How and When to Hire an FP&A Manager

When the forecast starts driving decisions. Until then, budgeting and variance analysis sit comfortably inside controllership. Once pricing, hiring, spend and capital decisions depend on a forward view, someone needs to own that view full time.

The usual signals are unglamorous. The budget is built once a year and ignored by March. Nobody can say what a new customer is actually worth. The sales team and finance quote different revenue numbers in the same meeting. Someone asks what happens if growth slows by a quarter and the answer takes two weeks.

Worth noting what does not signal it. Revenue does not, headcount does not, and an investor asking for a monthly pack does not on its own, since that may be a reporting problem rather than an analysis one.

A fractional CFO or an external FP&A resource can carry a good deal of this before you hire. They are effective at building the first model, establishing a forecast cadence and preparing board material. What they cannot do well is the continuous part: being in the pricing discussion, catching a bad assumption in a sales meeting, or knowing why a cost line moved before anyone asks. When that continuous work is what you are missing, the arrangement has run its course.

Almost always someone whose main job is something else. Usually the founder, the CFO, or the Finance Controller or Manager doing it alongside controllership. Identifying who, and how much of their week it consumes, tells you what level to hire.

In early growth companies the Controller or Finance Manager frequently carries both functions, and that arrangement works longer than people expect. One person closing the books and building the forecast has a real advantage: they know exactly what the numbers are made of. It is only when the two jobs start competing that the arrangement breaks, and reporting deadlines always win. Analysis gets deferred to the week after close, then the week after that.

So the diagnostic is not whether someone is doing FP&A. It is whether they are doing it properly or doing it last. If your Controller is producing a forecast at 11pm after the close lands, you have your answer, and you also have a decision to make about whether to hire under them or beside them.

If the person carrying it is the founder, hire sooner. Founder time spent rebuilding a model is the most expensive analysis in the business.

One builds and reports the numbers. The other sits with commercial teams and changes what they decide. Most job descriptions describe the second and most interview processes test the first.

Most companies need the first before the second, and most write job descriptions for the second because it sounds more strategic. The result is a strong partner who arrives to find no reliable data, or a strong modeller who is asked to influence a sales director and cannot.

At senior levels the two converge. Below that, pick one deliberately.

FP&A owns everything forward-looking and everything that explains why. Controllership owns what happened and whether it is right. The boundary matters commercially, because FP&A functions that drift into reporting stop doing analysis and nobody notices for a year.

• Budget and reforecast cycles, including the assumptions and who signs up to them.
• Driver-based forecasting, and the model everyone works off.
• Unit economics: contribution, cohort behaviour, customer acquisition cost and payback, whichever apply to your model.
• Business partnering with commercial functions, and the analysis behind pricing, discounting and spend decisions.
• Scenario and sensitivity work, including the downside cases nobody enjoys presenting.
• Management reporting commentary. Not the production of the pack, which sits with controllership, but the explanation of what it means.

What should not sit with FP&A: the close, reconciliations, statutory reporting and audit support. Where one person carries both, as in many early growth companies, protect the analysis time explicitly. Otherwise the close will eat it every month, and the FP&A capability you thought you had will exist only in the job title.

The business model decides what the analysis is about, and that decides who to hire. A strong SaaS FP&A professional is not automatically useful in a manufacturing business.

Cross-model moves work. They take a quarter or two, and the candidate should be able to tell you what they would need to learn. Ones who claim it transfers straight across usually have not done it.

Under the CFO if you want partnering. Under the Controller if you want reporting discipline first. Embedded in the business only once the function is mature enough not to lose its independence.

This is treated as an administrative detail and it decides what the role can actually do.

FP&A reporting into controllership gets close proximity to the data, which is genuinely useful, and inherits controllership's priorities. When close and forecast compete, close wins. The function stays accurate and gradually stops being consulted before decisions.

FP&A reporting to the CFO keeps its independence and its access to commercial leadership. It also needs the CFO to actually back it, because an FP&A manager telling a sales director their pipeline assumption is optimistic will not survive long without support from above.

