The question is almost never asked in the right order. Founders ask what a CFO costs, then work out whether they can afford one. The better sequence is to ask what decisions the business is currently getting wrong, and then work out who is cheapest to stop that happening.
Both answers land somewhere sensible. But only the second one tells you whether the seat needs to be full-time.
What each one actually is
A full-time CFO is a permanent executive who owns the finance function, sits in every leadership conversation, hires the team beneath them and carries the relationship with the board and the bank. Their value compounds with time in the building.
A fractional CFO is the same discipline bought by the month. You get the judgement, the board pack, the forecast and the diligence readiness. You do not get someone in every meeting, and you do not get a manager for your finance team unless you buy that too.
The distinction that matters is not seniority. It is presence.
The cost, loaded rather than headline
A startup CFO base salary in the US typically runs $180,000 to $240,000. That is not the cost. Add payroll taxes, benefits, recruitment fees amortised over an expected tenure, and the equity you are handing over, and the real annual number is usually somewhere near $260,000 before anyone has produced a board pack.
| Full-time CFO | Fractional | |
|---|---|---|
| Annual cost | ~$260,000 loaded | $21,000 to $72,000 |
| Time to productive | 3 to 6 months, plus search | 2 to 3 weeks |
| Equity | Typically 0.5% to 2% | None |
| Notice | Severance, and a rehire | Monthly |
| Presence | Every meeting | Scheduled cadence |
| Manages a team | Yes | Only if scoped |
Our own retainers start at $1,750 a month. The wider market for fractional CFO work sits roughly between $3,000 and $12,000 a month depending on scope and how much of the close is included. Even at the top of that range, the fractional seat is a third of a loaded permanent hire.
That gap is not the argument for fractional. It is the reason the argument gets made badly.
Where the crossover actually sits
Cost is the wrong axis. The seat becomes full-time when the volume of finance decisions exceeds what a scheduled cadence can absorb. In practice we see that happen at 3 thresholds, and it only takes one.
- Around 40 to 60 employees. Below that, payroll, hiring plans and departmental budgets can be run monthly. Above it, someone is fielding a compensation, contract or budget question every day.
- Around $10m to $15m of revenue, or earlier with multi-entity or multi-currency structure. Consolidation, intercompany and revenue recognition stop being a monthly task.
- A live transaction with an internal counterparty. A raise can be run fractionally. An acquisition you are integrating, a lender covenant you are negotiating monthly, or an audit with a real materiality threshold cannot.
The seat goes full-time when the questions arrive faster than the cadence, not when the revenue crosses a number.
The failure mode of hiring out of order
The most expensive version of this mistake is hiring a CFO onto an unreliable ledger. A CFO whose first 6 months are spent cleaning up bookkeeping is a $130,000 bookkeeper, and they usually leave.
The order that works is: accurate books, then a controller or a close discipline, then the CFO layer. If your close takes 3 weeks and the numbers move after you have reported them, the problem is below the CFO seat and a CFO will not fix it faster than a bookkeeper would.
The reverse mistake is cheaper but slower. A business running fractional 2 years past the crossover accumulates decisions that were deferred to the next scheduled session. Nothing breaks visibly. The plan just stops being ambitious, because nobody had the bandwidth to model the ambitious version.
When to stop using a fractional CFO
We tell clients to start the search when three of these are true at once, and we will say so unprompted:
- You are scheduling extra sessions most months because the cadence is not enough.
- There is a finance team of three or more that needs a manager, not a reviewer.
- The board is asking for scenario work between meetings rather than at them.
- A transaction is running that will still be running in 6 months.
- The company is materially over 60 people.
A good fractional CFO should be able to hand over cleanly, because the artefacts already exist: a documented close, a maintained model, a board pack in a stable format. The handover is the test of whether the arrangement was any good.
A short self-test
Open your last board pack. If you can answer every question it raises from the file itself, without a call, the cadence is working. If you cannot, count how many of those questions arrived between meetings. That count, not your revenue, is the number that decides this.