[[đď¸đ§ vikash]]'s [[đ˛đ§Źbayes_evol]]
REV: APRIL 26, 2018
HONG LUO GARY P. PISANO HUAFENG YU
# Institutionalized Entrepreneurship: Flagship Pioneering
As he looked out onto a frozen Charles River from his Cambridge office, Noubar Afeyan contemplated the future of Flagship Pioneering, the firm he started 17 years earlier. Unlike a traditional venture capital firm, Flagship created and funded its own ventures based on science and intellectual property developed in-house. Flagship rarely invested in ventures started by independent entrepreneurs or third parties. Flagship was, in Afeyanâs words, a âfully integrated life sciences innovation enterprise.â A unique aspect of Flagshipâs strategy was its focus on investing in areas of novel and unique science. Afeyan explained: âLast year, we changed our name to Flagship Pioneering to reflect our philosophy. Like pioneering species in biology, we innovate only into spaces that are not occupied by others.â
Since its founding in 2000, Flagship launched more than 96 new ventures pursuing opportunities in therapeutics, health technologies, and sustainability (see Exhibit 1). Together, these companies generated an aggregated value exceeding $20 billion and conducted more than 50 clinical trials for novel therapeutic agents. Flagship and its ventures patented more than 500 inventions. As of September 2017, 27 of Flagshipâs portfolio companies had been acquired and 13 others had gone public. According to Afeyan, the firmâs financial performance across the previous decade (Funds IIIâV) stood at 35% net IRR (internal rate of return) with a multiple of 3.6 times the $800M of capital deployed.
Between December 2016 and December 2017, Flagship closed two new funds, bringing the overall capital under its management to $2.32 billion. 1 While Flagshipâs earlier funds had largely financed early-stage ventures, one of the two new fundsâthe $285M Opportunities Fund Iâwould invest in the growth stage of a selected number of Flagship portfolio ventures. Afeyan noted that investing in later stages of its ventures would create new challenges. âWe think we have developed a reproducible system for hatching new ventures. Now, we need to think about how our model can be applied to systematically taking ventures through the development and commercialization stage.â Afeyan believed that if Flagship could become as systematic in its approach to scaling ventures as it was in spawning them, the next 17 years of Flagshipâs existence would prove even more fruitful than its first.
## The Origins
After graduating from McGill University, Noubar Afeyan moved to the United States to undertake his Ph.D. in Biochemical Engineering at the Massachusetts Institute of Technology. Shortly after earning his Ph.D. in 1987, at the age of 24, Afeyan started PerSeptive Biosystems, a company that developed instrumentation systems for life science research. Over the next 10 years, he served as the CEO of PerSeptive and contributed to the founding of five other biotech companies, including EXACT Sciences and ChemGenics Pharmaceuticals (acquired by Millennium Pharmaceuticals).2
At the end of the 1990s, when all of those companies either went public or were acquired, Afeyan reflected upon his past decade of experience as a serial entrepreneur and contemplated the path forward. âEntrepreneurship,â Afeyan commented, âwas often viewed as an âartâ that relied more on luck and spunk, rather than on systematic processes.â He noted how the word âentrepreneurshipâ itself, with a suffix that commonly indicated a state or condition in the English language, failed to capture the reality of entrepreneurship as an active process. âIâm an engineer, but I donât do âengineership.â I do engineering because thatâs the ending youâll use when youâre doing something. I think once you can figure out how to do it, then the appropriate term should really be âentrepreneuring.ââ His thesis was that entrepreneuring was an activity that could be institutionalized.
In 1999, Afeyan started NewcoGen, short for âNew Company Generator,â to implement his vision of institutionalized entrepreneurship. Although Afeyan never intended to start a venture capital companyâcapital was merely an input in his mindâhe realized that investment was âthe only systematic activity in the startup world.â The way in which venture investment was institutionalized provided a reference point for his vision about parallel entrepreneurship: capital as a relevant input to fuel the internal innovation engine that attempted to systemize the life science venture-creation process. The company changed its name to Flagship Ventures in 2002 and added a managing partner, Ed Kania. Kania was a former OneLiberty Ventures general partner who had deep venture industry experience and had been an early investor in several of Afeyanâs previous startups.
