More Brains

We all know the current VC structure used to fund cleantech companies is broken. One major aspect of this is the more rigorous technical challenges of the cleantech world. The VC structure, formed by funding tech companies, places very little emphasis on the idea and, as such, just isn't capable of accurately assessing the technological challenges and potential of cleantech companies.

If this is true, biotech VCs should face the same challenge as biotech startups face at least similar levels of scientific and engineering intricacies. Looking into the status of biotech VC funding, the trend is starkly similar to cleantech. Looking to move their model into new areas, VCs went crazy in the biotech field and now are scaling back as their startups fail to meet returns under the tech paradigm.

How can VCs change their structure to accommodate cleantech and biotech? The best place to start is at the beginning: pick better companies to fund. To do this VCs need to re-prioritize what they look for in startups. But even with a higher priority on the idea VCs will be in a tight spot. When being exposed to  hundreds or even thousands of ideas a year, how can a VC with possibly no technical background make the right call?

Lets look at Khosla Ventures, they have a portfolio of more then 100 companies and only a little over 20 employees. To reach a portfolio of 100 companies they surely looked at more then 1000 companies. 20 people is simply not enough people to fully vet so many businesses when giving real preference to the idea over the team and go-to-market. If you are a member of the investment team responsible for energy storage and you have 20 different battery chemistries in front of you, how can decide which to fund based on the technology? Let alone if you should be looking at battery, thermal, pressure, or hydro storage.

If you want to maintain an average 10X return in cleantech or biotech you need a method that does the following:

1. is able to screen a huge number of idea

2. is able to consistently select technically feasible ideas with real-world applications

3. is able to champion ideas with the potential of yielding required returns

The current system is setup around taking a few very smart individuals and bombarding them with ideas. It is able to accomplish 1, it can sometimes accomplish 2, and utterly fails to achieve 3. In short, in cleantech, you need to bring more brains into the process. 

You could take the government grant review approach of farming the technical review out to experts in the field. This certainly would give you feedback on each technology. Unfortunately, the experts are likely to steer you towards the technologies least likely to give you the returns you are looking for. If, as an investor, you want to invest in the incremental technologies most unlikely to meet 10X returns, take this approach. 

In my mind, ARPA-E got very close to the perfect model, before completely botching it. There, they take a fesh crop of the best and brightest students for only a few years to serve as technical fellows. The fellows are, theoretically, responsible for informing the technical consciousness of ARPA-E. What's new, different, and about the make a big splash? The fellows are, again theoretically, perfectly poised for this task because they are non-experts. They are fresh out of school, come from a diverse background, don't come with any biases, and can champion ideas they think will make a big impact. 

If this system is so perfect, why does ARPA-E still have so many problems picking incremental boring projects (see most recent ARPA-E storage SBIR). It goes back to the theoretical statements about the fellows. 

First off, if you have ever been in an ARPA-E meeting, you know that the fellows and lower level employees do very little informing at ARPA-E. They are more like interns used by the program managers. The program managers drive the direction of the funding and the selection of the grants, completely negating any unbiased input the fellows and lower level employees could have introduced. 

On top of that, I doubt the fellows could perform their theoretical task even given the opportunity because of how they are picked. To start, the fellows are required to have a PHD, guaranteeing they come in with a tremendous set of biases built up after 5-7 years defending the existence of their thesis. Furthermore, they are hand picked from the top research universities like Berkeley, Caltech, MIT, and Cornell. Having been through one of these schools myself, I can tell you first hand that those programs do not necessarily produce the most opened minded scientists. In the end, the fellows are just as biased as the program directors and the entire program suffers. 

So, how can we fix this system? I think there are a couple steps to taking ARPA-E's model and turning it into the most effective way of picking better cleantech investments. 

1. Bring more brains - At most VC firms, there are one or two people that make the real decision to invest in any given area. Sure there may be a group of partners that make a joint "call", but it really comes down to one or two partners who are the champions of one technical area. At ARPA-E there are theoretically more brains looking at ideas, but in actuality it is more like a VC firm, the program managers make the call. When looking at a huge swath of idea, you cant trust just one brain. You need 10-20 non-experts churning through them. 

2. Forbid bias -  It sounds easy, but its not. People are biased. The best way to get rid of it is to build your team around young intelligent non-experts. The best way to do this is hire students directly from undergrad. PHDs and MSs build experts. That's great for companies and universities that need experts to develop technology, but you need non-experts to find disruptive technologies. If you have chosen the right people, they can learn how quantum tunneling works as they go, they can't unlearn bias. 

3. Diversity - Dont hire the best and brightest from MIT, Berkeley, and Caltech. Hire the best and the brightest from all over the country / world. The top tier schools dont produce the most open minded people and dont have a monopoly on smart people. Go to the unknown schools. Build you team with members from outside science and technology. Include non-traditional backgrounds, and educations. 

4. Let the crowd rule - If the team has the right composition, it will be much smarter together then you every could hope to be. Trust it. An anecdote: once in a business class I was taking, a speaker came to talk about the wisdom of crowds. At the beginning of his talk he passed around a large jar full of marbles and asked us all to write down on a note how many marbles we thought were in the jar. Like the arrogant engineer I was I thought "these business people dont stand a chance, I can just calculate it!". When the numbers were all collected, the professor reported the findings: "if we disregard the one outlier, the average of the class is within 5% of the real answer!". He then explained that most of the guesses were about 10-20% above or below the answer, but one was more than 60% above. This value was of course my calculated value. This was a great lesson showed me three things: a) I am almost always wrong, b) arrogance and technical expertise are almost always worthless and c) crowds of the right individuals produce better results then the individuals alone. 

5. Focus on turnover - In abiding by 2, you can't allow the team to exist forever. That means after 2-4 years, your team members should be looking for an exit. As the facilitator and creator of the team you have a responsibility to make sure the membership on the team was a stepping stone in a career. 

6. Make the money work - This one is easy. Since everybody on the team is contributing, everybody should be compensated equally. Nobody on earth needs to make as much money as VC partners do. Share it and focus on successfully championing winning technologies. 

7. Kill the pitch - Do not allow entrepreneurs to pitch to you under any circumstance. Don't invest in a company because they got in someones network. Systematically pick companies based on idea before team and team before go-to-market. This necessitates a new type of review process based on many casual conversations without pre-rehearsed pitches, slogans, or simplifications. 

8. Remove social constructs - Dont let anybody ever use Occam's Razor or the Devil's Advocate. These social constructs dont help find disruptive plays. Encourage critical examination of ideas and the search for simpler explanations, but don't let imaginary social constructs decide what companies you invest in or not. Remember that all disruptive technologies seem complicated when they are first created.

As a 27 year old with a 10 month old pre-seed phase startup, you are probably saying "shut up". Thats fair. But think of it this way, we know non-experts are better at finding disruptive improvements to systems. And we know experts are better at finding incremental improvements to systems they understand very well. Does cleantech just need a few incremental tweaks or a disruptive overhall? I think the latter.