Whatever venture capitalist Bryan Roberts is doing, it’s working. Roberts, a general partner with Venrock, has had nine of his portfolio companies reach a valuation of a billion dollars or more. Those investments were in digital health (AthenaHealth), genomics (Illumina) and biotech (Receptos). Now Roberts focuses mainly on digital health. Read my questions and his responses from our January 11 fireside chat at Digital Medicine Showcase in San Francisco. On January 25, Venrock closed on a new $450 million fund.
Bryan Roberts. Photo courtesy of Venrock
Steve Dickman: Independent of returns, which of your investments have been the most satisfying to you so far?
Bryan Roberts: I get the most satisfaction out of doing things that, when we start with them, nobody else likes. That no one else thinks will work. They don’t like the people, the idea. Nothing. It’s not just that the timing is off. Regardless of timing, they believe it’s a silly idea forever. And I feel like I spend years going, “I don’t know why everyone in the world doesn’t love this.” But it takes a long time. And then they sort of flip. Other investors start to like them. That is the part I like.
Dickman: Have you gotten to the point where you are using data science and machine learning to help you make investments?
Roberts: We do not use it at all in choosing deals. For the most part, our investments are made in things where there are way more variables than there are equations. You are making decisions in the face of ambiguity that is pretty pervasive. So you can’t actually use data or even diligence to get yourself to an endpoint. By contrast, data science has become indispensable for companies doing product development. It’s like the electrical grid for them.
Dickman: You have always had an amazing network. What do you learn from big companies in pharma or digital health that you see as potential acquirers of the ventures you invest in?
Roberts: I do not look at big companies that way. I have tended to focus on orthogonal solutions to what I think are big problems. My thesis on most big companies is, what they will consider acquiring in five years will be different from what they would consider acquiring today. Either it will be different people or a different strategy or something. I don’t know what they will consider acquiring but it will be different. There is actually no “signal value” for me in what they like today, except on the investments that I made five years ago. I used to come to this conference and spend lots of time with those folks and I do not do that anymore. It’s the portfolio companies that have to spend time with those folks, not me. Because in my thinking I need to be out five years.
Dickman: What drives your investment decisions in pharma and biotech and in digital health?
Roberts: For me, those opportunities are chosen by entrepreneurs I get attached to intellectually and they are about an orthogonal approach to a big problem. Back in the day, Sirna Therapeutics [then called Ribozyme Pharmaceuticals] was trading at $6M market cap because it was about to go out of business. And on its old business model it should have gone out of business. That was at the same time as when its competitor Alnylam was doing its series B financing. We actually looked at both of them and made a decision to invest in Sirna. There was a huge price difference–Sirna’s price was much lower. The thing that was interesting to us about Sirna was, they had done 18-24 months of work in the space of RNA interference. And they had intellectual property and manufacturing. It was that that drove our decision more than, say, what Merck was thinking about at that time.
Dickman: In around 2012, you kept doing some biotech but you really dove into digital health. Was that the right move?
Roberts: I am hugely intellectually selfish. I work on stuff that is interesting to me from an intellectual perspective. The sea changes going on in how people pay for and organize healthcare have been fascinating to me. We’ll see whether in 15 years whether it is as robustly successful an ecosystem as biotech or genomics.
Dickman: But in a venture context, do you really have 15 years? Don’t you have to exit in less than 10 years?
Roberts: I sit on boards for an average of about 12 years. We may have a slightly longer time horizon. That’s both good and bad. I think you’ll find that health IT businesses have shorter cycle times than life sciences businesses, absent super-hot (or -cold) markets. From how long it takes for product development and then commercialization until you see if you actually have a differentiated product in the marketplace, my bet is that the cycle times will be materially shorter for health IT businesses than for drug discovery, medical device or early-stage technology businesses.
Dickman: Is that how it is working out with your portfolio companies Doctor on Demand, Grand Rounds? Those short cycle times?
