Wednesday, June 3, 2015

Do Venture Capitalists Value Patents?

This is a simple, but important question. Do venture capitalists value patents? You would think the answer is an easy yes based on survey data, as well as my own findings from the Kauffman Firm Survey that firms with patents are about ten times as likely to have venture capital funding.

But I get pushback on this. See this TechDirt post, for example, called: No, You Don't Need Patents to Raise Money:
While some of them are filing for their own patents, a key point was that their investors definitely didn't require it or push them in that direction.
None said their investors had pushed them to file for patents.
When I speak with people who espouse this view, and tell them of my 10x finding, the response is almost always: "Well, that's just people getting patents after they have money, or IP firms telling them to do it."

So, here we have an apparent conflict between stated preferences and revealed preferences. On my to do list for two or three years now has been a study of all startups, with an examination of who got patents and when. But now I don't have to, because Celia Lerman (a Fulbright Scholar at Stanford, among other things) has done the study, called Patent Strategies of Technology Startups: An Empirical Study:
How does a patent strategy affect a tech startup company’s growth? This is a fundamental question for technology entrepreneurs, investors, lawyers and the innovation system as a whole. In this study, I shed light on this issue by conducting an empirical analysis of the patenting strategies of technology startups, examining the relationship between a company’s patent applications and different events over the company’s life: rounds of investment received, company acquisition and closure. I provide the first comprehensive cross-industry analysis of this question, by analyzing the patent portfolios of United States startups listed in CrunchBase, a crowd-sourced registry of tech companies used by the startup industry. By looking into these companies’ public patent applications from the United States Patent and Trademark Office (USPTO) database between 2008 and 2012, I examine the patenting patterns of startups as they progress through funding rounds.
Through a quantitative analysis, I find that companies based in California tend to patent more than in other states, and that companies that are venture-backed patent more than those who are not. I also unveil that most start-ups that patent file their first application before even receiving any reported funding. Moreover, I find that there is a significant positive relationship between patent protection, and receiving investment and being acquired. I further find that the number of patents (and not merely the fact that a company has patents or not) contributes to higher total funding. I finally observe that patenting early is also associated to higher funding, and that early may be more important for start-ups than what some views in venture capital may predict. I also conclude that while more patents are associated with higher funding, patents account for a relevant but small portion of a company’s success.
The study provides novel insights on startup patenting strategies. It lays empirical groundwork on key circumstances under which patents can contribute to a startup’s growth, to provide important guidance to the legal and entrepreneurial communities.
The study finds that startups patent before their first funding round, from a low of 50% in software, to 64% in IT/Hardware and 67% in medical. It also finds that firms with patents are funded more often and for more money. More discussion on this after the jump.

Lerman performs a regression on the amount of funding (and also the amount of funding in total) based on the number of patents in each round. The study finds that a) having patents leads to significant increase in the amount of lifetime funding, and b) having patents before the first round leads to increase in the amount of funding in that round (and also in future rounds).

These are both useful results. They answer the stated/revealed preferences problem: while people say you don't need to have patents (and you don't - many firms, especially in software, got funding without patents), the money seems to move toward the patents, rather than cause the creation of the patents.

But that said, the paper left me wanting a bit more - at least to answer my binary question. What is the increase in odds of getting funding if you patent before the first round versus not? Or if you don't patent until your second round? That question remains unanswered, and the descriptive data doesn't give us a clear answer. On the one hand, most firms with patents got them before funding. On the other hand, most firms don't get any patents at all. Further, many firms get no funding at all (and the paper has scant details on that).

To answer these questions, at the very least I would love to see a crosstab of funding versus early round patenting to see what portion of the funding moved to the early patentees v. not. I suspect it's a mixed bag in software, not so much in other industries, which is why we are seeing such divided lines in patent reform. Even better, it would be great to see a logistic regression with funding (yes or no) as the dependent variable and patenting and other variables as the independents. I don't think amount of funding gets us there, because so many firms got zero funding or had zero patents.

Even without this extension, though, the data gathered in this paper will be extremely useful for understanding startup patenting activity on the ground, and I look forward to future analysis.