Built to Be Handed Over: How a Fractional CFO Engagement Should End

The first conversation with a founder is always about the problem in front of them: a raise in six months, a cap table nobody trusts, a board asking questions the numbers cannot answer. The engagement letter that follows describes a monthly retainer, a list of deliverables and a start date. What it almost never describes is the end. There is no exit criterion, no date to review whether the engagement is still needed, and no picture of who will run the work afterwards.

I think that is the single most common design flaw in fractional finance, and it is a flaw on both sides of the table. The founder does not raise it because they are relieved someone has finally taken the numbers off their plate. The fractional CFO does not raise it because a retainer that never ends is a comfortable business. Two years later the company has grown past the point where a few days a month is enough, the CFO has become the only person who understands the model, and nobody has a plan for changing either of those facts.

This post is about designing the end of an engagement from the beginning: the ways it can end, what has to exist for a handover to work, who should receive it, and how the overlap should run. It follows on from the first 90 days of a fractional CFO engagement, which covers the other end of the same timeline.

Three ways an engagement ends

In my experience, fractional engagements end in one of three ways, and only two of them are good.

The first is graduation. The company grows to the point where it needs a full-time finance lead, hires one, and the fractional CFO hands over the function and leaves. This is the outcome a fractional engagement should be designed for at most venture-backed companies, and it usually arrives somewhere between a Series A and a Series B, when the finance work stops fitting into a few days a month and starts needing someone who is there every day.

The second is step-down. The heavy lifting — the clean-up, the model, the raise — is done, and what remains is an operating rhythm that a bookkeeper or junior accountant can run with occasional senior review. The engagement shrinks to a quarterly board pack review and an annual budget, or ends entirely with an open door. This is the right outcome for a profitable small company that does not need a CFO every month, and it is a good outcome, not a failure.

The third is drift, and it is the one to avoid. Nobody decides anything. The engagement continues because stopping it would be inconvenient, the scope expands a little each quarter, and the fractional CFO becomes a permanent part-time employee without anyone having chosen that. Or the reverse: the CFO's attention fades, the monthly pack starts arriving late, and the engagement ends because a renewal was quietly not signed. Either way, the company ends up with a finance function that lives in one outsider's head.

A fractional engagement that works should make itself unnecessary. If the company cannot run its own close, its own board pack and its own model without me, I have not finished the job.

The exit belongs in the first conversation

The fix for drift is to name the end before you start. Not a date — nobody knows in advance whether the company will raise in nine months or eighteen — but a set of conditions under which the engagement will be reviewed. Mine are usually some version of these: the company has more than roughly thirty people, the monthly close needs more than a few days of senior time, the board expects a finance lead at every meeting, or a priced round has closed that brings an investor expectation of a full-time CFO. When one of those becomes true, we sit down and decide whether to hire, step down or continue for a defined period.

Writing those triggers down at signing does something useful to the rest of the engagement. It makes the deliverables about capability rather than output. If I know from the first month that someone else will eventually run the close, I build a close that someone else can run. If I know the model will be inherited, I build a model with an assumption log rather than one that only makes sense to its author. The same logic applies to the triggers in a hiring plan: you decide in advance what has to be true before you act, so the decision is not made by default.

What “handed over” actually means

A handover is not a meeting and a shared folder. It is a set of things that exist, are written down, and have been run at least once by someone other than the person who built them. When I hand over a finance function, the receiving person should be able to find and use each of the following without asking me.

The close runbook: the calendar from the five-day close, with every step, the system it happens in, the owner, and what a correct result looks like. A close that has only ever been run by one person is a close nobody else knows how to check.

The definitions: what the company means by revenue, ARR, burn, cash and headcount, written down as queries or formulas with a date and an approver. This is the definitions file described in the post on the data layer. Without it, the next finance lead will quietly redefine a metric in their first month and the board will see a break in the trend that nobody can explain.

