This year I celebrated my 10-year work anniversary with Atlas. Hard to believe how fast a decade flies by, and this milestone triggered some reflections on how times have changed in biotech and the early stage venture business.
As we all know, the biopharma industry has changed a lot. Wyeth, Schering-Plough, and Genzyme are now found primarily in history books. Once small-cap biotechs are now big players: Gilead’s market capitalization was $19B in 2005, its now $150B; Celgene’s was $7.5B, now $90B; Biogen’s was $14B, now $65B; and Regeneron’s was only $500M, now nearly $60B (100x increase!). The concept of immuno-oncology was only discussed on the fringes on the sector, along with gene and cell therapy. On the finance side, over the last few years, we’ve seen unprecedented access to the public equity markets. We’ve also seen new investors enter the mix (e.g., Alaska Permanent Fund) and the expansion of “cross-over” players. Preclinical M&A deals, like CoStim and Flexus, are more common today than a decade ago. Many, many other things could be noted.
Atlas has changed a ton, too. We were investing in both biotech and tech, had four offices, and a couple dozen partners when I joined, as I noted before (here, here). We’re now tightly focused on early stage biotech with our five-person partnership working out of our one office in Cambridge. All significant changes for the better.
Instead of talking about those more obvious changes, I’d like to reflect on a more nuanced change in the venture business – the evolution of how new biotech startups get formed and financed, also known as the venture creation process.
At least five rather striking changes in this process have occurred over the past decade:
- Entrepreneur-centric vs VC-collaborative startup formation. Back in 2005, it was common practice for startup entrepreneurs to come and “pitch” their NewCo to VC firms looking for Series A capital. The founders would have assembled a story, team, and strategy, and often some “friends-and-family” seed funding, and then would shop their proposal around in the venture community to see which firms would raise their hand to invest. Most startups would find standalone offices or labs to rent during this process. In that vintage the majority of young biotechs received their Series A funding in this manner. Fast forward to 2015 and it’s fair to say that the prior “on your own” entrepreneur-centric approach is less common, at least for early stage venture firms. A more frequent path today is for a venture firm, working with a talented pool of entrepreneurs-in-residence, to scan through great science or technological concepts that emerge out of academia and to push forward to form a company around it. The new startup is most often incubated or housed within the venture firm’s offices in its early days while it pulls itself together, often with VCs or their teams taking on acting operating roles until a seasoned team can be recruited. Other startups are co-incubating along side and a “startup community” within the firm develops. We’ve co-founded or incubated all but a few of the startups we’ve invested in over the past five years. In general, this “VC-collaborative” startup process is how Atlas, Third Rock Ventures, Flagship, and a few other early stage firms build the majority of their portfolios today. Traditional “standalone” startup funding approaches still exist and drive a meaningful number of financings (especially behind accomplished industry exec’s/entrepreneurs, e.g., Jeremy Levin at Ovid and Tony Coles at Yumanity), but as a whole this change in the venture creation process feels very real in the ecosystem today and has significant implications for how entrepreneurs engage with VCs.
- Willingness of early stage investors to go-it-alone without syndication. Back in 2005, the Atlas portfolio had a number of deals where we owned 15-20% and were one of a handful of independent venture firms in an early stage deal. This was typical of Series A financings, and having five or more venture investors involved by the end of the Series B round was common. This approach helped share the risk and bring more “expertise” and networks to the table; but it also diluted each firms’ returns, and occasionally complicated governance and decision-making during the formative early years of a startup. Although syndication still happens frequently today, there’s increasingly a trend toward “going it alone” in the seed/Series A rounds, or at the very least reducing the size of the syndicate of investors. Of course, forgoing syndication means that a firm needs both lots of dry powder (bigger funds) but also lots of intestinal fortitude to navigate, alone, the eventual bumps in the road. Only hindsight determines whether not syndicating a particular deal is “rational” brilliance or “unlucky” hubris. In many ways, though, cutting back on syndication helps solve for the Venture Capital Math Problem. If you only own 15-20% of your exits, getting to an attractive 3x+ overall fund return is very difficult, even in a bull market like the one we’ve been in. But having >40% ownership positions at an exit makes it very doable, as evidenced by some recent huge winners (e.g., Flagship owned 44% of Seres Therapeutics and TRV owned 47% of Sage Therapeutics after their IPOs). Beyond solving the fund math challenge, less syndication also means less contentious governance (as it’s harder to argue with yourself); however, the flip side to this for startup entrepreneurs is that there frequently aren’t other investor voices in the board room to create some balance. It will be interesting to watch this trend as it evolves. Another wrinkle in this syndication change is the increased role of corporate venture capital; as discussed on this blog previously (here, here), where today’s small syndicates are constructed they frequently involve corporate equity investors – a distinct difference from a decade ago.
