What does a fractional CFO actually do? The first 90 days, week by week

Most founders come into their first CFO conversation with a growth story ready. The CFO's first move is to set it aside. Before any forward-looking work begins, the books get read: the chart of accounts, the cost classifications, the liabilities sitting off the balance sheet. Not because the strategy does not matter to the CFO, but because the numbers it depends on are rarely as clean as they look. The infrastructure comes first, and the first 90 days are how it gets built.

This is the week-by-week breakdown of what a fractional CFO works on across that period: what gets produced, in what order, and what is materially different in the business by the time each phase closes. The previous two guides in this series covered what the role is and what sits on a retainer. This one is the sequence.

Founders who wait for that leadership to come to them are already behind the ones who built the infrastructure early.

Source: SBA Office of Advocacy, FAQs About Small Business 2026

Survival rates across all US businesses are sobering, but the failure mode at venture-backed startups is more specific: the financial infrastructure to catch problems early did not exist. A fractional CFO engagement in the first 90 days is, in practical terms, the installation of that infrastructure.

The three-phase arc

The 90 days divide cleanly into three phases. Each phase creates the conditions the next one requires. You cannot build a reliable model on inaccurate books. You cannot prepare for a raise if the model has not been stress-tested.

1. Month one: weeks 1 to 4

What a fractional CFO does in the first 30 days

The first four weeks are diagnostic. The CFO reads what exists: the chart of accounts, how revenue is being recognised, what is off the balance sheet. The same structural gaps appear in almost every pre-CFO finance function. Books on cash basis instead of accrual. Payroll costs in the wrong category, making gross margin look cleaner than it is. Deferred liabilities not reflected in the runway calculation.

One question that comes up before almost every engagement starts:

The month one diagnostic typically covers the following ground:

• Financial diagnostic complete. Every account category reviewed and mapped correctly. Chart of accounts restructured if needed.
• Books migrated to accrual basis if still on cash. GAAP-aligned accrual accounting is investor-expected baseline at any seed stage and above. Revenue is recognised when earned, not when cash is received.
• 13-week rolling cash flow forecast built from the actual bank balance. Burn rate and runway calculated precisely. The founder can state both numbers in under sixty seconds without opening three files.
• Gross margin recalculated correctly. Hosting, contractor, and support costs that belong in cost of goods sold are moved there. The resulting margin number is the one investors will work from.

A SaaS founder came into a Fintera engagement with nine months of runway on paper. The month one diagnostic found hosting costs sitting above the gross margin line. Moved correctly, gross margin dropped eight points. Runway recalculated to six months. The raise timeline moved up by ninety days. The number had been wrong for the better part of a year.

2. Month two: weeks 5 to 8

How a fractional CFO builds your financial model (days 30 to 60)

Once the books are clean, the forward-looking work begins. The CFO builds the three-statement model: income statement, balance sheet, and cash flow, mathematically linked so that a change in revenue flows through margin, burn, and runway automatically. A standalone revenue projection is not a financial model. This is the distinction that separates founders who get clean follow-up questions from those who get asked to come back in thirty days.

Founders often want to know how much of their own time this phase demands before it starts:

From there, the month two work runs largely without you needing to be in the room. The work the CFO is building in parallel looks like this:

• Three-statement model complete with monthly view for the next 18 to 24 months. Every assumption documented and defensible.
• Unit economics by channel or product line. CAC, LTV, payback period, and contribution margin calculated and segmented.
• Scenario planning: base case, upside, and downside. If growth comes in 20 percent below plan, what happens to runway? The answer should be prepared, not calculated live in a partner meeting.
• Burn multiple tracked. Net burn divided by net new ARR. Below 1.0 is excellent. Above 2.0 is a conversation to start before an investor does.

3. Month three: weeks 9 to 12

Board packs, data rooms, and investor prep: the final 30 days

By week twelve, the books are clean and the model is stress-tested. Month three is about turning that foundation into the outputs that carry the business through the next six months. The question at this stage is almost always about ongoing time commitment:

The deliverables that close out month three reflect that division of effort:

• Board pack template built and first edition produced. Burn, runway, ARR, churn, and budget versus actuals in a consistent format, ready 48 hours before every meeting. This recovers four to five hours of founder prep every month.
• Cap table modelled across the next two rounds. Equity decisions made at seed have real consequences at Series B.
• Data room skeleton built. Financial history, three-statement model, cap table, and legal entity structure in one place. Adding to it during a live raise takes hours, not weeks.
• Investor Q&A preparation complete. The financial questions every Series A partner asks first, answered clearly before the first meeting.

What if a raise is still 12 to 18 months away?

The 90-day arc serves the business regardless of whether a raise is imminent. The burn multiple informs hiring decisions. The unit economics inform pricing. The board pack makes every leadership conversation more grounded. The fundraise makes the work more urgent, not more valuable.

A SaaS business running a burn multiple of 1.8 used the three-statement model to evaluate a planned engineering hire. The model showed that at current ARR growth, the hire would push the multiple above 2.0 within two quarters without a corresponding lift in net new ARR. The hire was deferred by one quarter, structured differently when it did happen, and the multiple held. That decision had nothing to do with a raise. It had everything to do with having a model that made the consequence visible before the headcount decision was made.

The financial infrastructure is useful from the day it exists. Most founders build it six months later than they should have.

At Fintera, every engagement follows the same three-phase arc. The CFO partner assigned from day one is briefed on your stage and your fundraising timeline before the first call. The infrastructure built over 90 days is yours, documented, and accessible to you beyond the engagement. The model does not live on one person's laptop.