of revenue yesterday. What happened, right? So usually what happens is you go talk to the data engineering team and be like, hey, where's the data? Why does the data not look right? And this data engineering team, which is the users of Airflow and they're the ones tasked with actually delivering that data. They would have to go figure out where something went wrong and that could involve, hey, you have some homegrown tool that you're using. You have some scripts that are running, or maybe you have a drag and drop tool that you can't express your custom logic in in a very easy way. So when data wasn't delivered and the data engineer had to figure out what had to happen with it. That wasn't a particularly pleasant experience for the business. So when the state engineering team started adopting Airflow. What we saw happen is, they could serve the data needs of their business a lot better. They could say, hey, here's in code what needs to happen across all my different systems. Call it, my Salesforce API, my HubSpot CRM, so on and so forth. Getting all of this into my warehouse and then, modeling it in a way that's usable for different teams and different services. All in one code-driven system. That really lets data teams move a lot faster. Because they can reuse things, they can move faster, they can make sure that they're adopting best-of-breed tooling. Rather than being stuck in the confines of any proprietary system.

Pablo Srugo (00:14:01) :
And does that mean that you have fewer of those issues, where the data is not right or the data didn't load? Or that just when you have them, it's easier to debug and figure out what went wrong?

Viraj Parekh (00:14:11) :
It'll be a little of both, right? Because you're doing more, you're going to have more issues, right? Data is moving at a fast velocity through your organization. But, because you have a single tool that tells you where your data is at every point. All your workloads are defined by code, it's very easy to go and take action and fix that with Airflow.

Pablo Srugo (00:14:29) :
When you start hearing from people that you're kind of trying to sell this clickstream product and they're talking about the backend technology. Where's the business at? How many customers do you have at that point? Roughly, what kind of revenue are you doing?

Viraj Parekh (00:14:40) :
Yeah, so before we pivoted to Airflow. I don't even remember how much revenue we had. It wasn't a lot. 

Pablo Srugo (00:14:47) :
Like, $100k or sub? Like, ARR?

Viraj Parekh (00:14:49) :
Somewhere in the $100 to $400 range, right?

Pablo Srugo (00:14:52) :
Okay, but it was something and I ask that specifically. Because, when you have nothing, like you have zero, right? And everyone is saying no, and then they start talking about something else. You're going to naturally as a founder be like, okay, let's look at this other thing, it's going to be very easy. If you've got millions in revenues, it's going to be very hard. But in those hundreds of thousands, you're probably going to think twice about it. Because you're like, well, we've got a product, it is working, I'm sure there's customers that do like it. They're clearly using it and paying for it. So I'm just curious, how did that go internally? Once you start to hear these things, maybe dive deeper into what that phase was like and trying to think through what to do and what not to do.

Viraj Parekh (00:15:24) :
It's one of those things that hindsight's twenty-twenty, and when we look at where we are now. It looks like a great decision. At the time, I wouldn't say it was an obvious decision by any stretch of the imagination, right? This Airflow thing was cool. It was what people wanted to talk to us about. It's what our engineering team was more excited to build around. It was what we saw that we had some more defined expertise around. But because the project was so early, there was no defined market for it. Versus on the clickstream side, if it's like getting data to analytics tools and helping marketers, and mobile apps get data about user behaviors. That was a defined market. People knew what that was, people already had room in their budget to pay for that and, so on and so forth. So it definitely was not an easy decision. I think what helped us make that decision, right? Was, first off, when we talked to our customers. They wanted to hear more about how our product was powered. Because we said the Airflow word, than the actual outcomes we could drive with it directly. Second off, Segment was very good at what they did and we were always playing catch up. And when your job is to play catch up, you know, that's not something that makes it feel very sustainable. It doesn't really give you that much room to innovate. We built incredible expertise around Airflow. That was something that we had developed a ton of IP around, a ton of actual battle scars around, and so on. So we had our own internal conversations, but ultimately what really sealed the deal for us is when we went and talked to customers about it, right? We'd say, hey, like if we were providing an Airflow service, would you want to use that? And what we hear back is, hell yes, I'd love to have an easy button for this problem. I'd love to have an easy button to accelerate where I'm trying to go there. So that was what kind of put the icing on the cake, right? You could have every internal conversation that you want. You can kind of debate things to no end. You can look at studies about market size and, so on and so forth. But when people who are willing to pay say, I will pay you if you can give me an easy button for this thing. That's pretty much all the evidence you need.

