16 · Why a venture builder exists

The innovator's dilemma. Chapter 04 described Christensen's mechanism from the market side; here it has to be taken up from the organisational side, because that is where the failure lives. Established firms do not miss discontinuities out of stupidity: they miss them out of local rationality. Their resource allocation processes, built by listening to their best customers and pursuing the highest margins, correctly classify as unattractive the segments from which disruptions start; and when the discontinuity threatens the business model, the organisation has active interests in holding it back. The case condensed into a date: in 1975 Steven Sasson, a Kodak engineer, builds the first digital camera in the company's laboratories; the documented reaction of the management was not to push a technology that would have cannibalised film, the product on which the company earned. Thirty years later, digital technology developed by others did exactly that job, without asking permission. The organisational lesson: discontinuities need vehicles whose interest is to make them succeed, not to contain them; inside the organisation they have reason to threaten, that vehicle does not exist.

Exploration and exploitation. James March (1991) formalised the trade-off that runs through this essay: organisations distribute resources between the exploration of the new (uncertain, distant, diffuse outcomes) and the exploitation of the known (certain, close, appropriable outcomes), and systems that learn drift structurally towards the second, because exploitation returns rapid positive signals that reinforce the choice, while exploration returns above all initial failures that discourage it. The result is the competency trap: you get better and better at what matters less and less. The answer studied by the literature is organisational ambidexterity (O'Reilly and Tushman): separating structurally the units that explore from those that exploit, protecting the former with rules, metrics and horizons of their own, integrated only at the top. Ambidexterity works but is fragile within a single firm, because the protection depends on the continuing will of the leadership, and every cycle of difficult results calls it into question: the robust form is to give exploration an organisation whose mission is to explore, with an economic model built on that.

Open innovation. Chesbrough (2003) documented the premise that makes any internal laboratory insufficient: useful knowledge is distributed. Qualified people move, risk capital finances those who leave, universities and small companies produce growing fractions of the relevant knowledge; no organisation, however large, has within its own boundaries more than a minor part of what it would need. Open firms use inbound flows (licences, acquisitions, collaborations) and outbound ones (disposals, spin-offs) to innovate beyond their own boundaries; but openness requires a dedicated capacity for absorption and a structure made to receive: external problems, technologies and competencies enter only where there are doors, valuation criteria and contracts to welcome them.

The law of the minimum. The last failure concerns new companies, and it is the most everyday one. Chapter 01 defined the firm as an integrator of knowledge, capital, labour and relationships; the typical startup is born with one or two of these dimensions strong (usually the technology, sometimes the relationship with a market) and all the others to be built along the way. The agronomic principle of Sprengel and Liebig describes what happens: growth is determined not by the sum of the available nutrients but by the scarcest one; in a barrel of unequal staves, the water runs out at the shortest stave, however tall the others are. A business project behaves in the same way: it performs according to the competency that is missing (the legal expertise that was not there at the moment of the wrong contract, the tax expertise ignored until the penalty, the channel never built), not according to the average of those that abound. Startup mortality, read with this lens, is to a significant extent mortality from the short stave: and the short stave is predictable, because the functions of a company are known in advance.

That the systematic construction of companies can work at scale and in the most difficult sector is demonstrated by the documented case of Flagship Pioneering, the Cambridge (Massachusetts) organisation founded by Noubar Afeyan. Flagship does not receive business plans and does not select other people's startups: it generates its own with a declared internal process, which starts from systematic explorations of scientific hypotheses ("what if...?"), submits them to rapid validation in its own laboratories (the internal "proto-companies", financed with its own capital and killed in the majority of cases: the funnel of chapter 03 and the stops of chapter 07 applied to the letter), and turns into companies only the hypotheses that survive, providing them with a team, initial capital and a platform of common functions. The portfolio thus generated numbers dozens of biotechnology companies; the best known is Moderna, founded in 2010 within this process around the hypothesis of therapeutic mRNA, that is, exactly the technology that the case in chapter 09 showed reaching the world in 2020. The case serves two purposes. It shows the four properties in action: dedicated vehicles for hypotheses that no pharmaceutical incumbent would have cultivated with that patience, exploration as a mission with an economics based on shareholdings, science imported systematically from outside, a common platform of functions and talent. And it fixes the correct statistical expectation: even in institutionalised venture building most explorations die early and by design; the model does not promise to abolish the mortality of innovation, it promises to move it to where it costs little (the early phases, chapter 02) and to give the survivors a starting point that isolated startups do not have.

• J.G. March, "Exploration and Exploitation in Organizational Learning" (Organization Science, 1991): the article that gave the trade-off its language; dense, short, foundational.
• C.A. O'Reilly and M.L. Tushman, Lead and Disrupt (2016): organisational ambidexterity with the cases, including the conditions in which it fails.
• H.W. Chesbrough, Open Innovation (2003): the paradigm of permeable boundaries and its empirical premises.
• On Flagship Pioneering: N. Afeyan and G.P. Pisano, "What Evolution Can Teach Us About Innovation" (Harvard Business Review, 2021), which describes the process from the inside, together with the HBS case on the company.