What a fractional CFO actually does, and when you need one

"Fractional CFO" has become one of those terms that means whatever the person selling it wants it to mean. It covers everything from a retired controller doing eight hours a month to a full strategic finance function. That vagueness is expensive, because it makes it very hard to tell whether you need one.

The clearer way to think about it is not as a job title but as a ladder of four distinct jobs. Most businesses need them in order, and most problems come from being on the wrong rung.

The four jobs

Bookkeeping — what happened

Transactions recorded and categorised, accounts reconciled, a P&L and balance sheet produced. This is the foundation and there is no skipping it. If your books are wrong, everything built on top is wrong with more confidence.

Controller — is it right, and is it on time

Close discipline, accrual treatment, revenue recognition, a chart of accounts that reflects how the business actually runs rather than how the tax return is filed. A controller is why your numbers arrive on the 15th instead of the 23rd, and why you can trust them when they do.

FP&A — what happens next

Forecasting, scenario modelling, budget versus actual, unit economics. This is the first rung that looks forward rather than back. It answers "what does this decision do to cash in month seven."

CFO — what should we do about it

Judgement. Capital strategy, pricing, whether to take the contract, how to talk to a lender or a board, when to stop doing something that is working but not working well enough.

The mismatch that causes most of the pain

Owners rarely misjudge the first rung. They misjudge the gap between the second and third.

The common shape: a business grows past $2M, the bookkeeper is competent and cheap, the books close eventually, and nothing is obviously broken. But the owner cannot answer "can I afford to hire," "which of my service lines actually makes money," or "will I be able to make payroll in March" — because none of those questions are answerable from a historical P&L, no matter how accurate it is.

The clearest signal you have outgrown your current setup is not that the numbers are wrong. It is that the numbers are right and still don't help you decide anything.

What it costs

Rough market shape, so you can calibrate:

• Bookkeeping — $500–$2,000 a month depending on volume and complexity.
• Controller-level support — $2,000–$5,000 a month, usually bundled with the accounting rather than sold separately.
• Fractional CFO — $5,000–$16,000 a month depending on company size and how much of the function you're buying.
• Full-time controller — $80,000–$120,000 plus benefits. A full-time CFO, considerably more.

The reason fractional works at all is that most companies between $2M and $25M need CFO judgement a few days a month, but need the accounting and forecasting underneath it every day. Buying only the judgement, without the machinery, tends to produce expensive opinions built on unreliable inputs.

Four signals you're ready

1. You're making commitments you can't model. Hiring, pricing, a lease, a new location — decided on instinct because the analysis doesn't exist.
2. Someone external is about to ask hard questions. A lender, an investor, a board, a buyer. They will ask for things a bookkeeper cannot produce.
3. Your close lands after the 20th. By the time you see the month, you're a third of the way through the next one.
4. You cannot answer "how long is my runway" without building a spreadsheet. If the answer requires a project, you don't have the answer.

And when you're not

If your books are genuinely a mess, start there. Strategic finance built on unreliable accounting is worse than no strategic finance, because it produces confident answers that are wrong. Clean the foundation first — it is usually a one-off project rather than an ongoing cost.

If you're pre-revenue or very early, a good bookkeeper and a spreadsheet you maintain yourself is the right answer. The value of a forecast scales with the number of moving parts, and early on there aren't many.