Pablo Srugo (00:17:27) :
This Airflow thing. Obviously is not yours, you're using it. What was the top idea of what you could sell customers around this kind of backend?

Viraj Parekh (00:17:39) :
Yeah, so Airflow is an Apache project, right? Which means it does not have a single owner around the IP around it. There's all sorts of great companies built around Apache projects. Confluent publicly traded and it's built on Apache Kafka. Databricks is, I can't even remember how big the last round of funding they raised was but that's built around Apache Spark. There's plenty of existence proof that like, hey, you can build really sustainable, successful businesses around open source technology. What we heard in talking to our customers and going back to the theme of the show. And like finding product market fit, right? Is like, we heard that, hey, Airflow is great but I don't want to have to manage it. I'm a data engineer, and I want to focus on data. I don't want to focus on infrastructure and if I want to use Airflow right now. I have to go ask my DevOps team to spin it up. I have to go ask them to upgrade it when it needs an upgrade. I have to go ask them to maintain it and that slows me down. So in talking to those early users, what we heard is data practitioners want to focus on their data. They don't want to focus on getting their DevOps teams to do them favors. That's a bit of a blanket statement, right? But it's just giving teams that own the data and are responsible for the outcomes the ability to own the outcome all the way. So we kind of really focus on building a managed Airflow product as a result of that.

Pablo Srugo (00:18:56) :
And that's pretty common, right? The open source tech is just what it is, but there's really nothing. I mean, you have to host it yourself, maintain it, like you said, upgraded security. All these sorts of things you have to worry about and then tying it to whatever you're going to build on top. So you guys really build all that layer, made it easy for people to just, you know. The way that they're used to using other services where you just subscribe and pay. That kind of thing, but with Airflow.

Viraj Parekh (00:19:19) :
Yeah, there's a saying that running open source technology is like adopting a puppy, right? It's free to get, but it's a lot of work. Only once you got a puppy, right? And that's kind of where this idea of commercial open source comes in. It's like doing all of the things that you don't want to do and just focus on using the core thing.

Pablo Srugo (00:19:38) :
Tell me about selling in the beginning. How do you do it? Do you go, you build a product, do you just talk to customers, do you get somebody to pay for it up front, and then you build it? How did you? Because you have an existing business, so it's like, you're probably trying to figure out how much proof can you get. Before you go down this stream and abandon the thing that you've got that's at least somewhat working.

Viraj Parekh (00:19:56) :
Yeah, so I think there's different versions of this as your company grows, as your product grows, as your sales teams grow and, so on and so forth. I think in the early days, it's really about going into conversations with your biases turned off. It's like, hey, I want to focus on the problem this customer is bringing to me. I don't want to focus on trying to find some sort of confirmation bias for a preconceived idea that I have, and when you can turn off that knack. That kind of human instinct to try and find confirmation bias for a hypothesis that you have. What you'll find is you can actually have much more fruitful conversations with potential customers around what they're willing to pay for.

Pablo Srugo (00:20:34) :
How do you do that? Because I totally agree, but it's so hard to not. That sunk cost piece and frankly, the vision and belief that you have going in. And just kind of leave it aside, and be like, yeah, no, I'm blank slate.

Viraj Parekh (00:20:46) :
It's definitely an art, because on one hand. You don't want to be like, I'll do anything for you, right?

Pablo Srugo (00:20:51) :
Right, right. I'll mow your lawn, yeah.

Viraj Parekh (00:20:53) :
Yeah, exactly. I'll mow your lawn. So you don't want to come in completely without some sort of hypothesis. But you want to be able to really just draw the line between what the confirmation you're looking for and what the person's really trying to say. Unfortunately, the only way I found to get myself to be able to do it and other founders I've talked to be able to do it. Is just to do it nonstop. You really have to get yourself to get like five, ten, fifteen, twenty calls a day. Most of which you're getting bad news on and build that scar tissue. And build that internal resilience to be like, hey, this was the one key takeaway. That made all of those twenty no's worth it ten times over.

