CONCEPT
Day-to-day decision frameworks for founders who have become operators.
Operating a business is what begins once the thing works. Founding is a search problem: you are looking for a transaction someone will repeat. Operating is a control problem: you have found it, and the job is now to run it at volume without letting it degrade. The two require opposite instincts. Search rewards variance and cheap wrong answers. Control suppresses variance and makes wrong answers expensive. Most competent founders struggle at the transition not because operating is harder, but because the behaviour that got them here is now the thing damaging the business.
Founders bottleneck because every decision routes back to them. The usual response is to delegate tasks, which does not help — the decisions still come back, now with a delay. What actually creates capacity is delegating decision rights inside a stated boundary.
For every recurring decision, three things have to be written down: who decides, what constraint they are deciding within, and what triggers escalation. Discounting authority up to a stated floor. Refund authority up to a stated amount. Hiring authority within a stated band. A decision made by someone else inside a defined boundary is not risk exposure. It is throughput.
Jeff Bezos's distinction between one-way and two-way doors is the useful sorting rule: irreversible decisions deserve slowness and your attention, reversible ones deserve speed and someone else's. Most operating decisions are two-way doors being treated as one-way.
Busy is not the same as effective. The scarce resource for an operator is attention on the few variables that actually move the business, and an unmanaged calendar is shaped entirely by inbound requests, which means it is shaped by other people's priorities.
A usable filter for the weekly numbers: if you have looked at a metric three weeks running and it has never changed a decision, delete it. Reporting that cannot change behaviour is theatre with a spreadsheet attached.
"Work on the business, not in the business" — Michael Gerber's line from *The E-Myth* — is directionally right and practically useless as stated, because it does not tell you which in-the-business work is load-bearing. Some of it is the only source of ground truth you have: talking to customers, watching a sale close or stall, sitting in the support queue. Founders who remove themselves from all of it end up running an operating model built on last year's reality and wondering why the forecasts drift.
The better test is what the work produces. Does it generate information nobody else can get you, or does it generate output someone else could generate? Keep the first, route the second. The failure mode is not doing operational work. It is doing operational work that teaches you nothing.