We use cookies and other tracking technologies to improve your browsing experience on our site, analyze site traffic, and understand where our audience is coming from. To find out more, please read our privacy policy.

By choosing 'I Accept', you consent to our use of cookies and other tracking technologies.

We use cookies and other tracking technologies to improve your browsing experience on our site, analyze site traffic, and understand where our audience is coming from. To find out more, please read our privacy policy.

By choosing 'I Accept', you consent to our use of cookies and other tracking technologies. Less

We use cookies and other tracking technologies... More

Login or register
to apply for this job!

Login or register to start contributing with an article!

Login or register
to see more jobs from this company!

Login or register
to boost this post!

Show some love to the author of this blog by giving their post some rocket fuel 🚀.

Login or register to search for your ideal job!

Login or register to start working on this issue!

Engineers who find a new job through Blockchain Works average a 15% increase in salary 🚀

Blog hero image

How to scale 16x in 6 months on Ethereum

Anson Chu 30 July, 2018 (3 min read)

How to scale 16x in 6 months on Ethereum

Startup. Big vision. Fast moving industry. Product market fit. It’s time to scale. I lead the engineering team at Numerai. This is the story of how we grew our data science tournament, and how we scaled NMR to become the most used staking token on Ethereum.

Pick one key metric and set an aggressive target

Before you work on scaling, it is important to know what to scale. Ask yourself this question: what is the one thing I want my team to focus on in the next six months?

Monthly active users (Facebook)? Trips per week (Uber)? Transactions per second (Ethereum)? If you are not sure what metric to focus on, start with this lecture by Alex Schultz, VP of Growth at Facebook.

Numerai is a quant hedge fund controlled by a network of data scientists competing in a weekly machine learning tournament. The goal of the data scientists is to predict future movements in the global equities market. To win money from our weekly prize pool, data scientists have to make accurate predictions and stake Numeraire (NMR) on their own predictions.

The game theory behind staking and payouts helps Numerai 
gauge the confidence of predictions and elegantly mitigates 
sybil attacks — **slyfox**

The one key metric we chose to scale at Numerai over the past six months is the number of stakes per week. More stakes means more data scientists making predictions with confidence. More stakes means more signal for our meta-model to use in trading.

When I joined the company in January 2018, NMR was barely six months old and we had 59 stakes per week. We set out to 10x stakes per week by the end of June. It was an aggressive goal — within the realm of possibility but just barely. It forced us to think big. It forced us to focus.

If you want to learn more about setting goals and focusing your team’s efforts, I highly recommend reading about OKRs in Radical Focus.

Find and eliminate your bottlenecks systematically

You have your one key metric and have set an aggressive goal. Now it’s time to move the needle. But where do you start?

Any improvements made anywhere besides the bottleneck are 
an illusion. Astonishing, but true! Any improvement made
after the bottleneck is useless because it will always remain 
starved, waiting for the work from the bottleneck. And any
improvements made before the bottleneck merely results
in more inventory piling up at the bottleneck. — 
Gene Kim, The Phoenix Project

The first bottleneck was painfully obvious. The staking prize pool for round 93 was $6000, but the non-staking prize pool was $37180 (2000 NMR * $18.59/NMR on February 3). What started out as an on-ramp for new data scientists turned into the main focus. In round 94 we consolidated both USD and NMR payouts into a single prize pool for staking. Now there was only one game to play. Stakes were up ~2x at 143.

The second bottleneck we tackled was new user growth. With the non-staking on-ramp gone, it became difficult for new users to start staking since they had nowhere to get their first NMR to stake with. Further, while our website was functional, it wasn’t exactly easy to use or understand for new users. In March, we wrote a comprehensive tutorial and launched two airdrops for students and Kaggle users respectively. In April, we completely redesigned our website with a focus on new user experience. By round 110, stakes were up ~5x at 289.

While the new user base grew, the team shifted focus on a new dimension of staking growth — stakes per user. With an active and growing base of data scientists, we were confident that we could multiply the signal we get from each user by asking them to make predictions on multiple targets. In round 111 we launched multiple tournaments, stakes were up ~16x at 973.

Conclusion

NMR image 2.png

We have made a lot of progress in the past six months, but we are only just getting started. We have even bigger plans for the next six months. Stay tuned.

NMR image 3.JPG

Originally published on medium.com