Pablo Srugo (00:21:44) :
I mean, there's two pieces to it. Mainly one is, the questions you have to ask. That's really the art piece, which is you can really if your good at it. Frame things in a way, to ultimately hear what you want to hear and that's not going to do anybody any service, right? You go and you show somebody UI. And be like, how good do you think this UI is? Most people be like, it's pretty good. That's not going to get you anything. I mean, that's obviously a silly example, but I think the way that you frame questions is really important. The second one is what you touched on, which is getting to a place where you're comfortable with people telling you things as they are or as they think they are. Versus you might want to hear them and being totally okay with that distorting your reality. Because it doesn't feel good when you want them to be like, yeah, I'll pay a lot of money for this and they're telling you that it's garbage, right? But if you just ultimately, just want and you're seeking out the truth. And you know that getting as close to reality is what's going to lead to a big business but you just need reps. I think without the reps you just can't get there. We have tens of thousands of people who have followed the show. Are you one of those people? You want to be part of the group. You want to be a part of those tens of thousands of followers. So hit the follow button. So I took you on a tangent there, but back to the specifics. In your case, what did you wait to hear until you decided to build something? Or how did you manage that kind of build something, but get enough validation to build the next thing until you ultimately get a customer?

Viraj Parekh (00:22:51) :
Yeah, so the questions that worked for us, right? Was like, hey, if we had an easy button for that, would you pay for this? It's kind of a direct question and it kind of can put people in an uncomfortable spot. For folks that really had the problem or really were the right sort of. We call them design partners for that early stage. It was a really good determination, right? Because if somebody was waffling on it, then it probably wasn't urgent for them. But if someone's like, yes, without hesitation. Then you know that there's a higher chance that, that person's going to be with you on the journey.

Pablo Srugo (00:23:20) :
And were you putting a price on it? Would you pay $10,000 a year or $100,000 a year, or whatever it might be?

Viraj Parekh (00:23:25) :
Yeah, it depended, right? Like, different sized companies. We would kind of go in with different numbers around. If we're talking to a Fortune 500, then we'd say one thing.

Pablo Srugo (00:23:33) :
But you would put a number. It wasn't just, would you pay for it? It was like, would you pay this much for it?

Viraj Parekh (00:23:36) :
Yep, yeah, yeah and in the tech landscape. A really good guidance point for that, is to do it with a based on the salary of an engineer, right? You can say, hey, was this going to be worth half an engineer salary to you? And you kind of put a number on that, or is this going to be worth a full engineer salary for you, or five engineer salary for you? And that was the kind of yardstick that we had used. 

Pablo Srugo (00:23:55) :
I love that and I love yardsticks. Because one of the things that it helps clarify for a founder is sometimes. Especially as a zero to one founder and especially if you haven't been there, done that or sold big contracts. You can get a little intimidated by the size of the spend, right? You asked for $20k or $50k, or $100k, or $200k and you're like, wow, it's a lot of money. And then you think, okay, one engineer fully loaded, right? Depending on the city is costing, what? $100k, $200k, $300k, and then this company has a hundred of them. Depending on the size of the company, a thousand of them and you're like, I'm just telling you that instead of, you know, is it worth going from $100 to having $100.5? That's really what we're talking about here and it just contextualizes it in the sense that. If it's not worth that, frankly, it's not worth much. Because they're going to hire ten engineers this year. They're going to spend money and so the question is, are they gonna spend money on you? And so anyways, framing it that way relative to other, could be to engineers, could be to other budgets. 

Viraj Parekh (00:24:47) :
Yeah

Pablo Srugo (00:24:48) :
Right? But just forgetting the dollar figure and realize this is an enterprise. They spend millions of dollars a year. So what are they spending hundreds of thousands on, and are we gonna get a share of that. A meaningful share, or are we kind of wasting our time?

