The title is doing a lot of work and almost no standardising. Two companies can both say they have a fractional CFO and mean completely different things: one has a senior operator running scenario planning before a Series A, the other has an experienced accountant reviewing the month-end close. Both are legitimate. Only one of them changes decisions.
A fractional CFO owns the forward-looking finance work: the financial model, runway and scenario planning, board and investor reporting, unit economics, and fundraising support. They do not do the books. That is the bookkeeper's job, and it is the input a CFO depends on.
The work, concretely
Stripped of the packaging, the job is to turn a ledger into decisions. That breaks into eight recurring areas:
- Build and maintain the financial model, and keep it reconciled to actuals rather than diverging from them after month two.
- Own cash: a 13-week forecast, burn tracking, and runway under several hiring and revenue scenarios.
- Produce board and investor reporting, including agreeing metric definitions before they appear in a deck.
- Run the fundraise from the finance side: the model investors will stress-test, the data room, and the diligence question list.
- Analyse unit economics and cohorts, so growth decisions rest on which customers are profitable rather than on blended margin.
- Model hiring capacity, which is usually the largest and least reversible spending decision an early company makes.
- Set up the finance stack and close process so the reporting is repeatable without heroics.
- Coordinate tax strategy and credits with whoever files, so the planning happens before year end rather than after.
The first thirty days of a good engagement usually produce two artefacts: a 13-week cash forecast and a ninety-day roadmap. If neither exists after a month, the engagement has drifted into advisory conversation.
What they do not do, and why the line is firm
Some of the boundary is practical. A fractional CFO is not going to categorise transactions, chase receivables, or run payroll. Those belong to bookkeeping and operations, and paying CFO rates for them is the most common way an engagement wastes money.
The rest of the boundary is professional, and it is more rigid than most founders realise. Under the AICPA's rules on nonattest services, an outside provider cannot assume management responsibilities. Specifically, they cannot authorise or execute transactions, take custody of assets, design or maintain your internal controls, take responsibility for the fair presentation of your financial statements, or report to your board on behalf of management.
In plain terms: your fractional CFO can build the board package and sit in the meeting, but the numbers are still management's numbers and the controls are still your controls. Any provider implying otherwise is describing something they are not permitted to do.
What a month actually looks like
| Cadence | What happens |
|---|---|
| Weekly | Cash check-in: position, near-term commitments, anything that moved runway |
| Monthly | Close review, KPI pack, variance against plan, the investor update |
| Quarterly | Re-forecast, hiring plan, scenario work, board materials |
| As needed | Fundraise support, diligence, pricing changes, a major hire or contract |
Engagements are priced on this cadence rather than on hours, which is why asking how many hours a month you get usually produces an uncomfortable answer. The useful question is what lands on your desk each week and each month.
How to tell a good one from an expensive one
Three tests separate them. First, can they explain your unit economics back to you in a way that surprises you? If the analysis only restates what you already believed, it has not gone deep enough. Second, does the model reconcile to your actuals every month, or has it quietly become a separate document? Third, when the number is bad, do you hear it early and directly?
The last one matters most before a raise. A CFO whose job is partly to defend your numbers in front of investors has to be someone who has already challenged them in front of you. For what to have ready, see investor-ready monthly financials and what belongs in a board reporting package.
The first ninety days
Engagements that work tend to follow the same arc, and knowing it helps you tell early whether yours is on track.
The first month is diagnostic. A 13-week cash forecast gets built, the existing model is stress-tested against actuals, and the gaps in reporting become visible. Expect to be told things you did not want to hear about the state of the numbers; that is the engagement working rather than failing.
The second month is where the reporting becomes routine, meaning the board pack and metric definitions are settled and produced without a scramble. The third month is where the forward work starts: the hiring plan modelled properly, unit economics analysed by segment, and a view on when you should raise. If by month three you are still receiving retrospective commentary on last month's numbers, the engagement has stalled.
The two most common ways this goes wrong
The first is buying seniority to fix a bookkeeping problem. If the ledger is weeks behind, the first month of any CFO engagement goes on cleanup at CFO rates, and you have bought expensive bookkeeping. Fix the record first, or engage a team that owns both.
The second is a scope that never leaves the rear-view mirror. A fractional CFO who reviews the close each month and comments on variances is doing controller work. That is useful, and it is not what the title is for. The value sits in the decisions ahead of you, not the explanation of the month behind you.
Where Zinance fits
Because your books close daily, the CFO work starts from current numbers instead of waiting on a close. The same team handles bookkeeping, tax, and R&D credits, so the model, the ledger, and the return are built from one set of figures rather than three, and the questions that usually bounce between providers get answered in one place.