Most of the work of building an institution is invisible from the outside, and much of it is invisible from the inside as well. That is not a design flaw. It is the job.
A company can be built by a small number of loud decisions, made well, in a short window. An institution cannot. An institution is what remains when the loud decisions have faded and the practices they installed continue to compound in their absence.
The practices that compound
**A cadence the organisation defends.** Not a cadence the founder defends. The organisation. The tell is what happens in the week the founder is unreachable — the rhythm holds, or it doesn't.
**A decision rights map the room can articulate.** Every person in the room can name the decisions that belong to them, the decisions that belong to someone else, and the decisions the room makes together. When no one can articulate this cleanly, the institution is still, quietly, a company.
**A postmortem practice with teeth.** The postmortem produces operating changes, not sympathy. Six quarters in, the operating changes have compounded into a materially different organisation. That compounding is what an institution is.
A company is what the founder built. An institution is what the founder can leave without the room noticing for a month.
The tension
Institution-building is at odds with the incentives of most fast-growing organisations. It is slow to show up in the metrics that matter to funders. It is invisible in the launch calendar. Its most important work is often refusal — refusing to break the cadence for the announcement, refusing to bypass the decision rights for the deadline, refusing to spare the postmortem when the launch went well.
The organisations that get through that tension — usually because a small number of senior operators are willing to hold the line quietly for years — end up as institutions. The ones that do not remain, at best, well-run companies. The distinction only becomes obvious in the decade the founder leaves.
Where to start
Pick one practice — cadence, decision rights, postmortem — and defend it for four consecutive quarters, publicly, without renegotiation. That is not a change programme. That is the change.