Viraj Parekh (00:24:58) :
Exactly, right. And then just building on that a little bit. The term stage appropriate is thrown around a lot, but I think the way you think about value for an early stage product. You really have to make sure you're doing it in a stage appropriate way. So when we were based in Cincinnati and we were less than ten people. What was stage appropriate for us was to think about cost relative to that person hiring someone to do it, right? That was basically the only bargaining power that we had. What ended up happening, because of the work that we were doing, and other companies were doing to make Airflow better in the open source. The criticality of use cases people were running on Airflow started to wildly increase. In the early days, Airflow was used for very, what I call, analytics workloads. As in the sense that it would move data around, and ultimately the end result would be a number in a dashboard. And that number needs to be right. And if the CEO looks at that number, and it's not right. They're going to be like, what the heck? An executive might look at that number in the dashboard, once a month or something. What ended up happening is Airflow still does those analytics workloads, but it also evolved to be used a lot for operational workloads. So that's like, hey, if this regulatory report does not get sent at the end of every day. The CIO is personally liable, or hey, this is going to feed data into an application that an end user looks at, or hey, this is going to make sure that my models are trained on my most recent user data. So things directly applicable to a business's operations. Where if that doesn't work properly one day or that goes wrong. Then there is severe revenue implications for the customer and in that stage for us, the way you think about price and value is completely different. Because you're not thinking about, hey, what is the cost of an engineer doing it, right? You're thinking about what is that problem worth to a business and how would a business solve it without your product? That's another reason why paying attention to those early customers and how the life cycle of that customer changes as they adopt your product. And your product changes really matters. Because those early design partners are going to be the one leading edge on that. They're going to be the ones saying like, hey, yeah, I started using this for analytics workloads but look at this operational workload I'm doing now. That will empower you and whatever sales team you have. To go to those Fortune 500s and say like, hey, did you know you could do this? Did you know that this is a way that you can use this open source technology? Look at the value that could be driven if you go and do things that way. And you can kind of learn from those design partners in a very hand in hand way, to try and build what the journey you want for the rest of your customers to go down is.

Pablo Srugo (00:27:37) :
So tell me, after you had these conversations. Who was your first customer, and how did you? Did you do a design partner program, or how did you set up the first few implementations?

Viraj Parekh (00:27:46) :
Yeah, so as far as developer tooling goes in B2B SaaS world. I don't know of anyone whose first handful of customers hasn't been a bit of a knife fight. In the sense that you're just getting on the phone with them every day, trying to make sure what they're doing works and doesn't work. Our first customer that paid us for managed Airflow, was a data scientist who was working for a media company in Ireland. He was a former Googler and he said, like, hey, yeah, I had heard from my former Google colleagues that Airflow is a tool for this. I just need my data loaded into BigQuery every day. So I can focus on data science and I don't want to have to think about that. I think I had talked to him no less than four times a day for the first two months, right? You talk on intercom and different sort of tools. Sometimes you're getting on a Zoom call and kind of working through things. But it really is something that you just have to get maniacal about and make sure that you are doing your best to make so the second customer doesn't run those same roadblocks. There's no substitute for elbow grease there.

Pablo Srugo (00:28:44) :
Did you shutdown the clickstream product right away?

Viraj Parekh (00:28:50) :
No, so one of the cool parts about it is that product actually got spun out into a whole different company and that company is still operating, and doing great today. So that market was still there for them. It's just we wanted to focus on the Airflow thing, but that didn't mean that the problem we were trying to solve with clickstream was relevant.

Pablo Srugo (00:29:08) :
And did you split up the co-founding team or how did you? Who took over that other business?

Viraj Parekh (00:29:12) :
Yeah, so some of the folks that were at Astronomer went down the clickstream route and they kind of spun out. And then the rest of us stayed to focus on Airflow.

Pablo Srugo (00:29:20) :
Yeah, you have to find a way to fully focus. So either you shut it down or you spin it out. But certainly you can't have a small team doing both. OK, so you've got this one and I assume that first contract with this one engineer was probably relatively small. 

Viraj Parekh (00:29:33) :
Yeah, I don't remember the exact number but it wasn't more than like $20,000. 

Pablo Srugo (00:29:37) :
Do you try to land other customers right away simultaneously? Or do you try and just use this one customer to get your product to a certain level first?

Viraj Parekh (00:29:46) :
Yeah it depends. Because you want to show traction and momentum. The other thing here was we were a Cincinnati based startup. That was not based in New York or San Francisco. It didn't have the Silicon Valley presence or didn't have the creator of the open source. 

Pablo Srugo (00:30:02) :
Had you raised money, up until that point for Astronomer?

Viraj Parekh (00:30:05) :
We had raised a little bit. A bit of like a pre-seed sort of thing. 

Pablo Srugo (00:30:08) :
Okay.

