The “startup versus management” myth rests on a substitution of terms: in it, the word “management” silently means only the management of predictable domains—established processes in which one categorizes a task and applies a ready-made method. A startup is indeed incompatible with such management, but that is only one meaning of the word. Uncertainty does not require abandoning managerial discipline; it requires the discipline of another domain—an experimental one: decisions through controlled experiments, small bets, and retrospective sense-making (in Cynefin terms, work in the complex domain). It is important not to overstate the claim: it argues for the necessity of a different kind of discipline, not that experiments guarantee success.
Eric Ries gives the claim its correct form: “Entrepreneurship is a kind of management.”
The myth has a second, independent refutation—institutional: by definition, a startup is an organization, a group of people (A startup is defined by uncertainty, not size), and an organization must be built and managed. “We do not need management” contradicts the very definition of a startup, even before the argument from uncertainty.
Two meanings of the term, whose conflation sustains the myth:
- Management₁ — administration of the predictable: established rules, categorization of the situation, and application of a ready-made method (the clear and complicated Cynefin domains).
- Management₂ — the discipline of working with uncertainty: proposing hypotheses, controlled experiments, small bets, and retrospective sense-making (the complex domain). Ries’s Build-Measure-Learn is the protocol of precisely this domain; a startup’s early stage without customers is a chaotic domain, from which action (an MVP) moves the situation into the complex one.
Management₁ also has its own definition of failure: failure to deliver the stated result is either a planning failure or an execution failure; there is no third option, and in either case someone is at fault. Under uncertainty, the dichotomy is no longer exhaustive: an experiment may be flawlessly planned and executed—and still “fail” by disproving the hypothesis. This is not blame but a product: in Management₂, such failure is a necessary step from which knowledge is extracted (for a revaluation of the price of failure, see A startup exists to learn how to build a sustainable business). The final scene of chapter 1 shows the collision of the frameworks in human form: after an unsuccessful experiment, the intrapreneur expects punishment because the CFO lives within the “failure = someone’s fault” framework—the myth in its corporate guise, in action.
The cost of the myth is symmetrical on both sides:
- A startup that believes the myth gets anarchy—[ET] the opposition “often serves as cover for chaos in processes (we’re a startup; we don’t need management).”
- A corporation that believes the myth stifles innovation by applying processes for the predictable to the complex domain.
The same pattern—“management as a cycle of hypotheses”—at other scales:
- Individual: Strategizing (ШСМ)—“a continuous cycle of proposing, testing, and updating hypotheses” about the target system, method, and project.
- Organizational, before Ries: The Fifth Discipline—a learning organization based on the principles of analysis, experimentation, and openness; The Reflective Practitioner, by Donald Schon—practice as experiments with reframing.
- Organizational, after Ries: Organization 3.0—the next reinvented form of management (decisions → goals → agent prompts).
A typical mistake is to “combine chaos with management.” For Ries, this is not a blend of old management and chaos, but an entirely new discipline.
How this claim supports Startup success can be designed—the right process can be learned
In two ways:
- It removes the main objection. The claim that startup success can be designed presupposes that managerial discipline is applicable to a startup at all. The “startup versus management” myth supplies a ready-made counterargument: if a startup is anti-process by nature, then no process can create its success, and there is nothing to learn. This claim exposes the counterargument as a substitution of terms: a startup is incompatible only with the management of the predictable, not with managerial discipline as such.
- It clarifies which process can be learned. Without this claim, the “right process” in the claim about designed success remains an empty space, easily filled with processes from the wrong domain (planning, regulations—and the result is cargo cult). Here, the “right process” receives its type: the experimental discipline of the complex domain—hypotheses, small bets, and controlled experiments.