- More focus on the science and creating innovative new therapeutics, less on markets and models. In my first year at Atlas, I saw a handful of (and built a few) spreadsheets with market models forecasting the sales of a startup’s potential drug – typically out in the distant future. Those models, of course, are always wrong and not actually very helpful as they are rooted with cognitive biases (especially against markets that don’t exist yet). I haven’t employed or seen a model like that in years. There are several reasons for this. First, these models were often for spec pharm reformulations or repurposing opportunities, or investments into “early stage companies with later stage assets”. These “low risk” products were expected to have shorter paths to market, though in reality often faced regulatory and differentiation challenges (here). Atlas hasn’t done a deal like that since 2006. We, like some other early stage VCs, focus on finding innovative science and medicine around which to nucleate a future great company; this patient-centric and science-driven investing approach is by definition geared to high impact innovation and is years from product approval. Most of our deals are drug discovery or preclinical at the time we make our initial investment decisions, so granular market models have too little accuracy and too much false precision to be either relevant or valuable – so we just don’t bother with them. No drug discovery stage startup needs to have a market model for what its sales forecast will be in the distant future; instead they need to have a credible thesis for how they will address an important medical issue in a transformative way, differentiated from other approaches, and a path for progressively derisking that opportunity. That’s fundamentally a scientific and clinical question: back great medicine and the returns (and the future Excel model) will follow.
- Shift from refining formal “business plans” towards practical hands-on business building. Very related to the point above, the classic MBA program emphasizes the value of the startup business plan. That might be very helpful in tech deals, but it’s just not as valuable in R&D-driven early stage therapeutics companies, where the specific scientific and clinical hypotheses are the more appropriate focus. We don’t need a formal business plan (like that described here) to evaluate or get excited about a deal. I used to see them all the time a decade ago – typically as an eager MBA worked with a team of scientists and insisted on crafting a long-form business plan. This is very rarely useful (largely a waste of time), especially in a world of VC-collaborative venture creation as noted above, and often detracts from the time spent on more crucial details of the R&D and operational strategy. We always write comprehensive investment memo’s for our internal process, but we don’t need a “business plan” in place to launch a new startup; instead, it’s more common today to see us focus on practical R&D-stage business-building with hands on engagement from a team of partners and EIRs with complementary skills.
- Geography matters even more today than before. In 2005, we had portfolio companies spread all over the US and Europe, with one in Japan. The concept of being hundreds if not thousands of miles away from a portfolio company wasn’t typically considered an issue. In the early stage arena – the funding a new startups – this has changed dramatically: more than 90% of the startups we fund are not only in this region, but most are in Cambridge. Sending them to Alewife or Watertown requires a “going away” party. I know the same holds true for a number of other well known early stage VCs; they also like to invest in startups that are nearby so they can engage and interact on a real-time basis with them. The consequence of this is clear: its harder for entrepreneurs to attract seed/Series A venture funding from Cambridge- or San Francisco-based firms outside of those geographies. The same is true for building talented teams; many of the best entrepreneurs increasingly want to work as near as possible to these top biotech hubs. We will often identify great researchers in other geographies (e.g., Dario Campania is based in Singapore and was founder of Unum; Paul Thompson was at Scripps Florida when we founded Padlock with him), but these often end up as local startups here in Massachusetts with research relationships back to their founders’ labs. A decade ago we were more flexible in putting companies in other geographies, and over time we’ve decided that flexibility had more frictional costs than benefits. This trend isn’t likely to abate: the gravity of the deeper, stronger clusters will continue to aggregate more talent, more capital, and more startup formation activity relative to other geographies.
These elements all combine to make the startup environment for funding and launching new biotech companies very different today than it was a decade ago. Startup formation tends to coalesce more around specific early stage venture firms, in hubs like Cambridge and San Francisco, via a collaborative, interactive, and incubation-oriented startup approach.
I’m sure the models will continue to evolve. As the bull market continues its long-term advance, and capital hopefully continues to flow robustly into the sector, at some point the trickle-down effect into the startup community may create a new set of forces, geographic pressures, and venture formation models. It will be interesting to see what the next decade brings.