Viraj Parekh (00:30:09) :
But we needed to show traction in a way that was set at a higher bar than if we were based in San Francisco or if we had had the creator of the open source project on our team. So for us at the time, we needed to show more traction than I think what our parallel sort of company would have. So we were trying very hard to get additional folks. As we were working with some of those early ones and we would do all sorts of things, right? We would say like, hey, we'll do Airflow training for your team as part of signing up for our product. We'll bundle in a bunch of professional services as a way for signing up for a product. We very much had to show that we were the ones who could build a company around Airflow. Despite the fact that we didn't look like any of the other traditionally successful

Pablo Srugo (00:30:51) :
Yeah, tell me more about landing, let's say. Let's call it the first ten, maybe fifteen customers. Because, I mean, this is obviously a critical part of zero to one and I think founders really want to go the tactics of. Ultimately, what were the things that actually worked and actually got you ten, fifteen customers to sign up and pay you?

Viraj Parekh (00:31:09) :
Yeah so I think, first and foremost there's a line between selling the vision. And making sure you're clear with what you can do for the customer. On that day or in that month and in that year. So you really have to be future focused and like, hey, here's where you're going. But what doesn't work is if you tell a bunch of people, here's where you're going. They expect that to be what you have on the truck that day and then they leave in a not so happy way. And then word spreads in a negative way.

Pablo Srugo (00:31:40) :
But you do need both is your point. You can't also just talk about here's where the product at today. Because it's just it's never going to be enough kind of on day one.

Viraj Parekh (00:31:47) :
Yeah, and that's why it's really important to find those early design partners, right? Because at the end of the day, what you're saying is like, hey, trust us. Here's where we are today. Here's where we're going. Here's when we think we're going to get there. Do you want to jump on for this ride or not, and a lot of times. I think humans respond really well to authenticity and I think everybody knows that your product's not going to do everything out of the box that you're claiming, right? No one's going to be totally shocked by that. Especially when they go look you up on LinkedIn and see that you're ten people or fifty people or whatever. Information is pretty transparent on those things. So I think just adding the human authenticity element to it. I'm just being like, hey, here's where we're at. Here's where we think we're going to go. We would love to make sure that we go in a way that aligns with your needs. But you can't go out there and promise that you can do 24/7 support all around the world. But you don't have anybody outside of your time zone, right?

Pablo Srugo (00:32:38) :
But how do you even get these? Just cold calling, cold emailing? What method did you use logistically to just get in front of these people?

Viraj Parekh (00:32:45) :
Yeah, for us, email campaigns worked really well, right? That was something that we found success in. During that early time, nowadays.

Pablo Srugo (00:32:53) :
Was this because Airflow was so top of mind. That if you just mentioned you had a managed service on top of Airflow. It would, you know, a lot of people would want to kind of learn more?

Viraj Parekh (00:33:01) :
Yeah. In fact, now that you're saying it. I think our most successful email campaign was. The subject line was, don't open this email unless you use Apache Airflow. That was the subject line we used. Yeah, we started an Airflow podcast as a way to just start to orient ourselves and give ourselves legitimacy within the community. We did all sorts of tips and tricks to get there.

Pablo Srugo (00:33:20) :
Finding something that's top of mind and riding it is massive. Because the reality of an early stage startup is you don't have any sort of brand. As much as you might hope that you do. Nobody knows who you are and so being able to tie yourself to something else that's top of mind can really change conversion and response rates.

Viraj Parekh (00:33:38) :
Yeah but at the same time there, right? I'm going to challenge that a little bit, in the sense that you don't have a brand but you're also building what your brand is. You're building what the feelings and thoughts that people want to leave your company with are. One super specific example I can give you that what we did. So fundraising in the early days was very tough for us. So we were very, very lean to the point where we didn't even have like company t-shirts. So when we'd go visit customers, because we didn't have a bunch of Astronomer t-shirts we could give them. One of the things we'd do is, we'd look up who we're visiting, find out what they majored in in college, and we'd go to like the used bookstore, and get a book about that topic. Because that was a whole lot cheaper than getting a bunch of shirts. We'd show up not with a company shirt but with a book, right? And that was a little thing that we could do. That showed that we cared more than the other people that we're talking to. It was out of necessity, right? But it was still very much like, hey, what sort of company do you want to be? We want to be a company that pays more attention to you. We need to be a company that listens to you more to solve your problem. Because we don't have the legitimacy of a Silicon Valley unicorn.