The model and its assumption log: every driver with its source, the date it was last reviewed and the history of how forecasts compared with actuals. The track record from scoring the forecast is often the most valuable thing to inherit, because it tells the new owner which assumptions have been reliable and which have not.

The control matrix and the obligations calendar: who approves what, from the minimum viable control environment, and every recurring deadline the company has — tax filings, payroll returns, statutory accounts, audit milestones, covenant tests, investor information rights, option grant approvals. These deadlines are easy to carry in memory and easy to miss when the memory walks out of the door.

The access: every finance system — bank, ledger, billing, payroll, cap table, data store — owned by a company account, with the fractional CFO as a user who can be removed rather than as the account owner. This sounds administrative until the day a company discovers that the admin login to its accounting software is tied to an outside adviser's personal email.

The key-person test

A useful way to check all of this is to ask a simple question every quarter, long before any exit: if I became unavailable tomorrow, what would stop? If the answer is “the board pack would be late”, that is fine; it will be late once. If the answer is “nobody could produce the cash forecast” or “nobody knows the bank password”, that is key-person risk, and an investor doing diligence will find it. It is also a risk the fractional CFO created, which means it is theirs to remove.

Who receives the function

When the trigger is graduation, the hard question is who to hire. Founders often assume the answer is a full-time CFO. Sometimes it is. More often, at the point where a fractional engagement starts to feel too small, what the company actually needs first is a strong controller or finance manager — someone who owns the close, the controls and the day-to-day, and who can be supported by senior strategic input on fundraising and board work for another six to twelve months. The full-time CFO hire then comes later, when the strategic load justifies it.

That sequence is cheaper and usually works better, but it depends on the fractional CFO being honest about which part of their work is strategic and which part is operational. A large share of what a fractional CFO does in the first year is operational clean-up that a good controller could own. If the company hires a CFO to do that, it is overpaying for the role and probably losing the hire within two years.

Whichever role it is, I think the outgoing fractional CFO should help hire it: write the job description from what the work actually is, sit in on interviews, and set a practical exercise based on the company's real numbers. Nobody knows better what the job involves. The obvious conflict — an adviser hiring their own replacement — is exactly why the exit criteria need to be agreed up front, so that the hire is the plan rather than a threat.

Run the overlap, do not just schedule it

The most common handover mistake is a two-week overlap made of meetings. The new person is walked through the model, shown the folders, introduced to the accountant, and then left to run the next close alone. The first time anything unusual happens — a late invoice, a payroll correction, a board member asking for a number in a different cut — they discover what the walkthrough did not cover.

A handover that works is measured in cycles, not days. The pattern I use is three months. In the first, I run the close and the board pack with the new person shadowing every step. In the second, they run both and I shadow, correcting where needed and writing down whatever the runbook was missing. In the third, they run both alone and I review the output before it goes out, the way a reviewer would. By the end, the company has seen the function run three times, once without me touching it.

How to tell it is time

Founders sometimes ask how they will know. The triggers written at the start are the formal answer. The informal signs are usually visible earlier. The fractional CFO is spending most of their time on operational work that a full-time person could do more cheaply. Decisions wait for their next scheduled day. The team has started calling them for answers that should be in a report. The company's investors have begun asking, politely, when there will be a full-time finance lead. Each of those is a sign the work has outgrown the model, and the right response is to say so rather than to quietly add days to the retainer.

The opposite signs point to a step-down. The close runs on time without senior input, the board pack has not needed restructuring in months, and the meetings have become reviews rather than work. That is the moment to reduce the cadence and save the company the money, not to find new projects to fill the days.

None of this is in tension with doing good work. It is what good work looks like in a fractional role. The measure of a fractional CFO is not how indispensable they become; it is how much of the finance function the company still owns and understands after they have gone. If the engagement was designed well, the last month should be the least eventful of all.

If you are partway through a fractional engagement with no plan for how it ends, or about to hire a first full-time finance lead and want the handover to hold, I am glad to help design the exit criteria, the handover pack and the overlap.