## The Flagship Model
Flagshipâs strategy focused on creating ventures in domains unoccupied by others: âWe will not develop a company knowing there are other companies doing similar things,â Afeyan stressed. He further remarked, âMost innovations, even those by startups, are in âadjacent spaces;â that is, they are extensions to what already exists in a given market sector. For example, most biotech firms are started in relative close âproximityâ to where drug companies are already operating because they want partnerships with those companies. They join an existing technology, product or market; they do not create new categories.â
Flagshipâs âpioneering strategyâ came with tradeoffs and had implications for how they approached innovation. âThe good news is that in the science business, if you can be the first and create intellectual property, you can at least protect the value you created. The bad news is that thereâs nobody to copy. So, we literally have to figure out from the ground up, the strategy, the team, the right metrics of success in each case,â Afeyan commented.
Over the years, Flagship developed and refined a systematic, disciplined process for originating ventures in-house. Afeyan explained: âTo deal with âuncertainâ terrains, you need a âcertainâ process.â He made the analogy to evolution: âI know this works because this is how biological evolution works. The biological process of evolution is very rigid. Itâs the same process over and over, and yet it generates an extraordinary diversity of robust life forms. Thatâs how we approach the process of venture creation.â
### The Venture-Creation Process
Flagship operated a unique innovation and company-creation foundry called âVentureLabs,â where it originated ventures following a systematic process. The four-phase venture-creation process (see Exhibit 2) began with Exploration, during which a group of two to three associates (the majority of whom had a Ph.D. or M.D. degree), supervised by a partner, brainstormed about opportunities within emerging areas of science and tried to identify a âventure hypothesis.â A âventure hypothesisâ was a statement about the potential value of a venture within a new scientific space, under the assumption the science in question was feasible. A venture hypothesis was essentially imagined. It did not need to be constrained by whether there was existing scientific evidence or data in support of it. As Afeyan explained:
A venture hypothesis is different from a scientific hypothesis in that, initially, we do not care whether it is true or not. It is completely a product of our imagination drawn from asking âwhat if?â and âif onlyâ types of questions. We are interested in questions that point to hypothetical solutions that would be valuable to the world, once discovered. The claims of our venture hypothesis, however, need to be internally consistent. And, it must be the case that no one has done this before.
A notable example of such venture hypotheses came about ten years ago when Flagship recognized that people were beginning to see clusters of microbes in the human gut. Afeyan explained:
What got us interested in the human gut microbiome space back in 2008 were emerging reports of dramatically greater microbial diversity both within and across human guts than previously imagined, implying biologically important activities of potential medical importance. At the time, we didnât have the proof, but we made some assumptions that, if true, would allow you to think about the types of things that could be possible five to ten years later. And instead of waiting to convince ourselves that this is going to be, in fact, the case, we just started envisioning diverse companies and started to develop them even while we were still gathering functional data.
Among other hypotheses, we pursued the possibility of collecting human stool in order to diagnose various GI (gastrointestinal) diseases caused by microbes (rather than analyzing DNA from human cells to detect colon cancer, which was already being pursued). Another hypothesis involved natural products that we posited could be produced by microbes in the gut. These products could then control distal functions, specifically inflammation. Inflammatory bowel diseases or repeated use of antibiotics can also disrupt normal gut microbes, and we hypothesized that we could restore them by adding back a targeted subset of organisms. During 2008â2010 when these explorations first occurred, there was no evidence that any of these approaches could be realized scientifically, medically, or commercially.
Another feature that distinguished Flagshipâs approach was that venture hypotheses did not necessarily need to address a problem. Afeyan stressed: âThe hypothesized solution can be matched to a problem as well as a problem being matched to a solution. This approach goes both ways.â Partly as a result, the outcome was often a âmulti-product platformâ with which Flagship could diversify risk beyond a specific product aimed at specific disease areas. Returning to the microbiome example, Afeyan expounded:
> Over the ten years since we started to explore opportunities stemming from the emerging field of microbiome research, Flagship launched four distinct ventures, each based on internally generated innovations. These were: Seres Therapeutics, Indigo Agriculture, Evelo Biosciences, and Kaleido Biosciences. In aggregate, these companies commanded a value exceeding $2.5 billion. Adding to this opportunity landscape, two additional companies are now applying newer microbiome innovations to human health and agriculture.
