Two work-management tools run side by side in the same company, and nobody remembers why they never picked one. Marketing tracks its campaigns in one. Operations logs its tickets in the other. Finance keeps its own spreadsheet, because neither tool produces the report it needs by the fifth of the month. Eighteen months ago, someone bought the licences, ran two training sessions, then moved on to the next fire. Adoption never crossed forty per cent. Nobody owns the decision to finish the migration, kill one tool, or admit the whole exercise failed. Everyone just works around it, quietly, the way people work around things that are clearly somebody else’s problem.
This is not a story about software. It is what happens when nobody designs the COO function.
The tool war is a fever, not the disease
Faced with a mess like this, the instinct is to fix the symptom: pick a winning tool, mandate its use, move on. It rarely holds. Six months later a new tool appears, backed by a different function, and the same argument starts again under a different name.
That is because the tool was never the actual gap. The gap is a missing function — not a missing person, a missing function. It owns the company’s operating system: which processes exist, who runs them, what tooling supports them. It also owns whether adoption happened, and who follows up when it didn’t. Call it the COO function, whether or not anyone holds that title.
Every company has a COO function, the same way every company has a finance function whether or not it has a CFO. The only question is whether someone designed it, or left it to assemble itself. Designing it means giving it a name, a mandate, and an accountable owner. Left alone, it does not disappear. It scatters onto whoever stands closest when a decision can no longer wait.
It scatters onto whoever stands closest
In practice, across very different companies, it lands in the same three places — and each mislodging has its own cost.
- Finance. Finance answers operational questions as budget questions — it is the only function that can say no to a purchase. The result is cost discipline without operational judgement: price decides the tool, not fit.
- The PMO. Portfolio governance dilutes into ad hoc firefighting, because a PMO is the nearest function that already tracks cross-functional work. The portfolio mission suffers quietly while the PMO absorbs decisions its charter never covered.
- The founder’s inbox. Every unresolved operational question eventually routes to one person: whoever never formally agreed to own the operating layer. That person is also the one least able to spare the attention it needs.
None of these are wrong people to involve. They are simply the wrong place for permanent ownership to default to. Nobody mandated them to hold it; they just had nowhere else to send the question. A PMO absorbing this workload starts to resemble an enlarged, catch-all PMO that nobody ever chartered as one. It is a different failure mode from the one that name usually describes — same underlying cause: accident deciding scope, not mandate.
Smaller companies feel it hardest, not least
It is tempting to file this under “problems for companies past a certain size.” The opposite is usually true. A fifty-person company with a genuinely strong, visionary founder often feels this most acutely. The visionary layer does its job well, generating direction and momentum faster than the company builds an operational layer to absorb it.
This is not a new observation; Gino Wickman’s visionary/integrator split names the same pattern from a different angle. What is worth adding is simpler: you do not need to diagnose a personality mismatch at the top to see the gap. You can see it in the tool war. It shows up in the recurring meeting where the same operational question resurfaces, unresolved. And it’s there in the pilot that quietly became permanent because nobody ever decided to end it.
Designing the COO function is a decision, not an org-chart exercise
Naming it does not require hiring a COO by title. It requires a decision. Someone with real authority names a person or a small group as accountable for the operating layer. The mandate needs to survive beyond the current fire. That decision can sit with a COO, an operations director, or an explicitly chartered equivalent role. The title matters far less than whether the mandate exists — and whether the rest of the organisation knows it does. A mandate without decision rights produces the same tracker-function problem here as it does anywhere else on this site.
What does not work is hoping the function assembles itself out of goodwill. Goodwill is exactly what keeps finance, the PMO, and the founder’s inbox absorbing the question in the meantime. That is precisely why the gap stays invisible for so long.
Who owns your company’s operating system — by name, not by accident?