When Should an Agency Hire a Finance Director

I get asked some version of "when should we hire a finance director" by agency owners roughly once a month, usually over a coffee that they have engineered specifically to ask it. They want a number — a headcount, a revenue line, a moment. I used to give them one. I don't any more, because the number is not the thing, and giving them a number sends a lot of agencies up the ladder a rung too early, or into a hire that fails for reasons that have nothing to do with the person.

So this is the decision guide I wish someone had given me before I started hiring up the finance function at my own agency. It is in the first person because I got it wrong before I got it right, and the wrong version is more instructive. I am deliberately not retelling the CFO story here — I have written that separately in the two CFOs, and the lesson there is the spine of this one — but this is the wider map: the whole ladder, what each rung can and cannot do, and the single signal that tells you it is time to climb.

The ladder nobody draws for you

There are four rungs to the agency finance function, and most owners only see the top and bottom of it. Drawn out, with what each one is actually for:

The revenue bands are deliberately soft, because they are the least reliable part of the table. I have seen a £4m agency that genuinely needed a full-time FD because its work was complex, multi-currency and lumpy, and a £9m agency that ran beautifully on a sharp management accountant and a fractional FD two days a week because its book was simple and repeatable. Size is a hint. Complexity and decision cadence are the actual drivers.

What the ladder is really measuring is not how much money flows through the agency. It is how fast you need to turn raw financial events into decisions, and how much judgement those decisions require.

What each rung can and cannot do

The honest version of each rung, including its ceiling.

The bookkeeper gives you accuracy. Invoices raised, bills paid, payroll run, the ledger reconciled, VAT filed on time. This is non-negotiable and you need it from the first month you trade. What a bookkeeper cannot do — and should not be asked to do — is tell you whether a project made money or whether a retainer is quietly underwater. That is not their job and most will tell you so plainly. The trap here is mistaking a tidy ledger for financial insight. A clean set of books tells you the agency is solvent and compliant. It tells you almost nothing about whether it is profitable in the way you think it is.

The management accountant turns the record into something a leader can read: a monthly P&L, a margin view, a cashflow forecast, variance against budget. This is the rung where most agencies first get genuine financial reporting, and for a long stretch — somewhere in the low millions of revenue — it is exactly the right rung. The ceiling is in the word monthly. A management accountant produces a closing pack, and a closing pack is, by its nature, a report on a period that has ended. It is reconstruction: assembling what happened after it happened. Excellent for understanding the last quarter. Structurally unable to catch a project bleeding this week, because this week isn't closed yet.

The fractional or part-time FD is the first rung that buys you judgement rather than record-keeping. A day or two a week of someone senior who has seen agency economics before, can sanity-check pricing, can sit in on a difficult client decision, can tell you whether your margins are normal or a warning. For most agencies between roughly £3m and £8m this is the highest-leverage hire on the whole ladder, because you get genuine financial seniority without paying for a full-time executive you cannot yet keep busy. The ceiling is availability: a fractional FD is, definitionally, not there most of the week. They can interpret the numbers you bring to the Tuesday session. They cannot be in the room for the Thursday decision.

The full-time FD or CFO owns the financial strategy of the business — pricing, capacity, the shape of the book, the conversation with the board or the buyer. This is the rung you reach for when financial decisions are dense enough, and consequential enough, that you need someone whose full-time job is to have an opinion about them. The ceiling here is the most expensive one to hit, and it is not about the person at all. It is about whether your tooling can feed them a live number. Which is the whole point of this article.

The trap: a senior hire into a job the tooling makes impossible

Here is the mistake I made, stated plainly so you can avoid it.

I hired senior finance leadership to interpret the agency's economics in real time — to catch problems mid-flight, to make pricing calls with current data, to walk into a Monday review with a view on what was happening now. But the only data the agency could produce was a monthly close that took the best part of a week to assemble and landed roughly three weeks after the period it described. I had hired a job that the tooling made impossible. The interpretation I wanted required live numbers. The numbers were three weeks old by the time anyone could read them. No amount of seniority closes that gap; you cannot interpret your way out of a data lag.

This is the single most common way an agency finance hire fails, and it almost never gets diagnosed as a tooling problem. It gets diagnosed as a hiring problem — "the FD wasn't proactive enough", "they were too in the weeds", "they couldn't see the big picture" — when what actually happened is that you asked a senior person to do interpretation while only giving them the means to do reconstruction. They spent their week rebuilding the past because that was the only thing the data allowed, and then there was no week left for the future, which is what you hired them for.

So before you climb a rung, ask the unglamorous question: can my systems produce the number the new hire is supposed to interpret, at the speed the decisions require? If a project starts losing money on Monday, when does anyone find out — Tuesday, or the third week of next month? If the answer is next month, hiring a more expensive person does not change the answer. It just makes the lag more expensive.

A worked decision, anonymised

A founder I know runs a ~45-person agency, around £5.5m revenue, mostly retained development work with a handful of larger fixed-price builds a year. He came to me convinced he needed to hire a full-time finance director, because he "couldn't see what was going on" and a peer his size had just made the hire.

We went through it. His books were clean — a good bookkeeper, no issue there. He had a management accountant producing a solid monthly pack. What he was actually describing when he said "couldn't see what was going on" was a cadence problem: the pack told him in the third week of the month that a fixed-price build had gone over in the previous month, by which point the next build was already underway and likely repeating the mistake. He didn't need more financial seniority to read the pack. He needed the pack to stop being the only way he could see.

The decision we landed on had two parts, in order. First, fix the cadence: get project margin reporting off the monthly-close cycle so a build going over shows up in days, not weeks. Second — and only then — bring in a fractional FD, one day a week, to use that now-live data for pricing judgement and the occasional hard client conversation. A full-time CFO at his size and complexity would have been an executive salary spent largely on someone reconstructing a past he could already see clearly enough once the cadence was fixed. Eighteen months on, he is still on the fractional arrangement, the fractional FD is genuinely useful precisely because there is a live number to interpret, and the full-time hire is a "maybe at £9m" rather than an "urgently now".

The order mattered. Had he hired the full-time FD first, into the old monthly cadence, I am fairly sure I would be writing him into a version of the story I have told elsewhere about hiring the wrong job.

The signal that actually means "climb"

Strip away the revenue bands and the org charts and there is one signal that reliably means it is time to move up a rung: you keep needing an answer that requires interpretation of a live number, and the rung you are on can only give you a clean record of an old one.

When the questions you are asking finance shift from what happened last month to what is happening right now and what should I do about it, you have outgrown reconstruction and you need judgement. That is the moment to climb — and the moment to check, hard, that your systems can actually serve a live number to the more senior person you are about to pay for. Hire the judgement and fix the cadence together, or the judgement has nothing current to judge.

If you want to understand why I am so insistent on the live number — and why I think the finance hire and the tooling decision are the same decision — the two CFOs is the long version of how I learned it, and the about page explains why I ended up building the thing I needed instead of hiring around the gap a fourth time. If you would rather just see what a live margin number looks like on your own work, the demo runs for fifteen minutes on your real Jira data and doesn't ask for a card up front.

The short version: don't hire by headcount or by what your peer did. Hire by the question you keep failing to answer — and make sure the answer can arrive fast enough for the person you're hiring to do the job you think you're hiring them for.