“Our thesis, at Hatcher+, is that people should invest in venture the same way they are investing in stock market or hedge fund,” John said when we visited him in his office in Singapore. John is a serial entrepreneur, investor, and the architect behind the firm’s proprietary research and technology platform. With the use of artificial intelligence and machine learning, Hatcher+ is set out to redefine the practice of venture investing.
It was not without cause that Hatcher+ arrived at such bold vision for the venture capital community.
Looking for an algorithmic way to run a VC company
“We did our first fund (Hatcher, launched in 2015 in Singapore) and we invested in 13 companies. we made 21 investment into these 13 companies, and we ended up with eight surviving companies that are still alive today. All of these companies are funded, exited, or in good shape. We are happy with the progress. At the end of that fund, however, we realised that we spent probably 70-80% of our time on the companies that fail and maybe 20-30% of our time on the company that succeeded. So, we thought this is completely backward and we are determined to be more disciplined about how we approach this in the future,” John explained.
“That set us on a two-year journey to start understanding what are the most successful strategies in venture capital. so we went all around the world, and we met with venture capital firms in Shanghai, Tokyo, San Franscisco, London, Paris, and Dubai. We ask everyone the same questions — How do you run your company? What processes do you use? What strategies do you use? What are the most successful ways to run a venture capital company? And we discovered that there weren’t too many very clear answers. So we decided to do a lot of research. We looked over 300,000 investment events over the last 15 years and we tried to see if there are any algorithmic way we can run a venture capital company.”
At the end of the two year, the search was concluded with four key concepts, and these concepts have helped shape Hatcher+ as we know it today.
1. Invest early
“You have to invest early. When you are investing early, you are usually investing at a valuation that are over-discounted relative to the risk. The return from the very early stage investing can be very high — it can be double digits. it can even a be triple digits return in some geographies. Investing early is essential if you want to make money out of venture. if you invest, for example, in post Series B or Series C companies, what we have found is that the return from those investments track the stock market. So you might as well invest in NASDAQ as opposed to investing in Series C venture companies. So if you are going into venture is very smart to go in early,” John revealed the first conclusion of the firm’s research.
2. Have a large portfolio
“Large portfolios like Y Combinator, Sequoia, and 500 Startups, they do better as a general rule than small companies. there are some outstanding small portfolios but as a general rule, if you have a large portfolio the return is better,” John stressed the importance of having a large and diversified portfolio.
“It is actually an application of modern portfolio theory, which was invented in the late 50s. It was applied to the stock market in the early 90s. It is never been applied to venture capital. So we are actually taking modern portfolio theory and applying it to venture capital. The simple explanation of portfolio theory is that — when you have a diversified portfolio, some companies will fail but the successes of other companies will outplay the effects of those company failing. Imagine if you have only four portfolio companies and three fails, that a tragedy right? If you have 300 companies and 100 of those fails you still got a lot of companies in there you can potentially make up the difference.”
3. Get to the top 1% quality startups faster
“The third thing that we discovered is that on average accelerators and VC companies accept only 1% of the business plans that come through the door. So only 10 out of every one thousand business plans are accepted and only four of those get funding. The actual rate of business plans successfully make it into venture capital companies are extremely small. That was also quite surprising to us,” John would soon explain how Hatcher+ is using their system, DART, to accelerate the screening process to arrive at the 1% more effectively and efficiently.
“We are using AI and machine learning to help us filter through those thousand of business plans and find the 1%. that is part of our special sauce and what we bring to the equation. So we team up with accelerators, we give them this technology and we co-invest based on the results.”
DART stands for Diligence, Accuracy, Responsiveness and Transparency. With DART, Hatcher+ is simplifying the due diligence process and help improve the investment decision-making process.
“What we are really looking for are indicators that the founders really have domain expertise and solid knowledge of the area. For example, if you want to start a new streaming media company, we will look at evidence that you know how to create content in that area, we look at the kind of terminology that someone would use coming from the TV background or streaming media background. The same thing with medical devices, if someone said in their business plan that they have an amazing new technology, we remove those words (amazing and other adjectives). There is no data for us in those words. If someone said we have an FDA approved prosthetic device that has 19 degrees of movement and it is computer-controlled and powered by fuel cell, we are going to pick up from 90% of those words and they are going to give us an indication that this is something that would fit one of our special accelerator and it is a candidate for investment,” John described how DART works.
Among other things that the DART system analyses are founders’ understanding of market, scalability, financial performance, product delivery, and even founders’ relationship.