Roberts: The life sciences biotech companies have a journey of seven to 10 years of product development where they plow a huge capital hole in the ground and then they come screaming out if they are successful. In digital health, it’s different. I would call Doctor on Demand, Grand Rounds and some of our other digital health companies “software and services businesses,” since they have a big services component. If these companies initially succeed, they will spend a couple of years doing what I call “hard” product development. After that, they generally have a first-generation product. That product has value to individual customers who are willing to pay for it. Then they spend two to three years building up their competencies and improving the product features. This leads to a revenue stream. But that is just the first phase. At that point, if their product is selling, they will have a group of customers and users. What they can then do with that ecosystem of customers and users that they have created represents something interestingly larger.
You have then created a network of customers and a bunch of data. There is a very small number of businesses across these industries that sort of spring fully formed from the head of Zeus, right? Workday is one example. They raised a lot of money and just made this huge, enormous software product. Boom, off we go. Most other ones find a really needy niche and then, if they are successful in penetrating that niche, okay, now we’ve got these people who are our users. We’ve got this data and then we can do something more with these users. An example of that is Flatiron Health in New York. Google is an investor. When the cofounders started the business, their motto was “better information for oncology treatment.” That meant that their customer was a single doctor who has a single patient who can benefit dramatically from that. That’s the first phase. For the second phase, once you get a whole bunch of doctors and patients making better decisions based on Flatiron’s software, all of a sudden you’ve got a data business. You can start selling that data to pharma.
Dickman: You said you mostly are starting your own companies in digital health these days. If an AthenaHealth sought investment when it already had $10 million in revenues or even $1 million in revenues, it would not be available to you as a venture investment. The price would be too high…
Roberts: Correct. We have done pretty much all seed and Series A businesses over the last 4-5 years. These investments are by and large the product of a series of conversations and the work we do with an entrepreneur who we think is terrific. So I don't really consider that we started the business, but hopefully we had a hand it getting it to an interesting place.
Dickman: Doesn’t investing that early mean you have to take much more risk in anticipating what the product is and how the market will adopt it?
Roberts: When my portfolio company Castlight Health went public and the stock was trading at about $48 a share, I went up to the company and gave a talk and said, you guys, just know that it will be single digits at some point. They were like, “No, are you kidding me? The stock is at like $40 or $50! Are you crazy?” But it always happens. [Castlight’s stock price was recently at $3.50 a share.] It happened at Illumina and at Zeltiq, just to name a couple.
Castlight was the first investment I have been involved with that very quickly changed the discussion in an industry. When we started the business, venture capitalists did not want to invest in it. Their first product was offering employees of large corporations the ability to understand what pricing would be for a variety of medical services. I did not know this before we started Castlight but within a stone’s throw of here, if you want to go get an MRI, there would be a 4x difference in price for no particular reason at all. It was about uncovering price discrepancies which were rampant across the industry.
When we started the business, people were like, you will never get the data. None of these prices were known! Very quickly, about 18 to 24 months in, there was an article every week about pricing discrepancies in healthcare distribution. That was very different than most of the businesses I invest in where you slog away and nobody cares about you for a super-long time. Here it was suddenly in the spotlight. And then you have to figure out, how do I build a business under that? And what are all the different pieces of it?
Dickman: So if you are not investing in existing companies, what are you actually doing when you make early-stage investments? Where is Venrock’s value-add?
Roberts: People talk about company formation like it’s the Big Bang theory of the universe. In fact, it’s a very time-intensive, wandering path. When we started Castlight, Todd Park and I were on a float in the middle of a lake in New Hampshire saying, “Oh, yeah, we should definitely start a company.” Well, that doesn’t really count as founding a company. That counts as two jokers passing time. The notion was, oh, yeah, AthenaHealth was “the web and healthcare and doctors,” what if we did “the web and healthcare and individuals”? That’s worth about a nickel. Then there is nine months of, "How are you going to make any money? How are you going to go to market?" We spend a lot of time helping people work through that sort of stuff.
Some of the reasons not to do later-stage investments is less us and is as much, there is so much money out there that an entrepreneur who is just looking for dollars isn’t going to come to us because they can get it somewhere else at what is in all likelihood a substantially lower cost of capital.
