Hitting a 7-figure ARR within two years by focusing on quality and letting the product speak for itself

Yu-An (Alan) Chan dropped out of school, built a product, rebuilt it, joined Y Combinator, launched, and then iterated like crazy.

Less than two years later, Heptabase is bringing in seven figures per year.

Here's Alan on how he did it. 👇

College dropout x2

I studied Physics and Math at National Taiwan University in 2016 and Computer Science at Minerva University in 2019, but voluntarily dropped out from both after my freshman year. I left NTU to explore a broader range of interests and deepen my understanding of the world, and I left Minerva to focus on building Heptabase.

I really love learning new things and experiencing the "aha" moment when I grasp a deeper understanding of topics I'm interested in. I wanted to make this experience easier for people, including me. That's the original motivation for building it.

Heptabase is a visual note-taking tool designed to help people learn and research complex topics more effectively.

We participated in YC’s W22 batch and launched our 1.0 in September 2023. Now, we're making a few million in ARR with >80% gross margin and >20% net margin. And we won the 2023 Golden Kitty Award in the Personal Productivity category on Product Hunt.

A half-baked product

After I dropped out of Minerva, I spent a few months exploring several ideas for tools that could help people acquire deep understanding of things they care about. The original idea was to design a browser + note-taking app + publishing platform that people could use to learn complex topics.

The first few months went poorly because we spent too much time trying to design and build a perfect product, which made everything half baked and unusable. It was a rookie mistake.

Since we had no funding and very little savings at the time, I soon realized that this way to build things was not going to work, so I decided to start from scratch and build a simple MVP that focused on the idea of reusing atomic knowledge across different visual contexts. Users could create note cards and reuse them across different whiteboards.

I launched it in a week and, after a few thousand iterations, that became the current version of Heptabase.

Here's the stack:

• Language: TypeScript.

• Frontend: React, Redux, Tailwind CSS.

• Backend: MySQL, Node.js, AWS.

Most of our core components, such as the whiteboard, card editor, and sync mechanism, were built in house.

Organic growth

Communities

In the beginning, I went to the Discord communities and Telegram groups of several note-taking apps, searched for people who mentioned keywords like "visual," "whiteboard," and "mindmap," and asked them to play with my prototype. That's how I acquired my first ten users.

Word of mouth

After that, I started a Discord community to chat with these users every day, refine my vision based on their feedback, and iterate the product toward a refined vision. The product kept getting better, day by day, and as we built more powerful features, users got excited. They told their friends to try it out.

I think the reason we can rely on word-of-mouth growth is pretty simple: Our product is very useful, a lot of users love it and use it every day (including me), and they want to tell their friends about it. This is probably the most important aspect of growing a consumer-focused business.

Affiliates

We didn't spend a dollar on marketing until we started an affiliate program near the end of 2023. Some of our users are content creators, and they asked us for affiliate links. Now, 10% of revenue comes from our affiliates.

A simple strategy

So our growth strategy was very simple:

• Create a community

• Talk to our users frequently

• Keep shipping and making the product as useful as possible

• Charge early to get focused

• ​Let word of mouth drive the majority of growth and acquire affiliate partners along the way from our customer base

Improving the product

In the first few months, I asked every new user to schedule a user interview with me before letting them try our product. During the interview, I told them to share their screen so I could see how they used the product for the first time.

I asked them a lot of questions about their backgrounds and why they wanted to use our product. This allowed me to quickly identify the most important things to build and who we were building this for. We also use Amplitude to track the overall usage retention to have a sense of whether users will continue to come back to our product.

When we hit around 100 users, we put a paywall so new users needed to pay up front before even trying the product. If they don't like the product, we would refund them immediately.

We used the paywall to identify the background of our core customers — those who desperately needed our product and were willing to pay, even when the product was still in its early stage. This allowed us to iterate the product toward product-market fit faster.

Since then, what we're doing is pretty much the same: iterate the product, talk to users, analyze the cohort retention chart, again iterate again.

Know the problem

The most important thing is to figure out what exact problem you want to solve and why you care about the problem. Because the biggest mistake is to create a solution first without knowing what the problem is.

Take Heptabase as an example. "Taking notes visually" is not a problem; "learning complex topics more effectively" is.

Once you figure out the problem, the next step is to launch quickly and talk to users to get a deeper understanding of the problem. Don't overthink it; whatever you launch first is unlikely to attract many users, which means you can iterate very quickly without worrying about scale.

You'll learn more about the problem from these iterations. And the more you understand the problem, the better you'll be at solving it and creating a product with strong word-of-mouth.

Make sure you have a long-term vision of how your product can create a huge positive impact after ten years, but at the same time, keep talking to users and refine your vision according to your current reality.

Rewatch YC's "How to Start a Startup" course whenever you feel lost. It's the best material out there. Nothing beats it.

What's next?

Our company only has one mission: To create a world where anyone can establish a deep understanding of anything. This is the singular most important thing we're working toward, whether we grow fast or slow.

There are a lot of things we want to build in 2025, including using Heptabase to communicate and collaborate with others on complex research, using Heptabase to better work with different file types, and, of course, integrating AI with our UX in a way that can help users learn complex topics more effectively.

Since we've been profitable, we don't have the pressure to raise future funds from VCs. Nothing can really stop us from building useful things for our users.

You can follow along on X, my personal website, and my blog. And check out Heptabase, our wiki, our Discord, or this tutorial.