Pablo Srugo (00:34:39) :
Yeah, I mean, I think that's totally legit. Your actions are what build your brand. But I think the point is in the sense of getting that net new customer, they have no idea who you are realistically. They probably won't until that spreads into word of mouth and drive referrals. But that's kind of a longer, you know, it's probably not going to get you customer number five but it might get you customer number fifty, certainly five hundred.

Viraj Parekh (00:34:57) :
Exactly. Exactly.

Pablo Srugo (00:34:59) :
Tell me a little bit about, you mentioned, you touched on fundraising. Where do you get to, when you decide to go raise your first real, let's say, Seed round or Series A. Whatever it was, and how did that go?

Viraj Parekh (00:35:10) :
Yeah, so getting the first bit of money in the door to try and, stay operational, keep the lights on, you know? I know that was not an easy process. We decided to do it. Number one, out of necessity and number two, after we had a couple of actual things we could say that we did. We ended up raising a larger round in 2019, after the initial round of 2018.

Pablo Srugo (00:35:31) :
How much were these rounds? Like 2018, and 2019?

Viraj Parekh (00:35:33) :
I don't remember the exact amount of the Seed round. But in 2019, it was about $13,500 or something as a Series.

Pablo Srugo (00:35:41) :
And the Seed round was probably what? Like $2-3 million, let's say? Something like that?

Viraj Parekh (00:35:43) :
In that ballpark, maybe five. I don't remember off the top of my head.

Pablo Srugo (00:35:46) :
Okay.

Viraj Parekh (00:35:47) :
But what changed between then, right? Was we went and found all of the most active Airflow contributors and we got them to join the team. And we made it so that we could very much signal to the world like, hey, look at how much we're driving Airflow. We're the Airflow company and that was something that we have to earn the hard way. You know, we have to earn the right to do that. With our actions, not just words.

Pablo Srugo (00:36:09) :
Do you remember where you were at, when you raised that $3 to $5 million Seed round or whatever it was? Were you a million ARR or half a million, $2 million? Do you have a sense of kind of where you were at?

Viraj Parekh (00:36:17) :
We were probably in the half range, something between the half and the one. Granted, this was a very different fundraising environment than what we have now and like, I had said earlier, we had to show more traction. Because of who we were and where we were. Than the equivalent startup had to do at the time in New York or California.

Pablo Srugo (00:36:35) :
Did you go down to New York or California to fundraise? Or how did you do it?

Viraj Parekh (00:36:39) :
We did have some local investors in Cincinnati, but the folks that led our Seed round and our A round were Silicon Valley based.

Pablo Srugo (00:36:46) :
How fast did things grow in those early years? I mean at the beginning it kind of doesn't matter that your first few customers, but once you had the product right. How fast did you get to a million and then $5 million, sort of thing in ARR?

Viraj Parekh (00:36:56) :
After the A round, right? The Seed in A round, we actually got a good deal of growth from COVID happening. Just because everybody was trying to move to digital native and data budgets exploded. So there was a lot as a company we had to adjust to. Just like everybody else due to COVID and lockdowns. For data budgets, it made them more plentiful and that definitely aided our growth at the time.

Pablo Srugo (00:37:19) :
Do you remember what the ramp was like from one to ten? 

Viraj Parekh (00:37:22) :
I do not off the top of my head.

Pablo Srugo (00:37:23) :
How long did it take to hit, even like a million an hour?

Viraj Parekh (00:37:25) :
Getting to the first million definitely took over eighteen months. After we got to that, it definitely went faster but it was a slog to get there.

Pablo Srugo (00:37:33) :
Obviously, the faster you can get to a million. Just like anything else, the better but I really do find how long it takes to get to a million is not that meaningful. I mean, it matters a lot more how you grow after a million than before and that's just like an arbitrary line, drawing in the sand, right? Because depending on the complexity of the product, the market readiness, all these sort of things impact how fast you get to a million. But at a certain point, once you find product market fit. How fast you move after that is really telling of how real your product market fit is versus how much you're just going to have to slog it out for years and years to come. Kind of pushing a boulder up a hill.

