Operating Systems

The Day Your Decision System Broke (And Why Nobody Said Anything)

There is a point in a company’s growth you can almost pin down in hindsight, the day the way decisions used to get made stopped working. Nobody declared it. No single meeting failed in flames. But somewhere between fifteen people and fifty, or fifty and a hundred, the operating system started producing the wrong outputs: slower calls, recurring confusion about who owns what, a creeping sense that things are getting harder rather than easier.

The company is doing well. Revenue is up, the team is bigger, and the founder is tired in a way she cannot quite name, while the leadership team keeps having the same conversation about the same unresolved things. Every quarter closes with a fresh initiative to improve alignment. Every offsite produces commitments that half-land. This is the breaking point, and if you are a Chief of Staff there is a fair chance you saw it before anyone had the words for it.

What actually breaks

Informal decision-making is brilliant in small, close teams. Fifteen people share a channel, read each other’s signals, make calls in the corridor, and execution follows because everyone knows who does what. Then the team grows. The corridor now excludes half the people who need to know. The founder still makes the same fast, informal calls, but the execution layer is three times larger and the original recipients are managers themselves, so by the time a decision reaches where it is needed it has been interpreted three times and lost half its context. HBR’s May 2026 work on fast-growing companies names four fault lines where this fractures: alignment, operational complexity, financial discipline, and oversight.1 Cross one threshold and the informal system that made you fast becomes the thing making you slow. What makes it hard is the silence: nobody walks in to announce that the decision architecture has broken. Things just take longer, meetings multiply, the same issues resurface.

Three signs the system has broken

  • Meetings proliferate. When the decision system stops working, organisations compensate by booking more meetings. Status updates replace operating rhythm, and check-ins multiply because nobody trusts that information is flowing. 2026 data has employees in three times as many meetings as five years ago, a 192% rise, with 71% of senior managers calling them unproductive.2
  • Decisions drag. You can time this one. From “a decision is needed” to “the first action is taken” runs hours to days in a healthy system and weeks in a broken one, because nobody is sure who owns it and nobody wants to be wrong.
  • The leader cannot let go. When the system fails, leaders re-centralise, because trusting a machine that keeps producing the wrong output feels worse than gripping it yourself. HBR noted in March 2026 that many leaders are withdrawing under pressure, their agency eroding; some over-centralise, some disengage, both answers to the same broken system.3

The diagnostic

Before proposing fixes, spend two weeks mapping the operating reality. Which decisions are made at the right level? Which travel upward for no reason? Which take more than ten days? Where do the same questions keep getting asked? You will usually find three gaps: missing ownership, where decisions sit in a grey zone with no named owner; broken information flow, where the person who must decide lacks the context when they need it; and no agreed escalation path, where anything off-template has nowhere to go. Each wants a different fix. Missing ownership needs a simple map of who is responsible, accountable, consulted and informed for your key decision types. Broken flow needs a rhythm change, usually fewer and better-structured meetings. The missing escalation path needs a short conversation about what gets escalated, to whom, and when.

What to do this week

Pick one recurring item that keeps reappearing in your leadership meetings, the thing that should have been settled months ago. Trace its path: who first flagged it, where it went, who was supposed to own the decision, what stopped it being resolved. That trace is your operating-system audit, and one item will teach you more about how your decision architecture really works than any offsite. Take it to your principal as an observation, not an accusation: “I have been looking at how this kind of decision travels through the organisation. Here is what I found.” That framing keeps it diagnostic and fixable.

The breaking point is not a crisis. It is a growth stage every scaling company passes through, and most pass through it slower than they need to because nobody gives it a name. The Chief of Staff who sees it early, names it plainly, and brings a real diagnosis to the right conversation is doing some of the most valuable work there is. When your decision system breaks, the most useful thing you can do is say so, out loud, and then help build the thing that replaces it.

Sources:
  1. Harvard Business Review, “How Fast-Growing Companies Can Make Better Decisions”, May 2026[]
  2. Microsoft Teams workplace data and HBR survey of senior managers, 2026[]
  3. Harvard Business Review, “Leaders Feel Their Agency Eroding, and They’re Starting to Withdraw”, March 2026[]