What does a fractional COO actually do?
The title says how someone is engaged. It says nothing about the work, and the difference is where most of these arrangements succeed or fail.
“Fractional COO” describes how someone is engaged. It does not describe the work. It tells you the person is senior, part-time and temporary; it tells you nothing about which decisions they will take, what they will build, or what your company is left holding when they leave. Founders who hire on the title get an expensive adviser. Founders who hire on the work get an operating layer that stays.
So here is the work, as plainly as I can put it.
The job is to make the company legible
Companies at this stage are rarely badly run. They are illegible. Everything needed to run them exists, but it is distributed: in two founders' heads, in a spreadsheet somebody maintains out of goodwill, in what the team remembers from the last all-hands. That works up to roughly twenty people because everyone can hold the whole company in mind at once. Past that, it stops working quietly, and the first symptom is not chaos. It is slowness.
The first thing a fractional COO does is write the company down. Who decides what. What we are actually trying to achieve this quarter, in a form that survives being read back in twelve weeks. Where the work is, and where it is stuck. This is unglamorous and it is most of the value, because almost every argument a growing company has is really an argument about an undocumented assumption.
What actually gets owned
An adviser recommends. An operator owns. The distinction matters more than any other in this arrangement, and it is worth settling in the first conversation rather than the third month. In our engagements the owned list is usually some subset of:
- Decision rights. Which decisions belong to a founder, which to a lead, and which need nobody's permission at all. Most companies are surprised by how much of the third category they have been treating as the first.
- The operating cadence. The smallest set of meetings and written updates that keeps everyone pointed the same way, and the discipline to keep it small.
- Planning. A quarter that means something: a short list of outcomes, named owners, and an honest read on capacity before commitments are made rather than after.
- The hiring plan. Not the headcount number. The sequence, and the definition of what each role owns before the role is opened.
- Cross-functional execution. The work that falls between teams, which is where delivery actually breaks and where nobody's job description reaches.
- Board and investor reporting. Assembled from things the company already measures, rather than as a separate quarterly fiction.
If you cannot say which decisions the person owns outright, you have not hired an operator. You have hired a second opinion.
What does not get owned
A fractional COO should not become the company's single point of coordination. That is the failure mode of the role, and it is a comfortable one: it feels like value, everything runs through them, and eighteen months later the company cannot function when they leave. It replaces one bottleneck with a better-dressed bottleneck.
They also should not own your product direction, your engineering decisions or your culture. They should make all three easier to exercise by removing the operational noise sitting on top of them.
What the first ninety days look like
Broadly: understand, then design, then hand over. The first few weeks are spent finding out how the company actually runs rather than how the org chart says it does, which usually means sitting in the existing meetings and asking who decides. Then the missing pieces get built and put into use — not documented and circulated, but used, in the real meetings, on the real work, with the real disagreements that follow. Then ownership starts moving to the people who will still be here.
We have written about that quarter in more detail in the first 90 days after a seed round, because the post-raise version of it is the most common one we see.
How you know it is working
Not by how busy the person looks. By whether decisions are being made further from the founders than they were, whether anyone in the company can name this quarter's priorities without checking, and whether commitments live somewhere other than a founder's memory. Those three are observable from the outside, and a founder can check all of them in an afternoon.
When it should end
The engagement should have an end state described at the start, and it should usually be a person: a first operations hire, a promoted lead, or a founder who has taken the operating agenda back with a system underneath it. If you are choosing between fractional support and a permanent hire right now, that is a different question with a fairly clean answer, and we have set it out in fractional COO or first operations hire.
A good engagement makes itself unnecessary on a schedule. That is not modesty. It is the deliverable.