In order to arrive at that 1%, Hatcher+ needs a massive network. “Currently there are about 4.5K accelerators operating around the world. We look at about 400 of them, and we chose 40 that we want to approach and work with. For us, the most important thing looking for an accelerator is some form of specialisation. We already have deals in place with some very highly specialised accelerators. What we find fascinating working with these accelerators is that they generally have fantastic mentors, very strong backing from corporate venture capital group. And there is a specialist from that industry. So when they look at 1,000 business plans and pick ten. It is a high chance that those ten are going to lead us into that industry because the process has been so well informed by the people around it. If you take our platform and put it into that environment and get rid of 80% of the stuffs that does not deserve to be focus on and then hand the resulting 200 business plans to specialists that are good at what they do, the combination of those two things is going to work really, really well to create better quality investment.”
4. Tradability among VC fund units
“The fourth things we found is that most venture fund last for more than 15 years. For many investors, it is extremely unattractive having to lock up your money for such a long time. What we have been working on is the strategy to allow people to trade in and out of the units in the fund. If you are an investor, you can get into a large portfolio venture companies and get out again after a couple of years, as opposed to waiting for 15 years. We think that the combination of those four elements is quite revolutionary,” John said.
Bringing the VC industry up to speed
“We have seen an awful lot of support from the VC community, in fact, overwhelmingly positive support. I would say no one has been critical about the platform, some people have been skeptical about it, which is understandable because for everything new you are going to find skeptics. But when we show them the platform, we show them how we are actually judging the submission and the kind of algorithm we are using, the kind of data we are leaning on, people become supportive very quickly.”
“In a typical VC firm, there will probably be two or three people looking at all the business plans. If you tell them you don’t have to look at 1,000 business plan, we are going to take 800 of those and remove them from the fennels so that you can focus on the 200 best scoring, best quality business plans. That, by definition, is going to raise the quality of the output of those people and help them do a better job. So, what we are looking to do is not really to replace people with our AI and machine learning. We are looking to focus the human resources on better quality submissions, and better quality processes. Everything we are about now is introducing the system into an industry that largely rely on manual processes. 99% of people still doing manual due diligence.”
“A lot of time when you ask people, why do you need eight weeks to do due diligence on these super early stage startups, they can’t give you a good answer. And I asked —Why isn’t it eight minutes? Why isn’t it eight hours? Why isn’t it consists of just one meal with the founder, which they tell you their story and you decide or not? There are not a lot of data available for the super early stage companies. Yet people insist on doing weeks of due diligence. Sometimes, it is detrimental to both the companies and the value of the investment. Sometimes, it is better to act quickly, and get in behind an idea you think that is merit based on a simple understanding of what the risk are.”
“For latter stage’s due diligence, you probably have to go and do a little bit more. The reality is — if you are following a good performance monitoring programme, that should not really be necessary. If you have performance monitoring in place early at the accelerator stage, it should in theory become easier to do due diligence in the latter stage. Right now the problem is that most companies get to the Series A round, they don’t have financial statements, they don’t have written product plans, they don’t have very good HR processes or records of any type. The investors in that stage is forced to go into a lot of manual due diligence. What we help accelerator do is to pass on the data profiles from those early stage investors to Series A Investors, and enable them to see in its entirety what that business has been doing and get a better handle in their performance,” John expressed that with the data-driven model, it will make due diligence much easier at all stages of venture investment.
How soon can venture investing start resembling the stock market?
“It is going to be much, much faster than anyone think. When we look right now at the venture investment industry, it resembles the retail stock broking companies of the late 1980s, where you will buy a few hundred or thousand shares or stock through some retail guy in the retail environment. What happened quickly after that was that people move quickly within 10 years to investing into essentially the blind pool and this pool will trade stock between each other at a speed of four micro seconds and react to market information in a speed of also four micro seconds. So the stock market went from a very retail, human-driven environment to being a very much computer-driven trading environment based on 100% or close to 100% availability of information.
“We think that what we are doing in venture is going to lead to far greater availability of information which will create the condition for much faster and more active trading of private company stocks. This is the very beginning of a transformation of private company equity trading, which right now secondary trading is at a very nascent level but we predict in five years this is going to be very active market of secondary trading in company stock. You can already see it around the world, companies popping up everywhere that are starting to trade in private company stock. We think that is going to be a massive trend going forward. The more we can do to push data into that environment the better this exchange will function,” John concluded.
Having partnered with 16 best in class accelerators spanning all continents, Hatcher+ is planning to start co-investing with them in Q1 2018. Hatcher+ is currently raising a US$125 million fund. With this fund, they are going to invest up to US$250,000 in each startup. “We don’t plan on making Series A or later investment in companies. On rare occasions, we may make a Series A investment but typically we are going to invest in companies the earliest possible that we can. We just think that it represents really great opportunity, and that is when the companies most need the money, and that is when we can generate the most return. It is win-win for both of us.”