What were the key factors that enabled TraderPlan to overcome its scaling challenges? What unfair advantage did we have? How do you balance your investment in technology and marketing? I’ll address these questions in this post.
Firstly, how did I end up working for TraderPlan? In 2017, I was one of their users, and I also helped them architect a blockchain-based investment vehicle. One day, the Founder approached me, and asked, “how much can I pay you?”.
If I wasn’t discovered by TraderPlan, then I would have continued building my own software company, and I wouldn’t have gotten very far on my own. And TraderPlan might not have launched as quick as it did. Since January 2018, I’ve been working as the lead software engineer of TraderPlan. It was the perfect opportunity for me to gain experience in starting a FinTech company without making all the mistakes myself.
TraderPlan began as one person’s side-project. Today, we have a team of 10, with plenty of fans in the crypto community. We’re still as young as a baby. That’s why, when I was recently asked, “How did TraderPlan overcome it’s scaling challenges”? I was hesistant to share it. I didn’t believe I had much value to share, considering we’re still a small team.
However, I still think there is value to share the process you’re going through, rather than waiting after it’s finished. The short story of how we moved from MVP to a production application that could serve a wide user-base can boil down to 2 things:
1. Hiring the right people committed to the long term vision
2. Keeping the right balance of marketing and technology
The longer story is that we had an advantage. The Founder had consulting clients already, and he made video courses on YouTube. That’s how many of the early users came onboard. He validated and built the MVP based on what his consulting clients wanted. But when we reached 20 users, we needed to migrate the MVP (that was hosted on a single server) onto the cloud. That required refactoring much of the codebase into microservices.
We hired talent across the world, rather than limiting ourselves to a specific location. That allowed us to tap into the best talent at an affordable price. It was also helpful to have another software architect (besides the Founder) who was a coder, data scientist, accountant and trader combined. Having people with multi-disciplinary skills kept us moving forward at rapid pace and provided a big runway. It allowed us to self-organize and minimize management overhead. It helped a lot to be in the crypto and blockchain community as well. The community seemed to attract very smart people to the same place and make them easier to find (and hire).
After we re-architected the software to run on the cloud, we could infinitely scale our computing resources on demand. However, that wasn’t really true, because as an automated trading platform, we needed to send API calls to various exchanges. The problem is that some exchanges throttle how many API calls you can make each minute, and we quickly surpassed that limit, resulting in our server IPs getting banned. So, we had to change the way we did API calls.
The fact that bugs in a trading platform can cause people to lose money immediately meant that we could not iterate as fast as we’d like. If it’s just organizing information or a chat application, we could always rollback if it didn’t work out. But when your software is working with people’s money, there’s little room for mistakes. To minimize bugs in production, unit tests would not be enough. Since we were users ourselves, we would test against our own portfolio. If there’s a critical bug, you might hear one of the developers sigh, “Great! The bug just cost me $10k. That’s my vacation. Thanks.” At least that’s better than letting the bug affect all users.
We didn’t do much marketing, as engineering was the biggest challenge after we validated the MVP. Users were signing up without us even trying. We were actually hoping that people wouldn’t find out about us, until we were ready. So, staying in stealth mode was a better strategy. Now that our technological capability has caught up, it’s time to grow our market. We will grow until we hit another engineering roadblock, which is when we slow down on marketing until engineering catches up again.
UPDATE 26-Jan-2019: I left TraderPlan in July 2019 to pursue a better opportunity.