During the exploration stage, Flagship relied on a large network of external collaboratorsâexperts and non-experts from both academia and industriesâto help select and improve the companyâs ideas. Afeyan pointed out that âone of the most abundant natural resources out there is peopleâs unlimited capacity to tell you how stupid your ideas are. And we are more than happy to tap into that resource to make our ideas better.â A great majority of Flagshipâs unconventional ideas failed under the selection pressure, and the ones that survived were often variations of the originally proposed ideas. There was no âlossâ during the processâoftentimes, new ideas germinated from the faults and failures of the old ones. âAs Darwinian evolution, it turns out that if you do variation and selection and iteration, the survival of the fittest will cause winners to emerge,â Afeyan explained, âand itâs really hard to predict a priori, over generations, which species will emerge. Thatâs our model for innovation.â Afeyan further noted that if sufficient selection pressure was applied, and enough rounds of iteration were undertaken, the ideas that emerged from this evolutionary process could represent significant, and often surprising, leaps forward rather than incremental improvements. This, in turn, would give Flagship a significant and durable advantage over competitors in establishing dominance across new areas of innovation.
In three to six months, after a venture hypothesis stopped receiving skepticism and disbelief from the external critics, it moved on to the second phase, called a ProtoCompany (standing for prototype company). Once a venture hypothesis became a ProtoCompany, the focus shifted to conducting experiments to validate, modify, or kill the initial idea. âGenerally, we would like the proof-of-concept to be achievable within roughly no more than a year and no more than a million dollars,â Afeyan indicated. At this phase, âkiller experimentsââphysical experiments that might expose the fatal flaws of the conceptâwere conducted, mostly in Flagshipâs own labs, which amounted to about 50K-60K square feet across Cambridge, MA.
âIn our experience, an individual entrepreneur whoâs absolutely committed to building a company tends to avoid the killer experiment, and [if] the data arenât good, theyâre out of a job,â noted Douglas Cole, a managing partner at Flagship. For Flagship, the incentive was reversed. It was critical to conduct these experiments early so that the ideas would either âsurvive earlyâ or âdie experimentally,â and, in turn, time and financial resources could be more effectively directed elsewhere. Team members at Flagship had an incentive to uncover the truth quickly because there were usually many other venture opportunities in which to be involved. One intriguing way in which Flagship sought to embed its âfail in order to succeedâ mindset was that ProtoCompanies were deliberately given numbers instead of names, because, as Afeyan explained, âit turns out itâs easier to discontinue a number than a name.â
The length of the prototyping phase varied greatly, ranging from a few months to a couple of years. With a capital support from $500K to $1.5M (typically convertible notes), a ProtoCompany started to assemble a team of talents with relevant scientific and operational backgrounds. At this phase, even though it was still owned, operated, and financed completely by Flagship, the ProtoCompany began to incorporate external scientific advisors, who largely remained on the scientific advisory board when the company officially launched. This was also the phase during which many of the firmâs patents were generated.
If a ProtoCompany survived the killer experiment phase, and the venture seemed promising, Flagship could designate it as a NewCo (new company), the third phase of the venture process. At the NewCo phase, the venture would have a business strategy and a product plan. It began assembling a larger team of 20 to 30 people, as well as a board of directors. A Flagship partner always served as the interim CEO. During the previous ten years, Flagship had increasingly become the exclusive source of capital throughout the NewCo phase, providing $10-20 million in equity financing. In the early phases of Flagshipâs development, this phase included not only Flagshipâs funds, but also small investments from other firms such as Arch Venture Partners, Polaris Partners, and Third Rock Ventures.
According to Afeyan, a minimal number of ideas that were generated during the Exploration phase made it to ProtoCompanies, whereas many ProtoCompanies, more than 50%, successfully evolved into NewCos. âSince we donât do a ProtoCompany until we firmly believe that it has real potential and value,â Afeyan remarked, âď¸âthe cases when [the ProtoCompany] failed are either the ones that do not pan out experimentally [as proven by experiments] or those that take too long in the ProtoCompany stage.â
At the fourth phase of Flagshipâs venture-creation process, the Venture phase, a Flagship company generally recruited a proven CEO from the outside, began to operate as a fully spun-out entity, and attracted significant capital from external sources. Some of the early Flagship founding team members stayed with the company or rotated back to other Flagship ventures. Despite attracting an additional $100 million or more in private capital, Flagship maintained a significant ownership stake in the company and continued to be involved in the development of its science and intellectual property. The former founding CEO tended to remain on board as the Chief Innovation Officer to âkeep the flame of the original vision.â Afeyan said that such âcontinued connectivity of an innovation driving forceâ between Flagship and the new venture was beneficial to both parties. In particular, it afforded these emerging companies an opportunity to continue drawing on insights and expertise from across the broader Flagship network.
