The defining decisions of early-stage life are rarely made on evidence; more often, they are made on something that arrives before evidence is possible. An investor commits capital to a company that is still mostly a description of itself. A founder commits years to a market that does not yet answer back. The official vocabulary for how such decisions ought to be made is analytical: theses, diligence, models, frameworks. The vocabulary for how they are actually made is smaller and slightly embarrassed, and most of it is carried by a single word, intuition, usually offered as an apology. Around that word the discussion tends to stall at a single question, whether intuition should be trusted at all, and the stalling obscures something more useful. Investors and founders do rely on intuition, constantly and unavoidably. The question worth examining is in what ways, and what kind of knowledge is at work in each.
How these decisions are supposed to be made is well documented. Investors maintain theses, run diligence, build models, and benchmark against comparables; founders are urged to be data-driven, to test before building, to let metrics arbitrate. This is the official epistemology of the ecosystem, and it exists for honourable reasons: money and years are at stake, memory flatters, bias is real, and a decision that can be written down can be examined, challenged, and improved. Taken together, these practices describe a world in which knowledge means analysis, and analysis precedes commitment.
The limitation of this account is not that it is false, but that it describes how decisions are justified rather than how they are made. Watch the practice instead of the paperwork and a different sequence appears: the judgment tends to arrive early and whole, and the analysis arrives afterwards, to test it, dress it, or decline it. This is not a corruption of the method, it is the method, and it means intuition is not a flourish at the edge of these professions but a load-bearing part of both. The honest question is not whether investors and founders rely on intuition as knowledge, but where, and this shifts the discussion out of the language of legitimacy and into the quieter register of practice.
For investors, the reliance begins with people. What is assessed in a first meeting is character under uncertainty: how a founder thinks when pushed off the script, whether a team's disagreement is generative or rehearsed, the distance between the deck and the conversation. Nothing on this list has a metric, yet experienced investors read it with the assurance of a clinician reading a patient, and the reading is knowledge of a particular kind: perceptual, trained on hundreds of prior encounters, arriving as a verdict before it can produce reasons. A second reliance is pattern. Anyone who has sat across from a few thousand companies carries a compressed record of them, not as recallable cases but as a sensitivity, a sense of which shapes tend to survive; this is induction stored below the level of speech. A third is timing. Data can only describe the past, and early-stage investing is a judgment about the present, about whether a market is ready now rather than five years ago or five years hence; the feel for now has no source except immersion in the current moment. Beneath all of these sits a quieter reliance still: attention itself. Which introduction becomes a meeting, which anomaly gets pulled on, when diligence has said enough. Analysis can only examine what intuition has already selected, and it stops when intuition says it may.
For founders, the reliance runs deeper, because a founder's material is a thing that does not yet exist. The first form is knowledge of the customer, acquired not by asking but by dwelling. A founder who has lived inside a problem knows it by acquaintance, the way a resident knows a city; a survey knows it by description, the way a map does. These are not the same knowledge, and only the first can tell which of the customer's stated wants deserve to be ignored. The second is taste, the feel for what the product wants to become: which feature belongs, which detail matters, which request would dilute the thing into a committee's idea of itself. Taste is knowledge of coherence, and coherence cannot be specified in advance, only recognised on arrival. The third is timing from the inside: when to launch, when to persist, when to pivot, decisions that must be made at precisely the moment the data is most ambiguous, since by the time the numbers settle, the moment has passed. The last is the plainest. Founders operate where precedent ends, and for genuinely new things there is no base rate to consult and no comparable to borrow; intuition is not one instrument among several there, it is the only instrument that reaches.
It is important to note that naming these reliances does not place intuition beyond question, it only locates it. Each of the ways described is knowledge earned through exposure, and each can be counterfeited by confidence that skipped the exposure; a feel is only as good as what it was trained on. But that is a question of quality, not of category. In every case above, intuition is doing what knowledge does: discriminating, directing, deciding, in territory where explicit method cannot yet operate. Seen through this lens, the well-argued memo loses its status as proof of judgment. An investor's read of a founder, or a founder's read of a customer, may be more reliable than any model either could defend in writing, precisely because it was trained where models cannot go. The question, then, is no longer whether intuition counts as knowledge, but why articulacy is so often mistaken for it.