This is the third installment in our series on product-market fit, spearheaded by First Round partner Todd Jackson (former VP of Product at Dropbox, Product Director at Twitter, co-founder of Cover, and PM at Google and Facebook). Jackson shares more about what inspired the series in his opening note here. And be sure to catch up on the first two installments of our Paths to Product-Market series in this interview with Airtable co-founder Andrew Ofstad and this interview with Maven founder Kate Ryder.
We touch technology every day — from the ride-share service we call to get to work, to the credit card we swipe to purchase food, to the fitness app we use for our favorite workout.
But rarely do we pause to think about what’s happening on the backend. Whirring in the background of any operation — whether it’s a two-person startup or a Fortune 500 company — is a complex network of internal software that keeps everything running smoothly.
Without internal tools, like admin dashboards and customer support apps, companies can’t sustainably scale, serve their users or optimize their products. In other words, they can never achieve operational excellence. This, in essence, is what David Hsu believes Retool’s mission is: to help companies build the high-quality internal tools they need to grow quickly and effectively.
While Retool has hundreds of customers today and continues to grow at an impressive rate, it wasn’t always clear that his company was going to succeed. Hsu had to overcome many hurdles — from ignoring the naysayers who told him to abandon his idea to sending out thousands of cold emails — to get to where he is today.
But to understand the full journey, we have to rewind to Hsu’s Oxford days.
EXPLORING IDEAS
A natural-born tinkerer and thinker, Hsu attended Oxford to pursue a dual degree in Computer Science and Philosophy. During his time there, he worked on five to 10 different side projects with his friends. But no matter what he was trying to build, he always ran into the same hurdle: internal tools.
“We were spending probably half of our time on internal tools. And as lazy engineers, we were like, ‘There has to be a faster way of doing all this stuff.’ So that's where the core idea for Retool came from.”
From the start, he envisioned Retool as a drag-and-drop platform that helps developers build internal tools faster. But there was a problem that Hsu was well aware of, which is that developers hate drag-and-drop products.
“If you told me that I should build a website using drag-and-drop, I'd say that’s not for me. I'm a hardcore engineer. I write code. For Retool to work, we were going to have to reach people who already know how to code and change their minds about how they should code,” he says.
So, on the surface, the idea for Retool didn’t seem promising. And people weren’t afraid to tell him so. Hsu, who was part of Y Combinator’s Winter 2017 batch while working on Retool, had several founders from his cohort try to dissuade him from pursuing his idea — with some suggesting that he pivot to a no-code platform and target non-developers instead.
This suggestion made sense. After all, this was 2017, when the idea of democratizing programming was very en vogue. But Hsu, recognizing that following trends wasn’t going to lead him to success, stood his ground.
If you want to launch a startup, you need to have differentiated beliefs about the world. If your beliefs are the same ones that everyone else has, there's no reason why you're going to find success when others didn’t.
BUILDING THE EARLY PRODUCT
According to Hsu, all internal tools boil down to tables, buttons and text inputs.
So the initial version of Retool was, straightforwardly, a generic programming environment that allowed developers to build a table component, a button component and a text input component.
Hsu tested this MVP on other companies from his YC cohort. One of their first customers was an app that allowed drivers to work shifts with both Uber and Lyft at the same time. This startup needed a lot of internal tools to manage the billing process and work compatibly with the ride-sharing apps.
But the company didn’t have its own backend servers and therefore relied on public APIs for integration — something that Retool didn’t yet expose. In other words, their first customer couldn’t actually use the product. Hsu’s response was immediate:
“We told them to let us get back to them tomorrow. Then we stayed up all night to build it, and it worked. The early days of Retool were very fast iteration cycles like this,” he says. “In one day or one night, we'd come back with a new feature. Sometimes it worked, sometimes it didn’t. And we would just have to iterate from there.”
HONING IN ON THE ICP
While building the product, Hsu was still trying to fine-tune the company’s ICP. In the early days, he started off with two assumptions:
- The target user for Retool was FileMaker developers.
- The ideal company size for Retool was small startups.
Here’s the framework Hsu used to get crystal clear on the ICP for Retool:
1. Start with a hypothesis
Rather than throwing spaghetti at a wall to see what sticks, Hsu always started off with a strong hypothesis — a starting point that he could pressure test and re-shape as needed.
- Target user. When it came to finding their target users, Hsu needed to get clear on what type of developer they were trying to reach. “Since Retool is a product that’s kind of similar to FileMaker, Microsoft Access or Visual Basic, we thought we should go find similar customers and convince them to try Retool. This seemed like a very plausible path to finding product-market fit.”
- Company size. Similarly, the Retool team had ideas about what company size they were going after. “When we first started the company, we assumed it would be used by other two to three-person companies. We thought our product could be a fast way for a small company to build their internal tools.”
2. Test your assumptions
Hsu then tested his hypotheses — whether that was through extensive research or calls with potential users — to see whether his assumptions were correct.
- Target user. To get into the minds of his potential target users, Hsu “infiltrated” several LinkedIn groups for FileMaker developers and conducted extensive outreach. “We sent out a few hundred cold emails, received like three replies back and got one person on a call. That developer pretty much said that Retool is a horrible idea and wouldn’t ever consider using anything other than FileMaker.”
- Company size. Research revealed that the market for Retool was significantly larger than Hsu and his team originally anticipated. “We discovered that around 50% or 60% of all the software in the world is actually internal facing. And once we discovered that, we were like, ‘Wow, this is a ginormous market.’ If we can change the way that half of all software is built, that would be really incredible.”