In summary, Flagshipâs business model comprised originating first-in-category ventures based on internal innovations or licensed inventions from academia and developing and growing these ventures into best-in-category companies (see Exhibit 3). Over time, an increasing proportion of Flagshipâs portfolio companies were founded internally. According to Cole, Flagship had scaled up to launch six to eight companies a year, and about three quarters of them were internally founded. But even with externally originated ventures, Flagship became involved very early in shaping and forming the scientific and business concept behind the entity. âI think what is equally important is, how to conceive the idea in the biggest way,â Cole commented. Given the scientific discoveries, Flagship could often help reimagine the entire basis of the company with the potential to meet a significantly broader set of commercial opportunities and social impacts.
Collectively, Flagshipâs portfolio companies represented a substantial source of inventions that came to rival those of the largest pharmaceutical companies. Exhibit 4 compared the average number of PCT (Patent Cooperation Treaty) patent publications across the Flagship ecosystem (2015â2017) in relation to 2017 R&D spend with the equivalent numbers for a number of medium and large companies in the biopharmaceutical industry. Even though Flagshipâs aggregate R&D spending of $1.2 billion in 2017 was far smaller than that of a representative biopharma company, its patenting productivity was significantly higher (139 PCT publications per year for Flagship compared to 13-98 PCT publications per year for the other companies).
### Financing
Unlike the way it operates in all other respects, the way that Flagship raised funds was no different from other venture capital or private equity firms. A typical fund would receive investments from large institutions (e.g., pension funds and insurance companies) and high-net-worth individuals, commonly referred to as limited partners. The fund was managed by general partners, who financed companies across various stages of their development. Funds could exit investments and liquidate their stakes through the companyâs IPO and acquisition events. 3 Typically, the fund had a lifespan of 10 years; general partners made investments in the first five years and reserved the latter half for follow-up investment and exit. 4 General partners were typically compensated by an annual management fee (e.g., 2% of total committed capital at the beginning) and carried interest (e.g., 20% of the fundâs profits).5
As of December 2017, Flagship had successfully raised seven venture funds, totaling $2.32B in capital, including an Opportunity Fund that would be invested in a selected number of portfolio ventures in their growth stage 6 (see Exhibit 5). According to Preqin, out of 187 rounds of recorded investment made by Flagship for 66 of its portfolio companies, 64% were seed investments, series A, or series B (see Exhibit 6a). Exhibit 6b shows that for 76% of the portfolio companies, Flagship invested in fewer than three rounds (including seed investment).
Although the Flagship team was focused on success metrics concerning its innovations and company creations, Flagshipâs funds were among the top performers relative to all venture capital funds closed in the same vintage year. 7 Afeyan reported that as of mid-2017, each of its Funds IV and V was ranked in the top three among all venture funds of the same vintage. An important internal metric for funding efficiency was consistent with its approach towards conducting killer experiments early in the process. A dominant majority of companies that realized low or zero returns for Flagship were terminated when an investment of less than a million dollars was made.
The fact that Flagship was the founder as well as the main capital source of its companies meant that Flagship owned substantially larger shares of its portfolio firms than a typical VC fund. A typical VC fund would own 20â30% of an early-stage venture and eventually 10â20% at the time of an initial public offering (IPO). In contrast, Flagship typically owned 80â100% of the initial venture and 40â60% at the time of an IPO. This unique ownership position, which had been increasing in the last 10 years, contributed to Flagshipâs superior returns.
