Sramana Mitra: What is the fund size of rocketship.vc?
Anand Rajaraman: We are on fund two. It’s $120 million. Fund one was a $40 million fund. That’s fully deployed. We started investing fund one around late 2015. We have about 30 odd companies in the portfolio. We’ve just started deploying fund two.
Sramana Mitra: What is the stage focus?
Anand Rajaraman: We’re not focused so much on stage. We’re much more focused on where the company is in its evolution. The sweet spot is when a company has found product-market fit but before it has achieved massive scale. This might correspond to a Series A in some cases, pre-Series A in some cases, or even a Series B.
Sramana Mitra: Let’s dig down a bit on how you define and quantify product-market fit.
Anand Rajaraman: It would help to understand how we find these companies in the first place. We’ve built a big network. That network refers interesting companies to us. We talk to the companies and we invest a small fraction.
What we do at rocketship is somewhat different. We’ve built a huge database of startup activity all over the world. This database has a few million companies in pretty much every geography in the world. We collect a lot of data about each company. Then we run machine learning models on this data.
One of these machine learning models show us companies that seem to have product-market fit. What we mean is that the market is receptive to the product, and there are some early indications that customers are adopting the product. It’s different for a consumer company versus a B2B SaaS company.
For a consumer company, we may not care about monetization. For a B2B SaaS company, we may care about LTV. It’s hard to precisely define it, but you’ll know it when you see it.
Sramana Mitra: You’ll know when you see it, but how would your algorithm see it?
We have companies in our portfolio that have one customer to five customers. They’re in the early product-market fit phase. Nothing is available about these customers in public. How would your algorithm find these companies?
Anand Rajaraman: Our algorithms use a variety of things like social media and what people are saying. It includes data on employees and other investors to build proxy models. When there’s some activity, we track that as well.
We don’t claim to be 100% accurate. I don’t think we catch every company, and not all companies we catch have product-market fit. At some level, we have a filter that surfaces a lot of interesting companies. We talk to them and invest in some of them.
The nice thing about the filter that we use is that it runs in our network. At some level, we are not limited to companies that are referred to us by our network. We don’t have a particularly good network in Brazil, but we have investments in Brazil. We have investments in Europe. We have investments in Asia and North America.