3. Iterate
Once Hsu collected enough information, he iterated quickly and moved on to the next hypothesis. Lather, rinse, repeat. Until — finally — he got to the right ICP.
- Target user. Through this process of iteration, the Retool team discovered that their target users were frontend and backend developers who work mainly in React or JavaScript. And their initial obstacle around getting engineers to use a drag-and-drop platform turned out to be irrelevant. “It turned out that these developers hate building internal tools more than they hate drag-and-drop systems,” says Hsu.
- Company size. After uncovering the massive market for internal tools, Hsu did some more digging and discovered that the companies with the most pressing needs were large corporations. “One of our early customers was a Fortune 250 company with around 120,000 employees, and they were spending $400 million building internal tools every year,” says Hsu.
FINDING CUSTOMERS
One of the most impressive aspects of Retool’s journey was how quickly the company grew its customer base — scaling to dozens of logos in less than a year, including some major startups like DoorDash, Brex and Allbirds, as well as Fortune 500s. Hsu shared the tactics he used to grow Retool’s customer base:
Tactic #1: Lean on outbound emails
In the early days, the Retool team relied solely on outbound — in person and via email — to find new customers. In Hsu’s view, a cold outbound strategy can be an optimal approach for early-stage startups:
“It gives you a very good sense of how something resonates. On the other hand, if you get a warm intro, people will take the call and maybe ask for a demo. It’s very non-committal and vague, and I actually think that’s harmful to startups because it can make them think they have product-market fit when they don’t.” (Sprig founder Ryan Glasgow shared similar advice with us last year.)
Tactic #2: Get to 100 happy customers
Another factor that helped Retool win was its intense customer obsession. This is something that most startups claim but, with Hsu’s team, it was truly the North Star metric that drove all their decisions.
When we started Retool, our goal was always happy customers — specifically, how do we get to 100 happy customers? Because if we can find 100, then we can probably find 1,000, and then 10,000.
As an example, Hsu points to the in-house analytics platform they built that allowed them to collect data around what people were doing with the product. They connected this platform to Slack so that any time someone was active in the product, the team would get a notification.
“We'd immediately watch what the customer was doing in the product. And that was extremely helpful because we can't always be next to customers while they’re using the product, but these analytics got us closer.”
The team took it a step further by immediately reaching out to affected customers any time an error occurred in the product. “We would reach out to them with a call or even a text. From a customer experience perspective, their experience was really intimate in the sense that, if they run into any error while using the product, the CEO would call them a minute later. I think that built a lot of trust with our early customers,” says Hsu.
Tactic #3: Find language-market fit
Another key tactic Hsu used to find more customers was to focus on finding language-market fit. When they first started Retool, they positioned the platform as an “Excel sheet with higher-order primitives.” As you might have guessed, that didn’t land very well with potential customers.
“It turns out that nobody knew what that meant,” admits Hsu. So the team continued iterating on the messaging. One day, Hsu sent a cold outbound email to a startup called Rappi, positioning Retool as a platform that helps companies build internal tools faster.
“Within 15 minutes, Rappi’s CTO replied asking to get on a call. I was like, ‘Cool, how about tomorrow?’ And he was like, ‘How about today?’ The fact that the CTO of a thousand-person company wanted to get on a call and take 30 minutes out of his day to talk about his struggles with internal tools, that was the first sign that our positioning, which is our core messaging today, was finally resonating.”
LAUNCHING
Retool finally launched publicly in 2018. At this point, the company already had around 40 customers and around $2 million in ARR — an unusually strong position. Hsu unpacks why they waited so long to launch:
“We’re in the developer tool space, and developers have very high standards. We knew they weren’t interested in using a product that was buggy or had a crappy user experience. So we wanted to launch only when we were confident.”
You’d be forgiven for assuming that Retool had product-market fit well before their delayed launch. But, according to Hsu, it wasn’t obvious to him.
We talked to someone who said that finding product-market fit was so visceral that you immediately feel it — like a geyser exploding. We honestly never felt that. Every customer we got — whether that was number four or number fourteen — felt like the last customer we were ever going to find.
That’s because every customer used Retool for wildly different purposes. “DoorDash was building logistics tools for drivers. Brex was building tools to manage credit limits. And that was totally different from what a tutoring company was using Retool for,” Hsu says. “So I remember wondering whether we would actually be able to serve these customers and make them happy with so many different use cases.”
It wasn’t until they got to around 40 logos that Hsu started to feel confident that they could in fact support a wide range of use cases, while still scaling and keeping customers happy.
For Hsu, the takeaway was clear: getting to product-market fit can be more ambiguous for some than others. It wasn’t the strike of lightning that he anticipated, but rather a slow burn that took time to get going.
His advice for founders who are similarly struggling to pinpoint whether they’ve truly reached that high bar of product-market fit? Stay paranoid, and keep pushing. A potential signal you’ve reached product-market fit is when you feel like you can relax a bit without worrying about the rock rolling back down the hill.
It’s like rolling a stone up a hill — if you stop pushing, it'll start rolling back down rapidly. That’s what our experience of searching for product-market fit felt like until we had a few million in ARR.
THE PATH FORWARD
Since Retool’s launch in 2017, the company has firmly established itself as the go-to platform for developers who are building internal tools. Hsu and his team also recently raised a $45 million Series C round, backed by big names like John and Patrick Collison, Daniel Gross and Elad Gil.
The platform serves major tech companies like Amazon, Pinterest and Coursera, as well as corporations like the NFL, NBCUniversal and WarnerBros.
With a solid customer base under his belt and a new injection of funding, Hsu has big plans to double down on the Retool platform, grow his team and continue serving the company’s mission of helping its customers achieve operational excellence.