### People
By end of 2017, Flagship had a total of 43 full-time partners, principals, senior associates, and associates, 67% of whom had a Ph.D. or M.D. degree. Some of the associates remained with Flagship, focusing on venture creation (with a career path towards partnership), while others chose to work for portfolio companies. There appeared to be a fluid boundary between Flagship and its portfolio companies: people might work across multiple portfolio companies in different capacities and might even return to Flagship full-time afterwards. âThere is a set of skills required to make these companies move forward, and sometimes itâs only required for a defined period of time or only for part of the time,â Cole remarked. âOne of the ways that we get scale is by being able to deploy people across multiple companies for certain roles, for certain periods of time.â
In addition to a strong scientific background and genuine curiosity for science itself, the firm was also looking for people who had âthe right mix of creativity, humility and earned-confidence.â Afeyan further explained the notion of earned-confidence: âThey need to exhibit good judgment, and that comes through experience and is based on something that is actually real, versus, unearned-confidence, which is the kind of premature, stubborn confidence that, because this is your idea and you believe in it, you want to demonstrate resilience and persistence.â Afeyan referred to the ideal trait that Flagship sought in new hires as a âparanoid optimism.â âItâs a balance between paranoia, which means doubting things youâre saying and doing, and optimism, which means believing that your idea can work,â he explained.
Cole also indicated that, often, âthe scientific background can be somewhat anathema to imagination and creativity.â Therefore, the ability to âembrace very orthogonal ideasâ was critical to thriving at Flagship. Cole believed that, to an extent, such an ability could be learned. Afeyan described how, as outsiders to the fields in which they innovated, their people were initially unaware and later doubting of âdogma.â This, interestingly, allowed them to consider possibilities that others would have considered unreasonable and inappropriate. In fact, Afeyan and the Flagship team had come to believe that breakthrough, discontinuous innovations often arose from unreasonable starting points, while their improvements required reasonable steps.
## The Venture Fellows Program
Flagship initiated the Venture Fellows Program in 2008 as a way to expose graduate students in scientific or medical fields to its approach and provide them with a unique opportunity to participate in the earliest stages of Flagshipâs innovation pipeline. The three-month summer program recruited doctoral students, post-docs, and medical students. The 2017 cycle received over 400 applications for just 24 openings.
The program focused on the Exploration phase. During the three-month program, the fellows worked in groups of two to four people, with a partner and a couple of associates giving supervision and guidance to each team. âFor fellows, the hardest task was to come up with outrageous ideas,â Afeyan said, âDuring the program, they would have the chance to see some early ventures at various stages and get an overview of our venture-creation approach.â Throughout the program, Flagship also held lectures on different aspects of the Flagship business. By the end of each month, the teams would present their exploration results to all the partners and associates. âThe quality of the ideas they come up with, the depth of thinking that they convey, the clarity with which they can articulate these ideas is nothing short of extraordinary,â Cole remarked.
The idea behind the program was in line with Flagshipâs âinstitutionalized entrepreneurshipâ approach, whereby âone can, given the right culture and right approach, systematically generate radically innovative ideas.â According to Cole, Flagship benefited from this program in many ways: Apart from ideas that they could continue to work on, about 20â25% of the fellows could end up working at Flagship, and another 25â33% in one of Flagshipâs portfolio companies. More broadly, âWe want to be a part of broader activities and conversations that the sciences can potentially offer to the society,â said Cole, âAnd we also want to get other people excited about this approach.â
### Culture
The firm purposely nurtured a safe intellectual environment for the free exchange of ideas. During Exploration, in particular, every person, regardless of position and seniority, was encouraged to propose novel ideas. â[The culture] really encourages people to propose ideas that may appear outlandish, and [to] feel thatâs actually valued rather than criticized,â said Cole. âI think thatâs a very distinctive part of Flagship, and a very distinctive part of what has allowed us to be successful.â
With the majority of employees having a deep background in science, Flagship intended to cultivate a mindset with which people would resist the tendency to evaluate outlandish or seemingly impossible
ideas based exclusively on existing knowledge, but, instead, would mull over possibilities beyond the unknowns. In Coleâs words:
> Because almost by definition, when somebody proposes something that immediately sounds highly plausible to you, itâs incremental, itâs pretty close to what youâve already known is true. It only becomes interesting when itâs not the case. I want people who can listen to crazy ideas, and not immediately focus on all the reasons why that might not be true, but actually allow themselves to imagine what it will be like if only it were true, and to engage in thinking about what it would take to make that true.
