The midnight loop, and the shape of the real problem

By the time we met the founder, the team was forty people and the company was growing fast enough that every quarter looked different from the last. The founder was not avoiding decisions — they were making decisions all day — but the decisions kept reopening at midnight. A call made in a Tuesday standup would be re-litigated by the founder alone at eleven that night, re-confirmed on Wednesday morning, then quietly reversed on Friday after a customer conversation. The team had stopped trusting that any decision was final, which made them cautious, which pushed more decisions back upstream to the founder, which made the loop tighter and the nights longer.

The founder described the problem as decision fatigue, and they were right, but the fatigue was a symptom. The actual problem was that nobody in the company — including the founder — could tell which decisions deserved a slow careful process and which ones were cheap to make and cheap to reverse. So every decision got the heavyweight treatment, which exhausted the founder and trained the team to wait.

Two weeks of data, and a tiering nobody had thought to draw

We started by logging every decision the founder touched over two full working weeks. Not the meetings — the decisions. Each entry was classified on three axes: reversible or not, founder-only or delegable in principle, strategic or operational. The classification was deliberately conservative. Anything that could plausibly be argued either way was marked as needing founder input, so the data would not flatter our eventual recommendation.

Even with the conservative classification, more than sixty percent of the founder’s decision load was reversible operational decisions that the leadership team was demonstrably capable of making. Another quarter were decisions that needed founder input but did not need it in real time. Less than fifteen percent of the founder’s decisions actually required the heavyweight, slow, founder-owned process the company was applying to all of them. The midnight loop was not a discipline problem. It was a routing problem.

A three-tier system the team could run without the founder

We introduced a simple tiering. Type-1 decisions — hard to reverse, founder-owned — get a written brief, a 48-hour minimum window, and an explicit recorded rationale. Type-2 decisions — reversible but consequential, leadership-owned — get a one-page memo, a 48-hour clock, and a default of “proceed unless objected.” Type-3 decisions — reversible and operational, team-owned — require no founder involvement at all and no memo unless the team chooses to write one for their own clarity.

The hard part was not the tiering. It was building the leadership team’s confidence that a Type-2 decision, once made, would not be reopened by the founder at midnight. For the first month we ran a Friday review of every Type-2 and Type-3 decision the team had made — not to second-guess them, but to calibrate confidence in public. The founder learned which decisions they had been right to trust the team on; the team learned which kinds of decisions the founder genuinely wanted to be looped into earlier. By week six the review was a fifteen-minute check-in. By week ten it was retired.

What changed, and what stuck a year later

The most visible change was the calendar. The founder’s week shifted from reactive one-on-ones to weekly strategic blocks, and the leadership team’s calendars filled with the kinds of decisions they used to escalate. Nobody added a new tool. Nobody hired a chief of staff. The company added a frame, a habit, and a Friday review that paid for itself by retiring itself, and the rest followed.

A year later, the system still runs without us. The founder reports that the biggest unlock was not the time saved, although the time saved is significant. The biggest unlock was the end of the midnight loop. Decisions stay decided. Reversals, when they happen, are explicit and recorded rather than emotional and silent. The team trusts the process, which means the founder trusts the team, which is the only mechanism we have ever seen actually scale founder leverage in a company growing this fast.