When every consequential decision still routes through you

A full calendar is not enough to prove that the founder is the bottleneck. Some decisions should remain with the chief executive: capital allocation, ownership, enterprise risk, and choices that are difficult to reverse. The test is whether decisions that should move elsewhere repeatedly return to the founder before action can begin.

Look at the last 30 consequential decisions, not at the org chart. Record who discovered each issue, who developed the options, who had authority to choose, who actually chose, and how long the work waited. If the formal owner and the actual decider differ, the operating model is telling the truth that the chart hides.

• Count waiting time separately from working time.
• Mark decisions reopened after an executive had already made them.
• Identify thresholds that trigger an automatic escalation to the founder.
• Distinguish a request for information from a request for permission.

The usual diagnosis—people need to be more accountable—is incomplete. Authority returns for a reason. The executive may lack information, the decision boundary may be vague, incentives may punish an honest miss, or the founder may routinely reverse decisions after delegating them. Each cause requires a different fix.

Do not begin by announcing that the team is empowered. Begin by choosing one recurring class of decision and writing the boundary. State the outcome the owner controls, the financial or risk threshold, the facts that require consultation, and the conditions that require escalation. Anything inside the boundary is theirs to decide, including the right to choose differently than you would.

Task delegation preserves the founder as the operating system: someone else gathers the facts, builds the document, and runs the meeting, but the judgment still waits at the same desk. A real transfer includes the decision, the resources needed to execute it, and the obligation to report what happened.

Start with decisions that are important enough to matter and reversible enough to learn from. Agree on a review date instead of requiring approval in advance. At review, judge the reasoning and the result separately. A sound decision can produce a poor outcome; punishing it teaches the team to send every future choice back to you.

• Name one owner, not a committee.
• Write the decision boundary in language the owner can test alone.
• Give access to the numbers and people required to decide.
• Review outcomes on a schedule rather than intervening continuously.

Ask what would happen if you were unreachable for two weeks. The useful answer is not “nothing would go wrong.” It is a list: which decisions would proceed, which would pause by design, and which would become unsafe. If leadership cannot produce that list, the decision system is implicit.

Run the test in a bounded area before treating it as a vacation experiment. Choose a function, define its decision rights, and let the executive operate without pre-clearance for two weeks. The resulting decision log will show whether the constraint is capability, information, incentives, or your own difficulty letting a different judgment stand.