### Looking Forward
As 2017 drew to a close, Afeyan was excited by the next phases of Flagshipâs growth and the challenges ahead. He felt that the company had learned how to institutionalize venture formation, but that it now needed to apply the same discipline and systematic approach to venture growth and scaleup. He summed up:
> The more companies we originate, the more we are forced to think thoroughly about the growth and development of these companies, which requires a completely different skillset than forming the company. Right now, once our ventures hit the development phase, we are really no different than anyone else. It is all about hoping that the right team shows up and the right execution happens. If you want to be systematic about value creation, you canât pull out halfway through this race.
Exhibit 1
![[flagship_pioneering 2025-04-13-4]]
Number of Portfolio Companies Launched by Year (96 total)
Exhibit 2
Flagshipâs Venture-Creation Process
Exhibit 3
The Flagship Model: Science-based Innovation and Professional Venture Creation
Exhibit 4 Flagshipâs Patenting Productivity: Three-year Average (2015â2017) Number of PCT Patent Publications per $1B R&D Spend in 2017
Three-year average PCT (Patent Cooperation Treaty) counts from Derwent Innovation WIPO Applications database; and R&D spending from company reports for 2017.
Exhibit 5
Fund
Flagshipâs Capital Pools
Vintage
Closed Size ($Mn)
Flagship Ventures Fund I Flagship Ventures Fund II Flagship Ventures Fund III Flagship Ventures Fund IV Flagship Ventures Fund V Flagship Opportunities Fund Flagship Pioneering Fund VI
Exhibit 6a
Round
Flagshipâs Investment Rounds by Stage
Number
Percentage
Seed Series A/Round 1 Series B/Round 2 Series C/Round 3 Series D/Round 4 Series E/Round 5 Series F/Round 6 Venture Debt Growth Capital/Expansion Unspecified Round Total
32
42
46
27
13
6
2
1
2
16
187
17.1%
22.5%
24.6%
14.4%
7.0%
3.2%
1.1%
0.5%
1.1%
8.6%
100.0%
Source: Note:
Preqin, accessed July 24, 2017.
Based on investment-round data made by Flagship available in Preqin for 66 of the portfolio companies shown on the website of Flagship Pioneering.
Exhibit 6b
Total Number of Investment Rounds in A Portfolio Company Made by Flagship
Total Investment Rounds
1
2
3
4
5
6
Number
Percentage
14
18
18
3
7
6
66
21.2%
27.3%
27.3%
4.6%
10.6%
9.1%
Total
100.0%
Source: Note:
12
Preqin, accessed July 24, 2017.
Based on investment-round data made by Flagship available in Preqin for 66 of the portfolio companies shown on the website of Flagship Pioneering. The number of rounds includes seed investment.
Endnotes
1
Ben Adams, âFlagship raises $285M special ops fund, changes name,â FierceBiotech, December 15, 2016, accessed August 15, 2017 http://www.fiercebiotech.com/biotech/flagship-raises-285m-special-ops-fund-changes-name-again.
2
Wikipedia contributors, âNoubar Afeyan,â Wikipedia, The Free Encyclopedia, accessed September 9, 2017, https://en.wikipedia.org/w/index.php?title=Noubar_Afeyan&oldid=808150442.
3
Anthony Gambardella, âIBISWorld Industry Report 52391: Venture Capital & Principal Trading in the US,â accessed August 3, 2017. IBISWorld.
4
Metrick, Andrew, and Ayako Yasuda. âThe economics of private equity funds.â The Review of Financial Studies 23, no. 6 (2010): 2303â2341.
5
Ibid.
6
Max Stendahl, âCambridge VC firm Flagship targets $500M for latest fund,â Boston Business Journal, September 15, 2017, accessed November 18, 2017 https://www.bizjournals.com/boston/news/2017/09/15/cambridge-vc-firm-flagship-targets500m-for-latest.html.
7
Preqin, the Internal Rate of Return statistics were available for Flagship Fund 1 and 3, and they were among the 1 st quartile (top) and 2 nd quartile, respectively, compared to all venture funds closed in the same vintage year, accessed August 22, 2017.
13
This document is authorized for educator review use only by SCOTT STERN, Massachusetts Institute of Technology (MIT) until Mar 2025. Copying or posting is an infringement of copyright.
[email protected] or 617.783.7860