Viraj Parekh (00:38:06) :
I deeply agree with that. I think the lessons we learned in how hard it was to get to a million for us made it. So that we could better serve our customers and build the company, and growing past that. Because that's when all the humbling happens, and that's when all of your resilience gets built. And candidly, I think it's a pretty good thing for a company's long-term outlook to go through that earlier on. So that you know when things really start to work. It's because you found the right thing. What you don't want to happen is have a couple of early successes, and it turns out that those were flashes in a pan. And not something sustainable that you could do or a regular sort of problem you could solve for your customer base.

Pablo Srugo (00:38:48) :
And when would you say that you felt like you found true product market fit?

Viraj Parekh (00:38:52) :
Near the end of 2018, and early 2019. We would start to talk to customers and they would say that, hey, we've decided on Airflow. As the thing we want to do for a year and we hear that you guys are the best way to do that. It happened for us, because we got legitimacy as a company as Airflow became more and more prevalent inside of an organization's data stack. And, you know, if I fast forward to today. We have Fortune 100 companies coming to us and saying like, hey, we have all these AI initiatives. We're trying to get into production and it's like every time I hear it, it still feels a little surreal. Because I remember how much the opposite was if I just like turned back seven years.

Pablo Srugo (00:39:32) :
And then the second to last question. I mean, things obviously gone extremely well. You've raised over $350 million, you have hundreds of employees but was there ever a time in this journey over the last seven, eight years. Where you actually doubted that things would work and thought, maybe it would all kind of crash and burn?

Viraj Parekh (00:39:49) :
Yeah, I think like anybody who works in an early stage company and is dedicated to that company. Be it if you have the founder title or not. You're going to have emotional ups and downs, right? You're going to have this kind of sinusoidal experience. Oftentimes in the course of a day or an hour, where you're like, hey, this is going to make it so I can buy, I can do whatever. Versus like, hey, this is going to be the train has completely derailed and this is going nowhere, right? So you have those sinusoidal experiences all the time. I think when I look at what's happening right now. Two things make me feel like those down days aren't nearly as down anymore. Number one, it's the fact that Airflow is downloaded over three hundred million times last year and there's eight thousand plus companies using it. And that includes everyone from Seed stage startups to the biggest banks in the world, to companies like OpenAI and Anthropic. That are building incredible next generation technology. So when I look at that I feel energized and optimistic about all the problems that we have to solve day to day at a startup. I think the other thing on that, is just really the team around you. Everybody has ups and downs. Sometimes something that goes on in your personal life will make you feel a little more down at work. Sometimes something that happens positively in your personal life will make you feel a little more energized at work and sometimes it'll be the opposite, right? So that early team around you and the folks that you choose to like spend the bulk amount of your intellectual and emotional energy with day to day. I think at the end of the day, that's truly what'll keep you trending upwards. Without solid customers and use cases that you see every day. And without the right sort of people around you, I don't know how anyone would do it.

Pablo Srugo (00:41:27) :
And then last question, what would be one of your top pieces of advice for an early stage founder in this kind of finding product market fit stage?

Viraj Parekh (00:41:36) :
That's a good question. The biggest one would just be like, it's supposed to be hard. Hard is a relative term, but if it feels really hard that means you're doing it right. So that'd be number one is just like, take it with joy. Just remember that with what you're doing, but you're feeling like you're messing up every day. That is, that is what good looks like and the most important thing you would do is just try again with energy and enthusiasm. So that would be the biggest one. The second one would be, and I know I'm a broken record onto this. Is talking to customers and prospects is never a bad use of time. In fact, it's hard to justify what could be a better use of time, most of the time.

Pablo Srugo (00:42:13) :
Awesome. Well, listen, appreciate you jumping on the show. It's been great to hear the story.

Viraj Parekh (00:42:17) :
Awesome, yeah. I appreciate you having me. I appreciate you letting me wax poetic about some of the past there and, always happy to come on and chat again.

Pablo Srugo (00:42:26) :
Wow, what an episode. You're probably in awe. You're in absolute shock. You're like, that helped me so much. So guess what? Now it's your turn to help someone else. Share the episode in the WhatsApp group you have with founders. Share it on that Slack channel. Send it to your founder friends and help them out. Trust me, they will love you for it.