Dickman: It seems to me that there is no Google of digital health. So maybe there is no natural acquirer for businesses you start in digital health, even when they start to work.
Roberts: You are talking to a guy whose firm started making biotech investments in the late 1970s and early 1980s. So we invested in companies like Centocor and Idec. Idec later merged into Biogen which is now a huge company 30-odd years later. But if you get Bill Rastetter, the former CEO of Idec, up here, he will say, “I had to get someone to throw me a lifeline, a buck a share to stay in business.” Back then, pharma companies were more chemical companies and were not buying anything like Idec. I think you are frameshifted by two decades or something in the maturity of these two industries.
That said, you have an interesting set of people in the ecosystem for healthcare delivery for digital health. Health systems, insurance companies, and then you’ve got a small number of large healthcare IT companies: Epic, Cerner. AthenaHealth is maybe an adolescent one of those. What tech has and by extension digital health does not have is that those large companies–Google, Facebook, Apple, Microsoft–have really awesome EBITDA margins. Those companies mint money. One of the issues for digital health companies both from a customer and an acquirer perspective is that most of the businesses–the various customers, the health systems or whatever–don’t make any money. So the value that one’s product needs to bring to customers must be very concrete and short-term in its impact on the customer’s income statement. There is no room for products that might be helpful several years down the line.
Dickman: Do you typically invest in serial entrepreneurs?
Roberts: Something north of 95% of the investments I make are in companies that have first-time CEOs. Almost never do I invest in a second-time CEO. I think that some percentage of people who have been successful before are really good and really smart and continue to be really driven. And then there is a whole swath of people who were at a successful place but the success did not have much to do with them. Then you have people who were really good…and then stuff happened. Like they got rich. All sorts of stuff.
The world generally overestimates repeat CEOs and repeat entrepreneurs and underestimates first-time CEOs and first-time entrepreneurs. All of my concern about wanting to re-invent the wheel from an intellectual perspective every time I look at something, that is generally because it is unclear to me that prior experience is actually a positive or a negative for looking at any specific thing today.
Dickman: You told FORBES in 2014 that you “operate with a very healthy dose of paranoia–always trying to do new things, because the old things have been done.” Can you elaborate?
Roberts: Most of my paranoia comes from worrying that, as I do this longer, I am not getting smarter and knowing more. I worry that the pattern recognition that all of us end up relying on to some degree in our lives is actually constraining, not enabling. Because stuff that worked before won’t necessarily work this time. And stuff that didn’t work before won’t necessarily fail this time. I feel like business models in our business are really overrated.
N.B. Roberts’ responses were edited for clarity. A video link to the entire chat can be found here courtesy of EBD Group.
The Billion-Dollar Exits
Roberts is one of a very few venture capitalists who have successfully invested both in biopharmaceuticals and in digital health. He is a member of FORBES’ Midas List. Roberts’ biggest exits (in order of initial investment) include:
- Illumina: Led Series B at a valuation below $10 million in 1998; recent market cap $23.6 billion
- AthenaHealth: Led first institutional round in 1999; recent market cap $4.3 billion
- Ironwood Pharmaceuticals: Participated in Series A in 1999; recent market cap $2 billion
- Xenoport: led Series A in 2000; achieved market cap of $1.4 billion in 2007
- Sirna Therapeutics: led PIPE recap at $6 million pre-money in 2003; sold for $1.1. billion in 2006
- Ikaria: co-led Series A in 2005; partially acquired by Madison, Dearborn, a PE firm, for $1.6 billion in 2013; fully acquired by Mallinckrodt for $2.3 billion in 2015
- Castlight Health: led seed in 2008; company reached a valuation of $3 billion in 2014
- Zeltiq: led Series C in 2008; recent market cap $1.7 billion
- Receptos: led Series A in 2009; acquired by Celgene in 2015 for $7.2 billion
Roberts has nine current board seats at Venrock portfolio companies that include Doctor on Demand, Grand Rounds and Kyruus in digital health; 10X Genomics in genomic sequence analysis; Achaogen in antibiotics; and Intarcia in drug-device development. He continues to invest actively.