Embedding analysts inside business units gives the best partnering and the worst consistency. Two units end up calculating margin differently and nobody notices until the board asks. It works where there is a strong central FP&A function setting definitions. It fails as a first move.

Whether they have owned a forecast that people relied on, and whether anyone changed a decision because of their analysis. Years of experience and modelling skill are the weakest predictors on the CV.

Four things worth weighing.

Ownership of a number. Did they build the forecast, defend it, and answer for the variance? Or did they populate someone else's model? A candidate who has never had to explain a miss has not done the job.

Evidence someone acted. Ask what changed as a result of their work. Analysis nobody used is not FP&A, whatever the title said.

Business model relevance, per the table above. Closer is faster, and further is workable if the candidate is honest about the gap.

Comfort with imperfect data. FP&A never has clean inputs. Candidates from very mature environments sometimes struggle in businesses where the data has to be argued with first.

Consulting and audit backgrounds transfer well on rigour and badly on ownership. They have advised, and often not lived with the consequences. That gap closes quickly for the right person, and it should be tested rather than assumed.

Give them your actual numbers, anonymised, and ask what they would look at first. It reveals more in twenty minutes than a modelling test does in three hours.

Modelling tests measure whether someone can build. They do not measure whether someone knows what is worth building.

A format that works, and takes about an hour of everyone's time:

• Send four to six quarters of P&L, anonymised, with a one-page sheet of context: what the business sells, roughly how, and three or four operating assumptions.
• Give them twenty to thirty minutes with it, either before the interview or at the start of it. No spreadsheet build required.
• Spend thirty minutes discussing three questions: what would you look at first, what is missing here, and what would you want to ask the business?

Keep it to that. Multi-day take-home cases screen out strong candidates who are already employed, and they measure diligence rather than judgement. If you want to see modelling ability as well, ask the candidate to walk through a model they built previously with the confidential data removed.

Then listen for a view. Good candidates form one quickly and say what would change it. Weaker candidates describe a method, or ask for more data before committing to anything.

• What in this business would you expect to be the biggest driver of profitability?
• Which number here would you not trust without checking, and why?
• Tell me about a forecast of yours that was badly wrong. What did you miss?
• Describe a time you told a commercial leader something they did not want to hear.

The forecast-miss question does most of the work. Everyone in FP&A has been badly wrong. Candidates who cannot produce an example are either inexperienced or were not the ones accountable for it.

Most come from hiring a partner when the business needs a builder, or the reverse.

• Writing a business partnering job description when the underlying problem is that nobody trusts the numbers.
• Hiring a strong modeller and expecting influence. Model quality and the ability to hold a room are unrelated skills.
• Placing FP&A under controllership and then being disappointed that analysis is always late.
• Hiring senior FP&A into a company with no data infrastructure, where the first year will be spent on plumbing.
• Assuming a candidate from a much larger business will adapt to doing the work themselves.
• Testing modelling ability thoroughly and commercial judgement not at all.

If your close is late or your numbers are disputed, an FP&A hire will spend the first year on data quality rather than analysis. Fix reporting first, or hire knowing that is the job.

This is the most common disappointment we see with FP&A hires, and it is rarely the candidate's fault. The company wants forward-looking insight, the actuals arrive late and change after the fact, and the new hire ends up rebuilding reconciliations to trust their own inputs. Eighteen months later the business concludes FP&A did not work here.

Two honest options. Fix controllership first and hire FP&A afterwards. Or hire someone who has done exactly this before, tell them plainly that year one is remediation, and judge them on that rather than on the analysis you were hoping for.

The useful part of an FP&A search happens before it opens: deciding whether you need a builder or a partner, and being honest about the state of your data.

We scope the mandate before opening the search, test candidates on ownership and commercial judgement rather than modelling alone, and give shortlists with a written assessment of each candidate. FP&A is one of our most frequently run mandates across startups, GCCs, BFSI and manufacturing. More on how these searches run is on our FP&A recruitment page.