Here’s a sentence we hear in some form almost weekly: “We’re covered on the finance side — we have a great bookkeeper.” And the bookkeeper usually is great. That’s not the problem. The problem is the word “covered,” because it assumes finance is one job that one person either does or doesn’t do. It’s actually three different jobs, stacked, and most middle-market companies are running with the top of the stack empty — while the owner unknowingly works that job himself, badly, at midnight.
The three jobs are easiest to see as three different relationships with time.
Three jobs, three directions of sight
The bookkeeper records the past. Transactions entered, invoices sent, bills paid, accounts reconciled, payroll run. It’s the foundation of everything — bad bookkeeping poisons every layer above it — and a reliable bookkeeper is genuinely valuable. But the job is transactional by design: it produces records, not judgment. Asking your bookkeeper why margins slipped or whether you can afford a second location isn’t unfair because they’re not smart; it’s unfair because it isn’t the job.
The controller organizes the present. This is the layer companies add somewhere in the growth curve, often without naming it: someone who owns the close — turning the bookkeeper’s records into accurate monthly financial statements, on a deadline, under controls. Accruals booked, revenue recognized properly, statements that would survive an outside review, processes that prevent errors and, bluntly, fraud. A controller answers “what happened, exactly, and can we trust the numbers?” When your statements arrive six weeks late or get restated every quarter, the controller layer is what’s missing.
The CFO shapes the future. The top job is the one with no historical component at all: forecasting cash and results, modeling the big decisions before they’re made, setting pricing with actual margin knowledge, managing banking relationships and debt structure, preparing the company for a transaction. The CFO answers the questions that keep owners up: can we afford it, what happens if, which parts of the business actually make money, what will the bank say. Notice that nothing on this list is about recording anything. It’s judgment applied forward.
Your bookkeeper records the past. A controller organizes the present. If no one owns the future, the owner does — at midnight, by default.
The diagnostic: which layer is missing?
The stack fails in predictable patterns, and the symptoms identify the missing layer with surprising accuracy. If transactions are behind, reconciliations don’t happen, and nobody quite trusts the raw data — the bookkeeping layer needs help, and nothing above it can be fixed first. If the data is fine but statements arrive late, contain surprises, or get argued about — that’s the controller layer: the records exist, but nobody converts them into trustworthy monthly truth. And if the statements are timely and accurate but decisions still happen on gut feel, cash still surprises you, and the bank still makes you nervous — the CFO layer is empty. That last pattern is the most common in companies between roughly $5 and $30 million: good bookkeeper, maybe a controller, and a future that nobody owns except the owner.
The reason the top stays empty is arithmetic, and it’s legitimate: a full-time CFO costs a few hundred thousand dollars fully loaded, and a $10 million company doesn’t have forty hours a week of true CFO work. But the conclusion owners draw — so we’ll go without — doesn’t follow. The company doesn’t need forty hours a week; it needs the right few days a month, from someone senior. That’s the entire logic of the fractional model; the signs that it’s time tend to arrive in a recognizable order. The stack also clarifies what a fractional CFO isn’t: not a replacement for your bookkeeper (they’ll finally get direction), not a part-time data-entry upgrade, and not useful until the layers below are at least stabilizing — which is why real engagements often start with a few weeks of controller-style cleanup before the forward-looking work begins.
A practical way to use this article: list the questions that worry you most about the business right now. If they’re about whether the records are right, hire at the bottom of the stack. If they’re about what the numbers mean and what happens next — those are CFO questions, and the fact that you’re the one holding them is the answer to whether the top of your stack is filled.
Building the right finance stack — bookkeeper, controller, and CFO — is where Chief Perspective’s Fractional CFO engagements often begin. If your questions are running ahead of your finance function, let’s talk.
Common questions
Can one person do all three jobs?
In a very small company, briefly, yes. But the jobs pull against each other — transaction processing is interrupt-driven, close work is deadline-driven, CFO work needs uninterrupted thought — and the person capable of the third is expensive to use for the first. Growth splits the stack; pretending otherwise just delays it.
We have a controller who’s been promoted to CFO. Same thing?
Sometimes it works; often the title changes and the job doesn’t. The test isn’t the business card — it’s whether anyone is producing forecasts, modeling decisions, and managing the bank. A controller doing controller work under a CFO title leaves the top of the stack just as empty.
Which do I hire first?
Fix the lowest broken layer first — forecasts built on bad books are decoration. But “first” doesn’t mean “only”: a fractional CFO engagement frequently stabilizes the lower layers and fills the top one in the same motion.