When to Hire a Controller: The Question a $9M Firm Should Ask First

The problem: The bookkeeper closes the month accurately and on time, and the owner still cannot get an answer to a question in the days a decision actually allows.

The solution: Judge the finance function by how fast it can answer a question rather than by how much volume it processes, because that is the capability you would be buying.

Most owners approach this as a staffing question. Revenue is up, headcount is up, the bookkeeper is stretched, so it must be time to add above her. That framing puts the decision on volume, and volume is the wrong measure. Plenty of firms at $9 million are processed perfectly well by one competent bookkeeper and a good accountant at year end.

The tell is different, and it is easy to miss because it does not look like a finance problem. It looks like the owner making a decision on Thursday about something he asked a question about on Monday, and going with his gut because the answer was not back yet.

The real threshold is time to answer, not transaction volume

Watch what happens at a $9 million firm with 45 people when the owner asks a real question. Not "what did we bill last month", which the bookkeeper can answer immediately, but something like:

• Which practice area actually contributed margin last quarter after utilization and non-billable time?
• If we take this client at the rate they are asking for, what does it do to the margin on the team that would staff it?
• What does the next six months of cash look like if that one large engagement slips a quarter?
• Which of our people are consistently staffed on work that prices below what their time costs?

None of those are exotic. All of them are answerable in principle from records the firm already keeps. But each requires pulling time entries, engagement budgets, billing, and payroll together, and at most firms this size that means someone exporting from three places into a spreadsheet over a couple of days, then the owner questioning a number, then another day of reconciling. A week later there is an answer, and the decision was made on day three.

That is the threshold. Not that the work is not getting done. That the answer arrives after the window closes, every time, so the business is effectively deciding without it.

What a bookkeeper is and is not for

This is worth being fair about, because the framing "the bookkeeper is no longer enough" reads as a criticism of a person who is usually doing exactly the job she was hired for.

A bookkeeper records what happened. Bills go out, bills get paid, payroll runs, the month closes, the accountant gets clean books. That is a recording function, and a good one is genuinely valuable.

A controller does something different in kind. A controller is responsible for the numbers meaning something: how work is costed, what the margin structure actually is, what the forecast says, and whether the reported figure and the real figure are the same. It is an interpretation function.

Firms get into trouble when they ask a recording function to perform an interpretation function and then conclude the person is underperforming. She is not. She was never given the structure that interpretation requires, and building that structure is not her job either.

Why the answers are slow, and it is usually not the person

Here is the part that changes the decision. At most firms this size, the reason an answer takes a week is not that nobody in the building is capable of producing it. It is that the inputs live in separate places that do not know about each other.

Time entries sit in one system with a project code. The engagement's budget and scope sit in a proposal document or a spreadsheet the partner keeps. Billing sits in the accounting system, organized by invoice rather than by engagement. Payroll knows what people cost but not what they were working on. Nothing connects a person's hour to the engagement it was spent on to the rate it was billed at to what that hour cost the firm.

So every question that crosses those boundaries, which is every interesting question, becomes a manual assembly job. Hire a controller into that and you have bought a very expensive person to do the assembly by hand, faster and more reliably than before, but still by hand. Some firms do this and are satisfied, because faster and more reliable is worth something. Many find that the answers still take three days instead of five, and the decision window was two.

The order that tends to work better is to connect the inputs first, so the assembly is not manual, and then decide what kind of person you need on top of it. Sometimes that is still a controller. Sometimes it turns out to be a fractional CFO for two days a month, because the assembly is no longer the bottleneck and what you actually needed was judgment, not capacity.

What you are actually buying

Frame the spend as a capability rather than a box on the org chart, and the evaluation gets concrete. You are buying the ability to answer, inside a decision window:

• What each engagement type really returns, so you can price and pursue accordingly.
• What a new hire, a new office, or a new practice line would do to cash and margin before you commit.
• Where the firm's capacity is going, so utilization is a managed number instead of a year-end discovery.
• Whether the number the owner is about to decide on is right, which is the part nobody can do for themselves.

Once it is framed that way, the test is easy to apply. Take the last three decisions of consequence you made. For each one, ask what you would have needed to know, whether you could have known it in time, and what it would have been worth to know it. If the honest answer is that you would have decided differently at least once, the capability pays for itself and you are only choosing what shape to buy it in.

A look at a professional services firm

Consider a professional services firm doing about $9 million a year with 45 people across three practice areas. One bookkeeper, an outside accountant at year end, QuickBooks and a time tracking tool that do not talk to each other, and margin analysis that happens in a spreadsheet a partner rebuilds each quarter. The owner has been advised to hire a controller at around $110k base, roughly $130k with payroll taxes and benefits.

Suppose that before hiring, the firm connects time entries, engagement budgets, billing, and payroll into one picture, so an hour worked carries its cost, its rate, and its engagement with it. What you would expect within a quarter is not a new report. It is that questions which used to take a week take an afternoon, and a handful of things nobody was looking for surface on their own: an engagement type that prices well and staffs badly, a client whose scope has drifted twice without a change in fee, a practice area whose margin depends entirely on two people being fully utilized.

Then the hiring decision gets made with better information. If the firm still hires a controller, that person spends their first month interpreting rather than reconstructing, which is what you were paying for. At $130k all in, the position needs to return about $11k a month. At a firm this size, a single engagement repriced correctly, or one mistimed hire avoided, would plausibly cover a year of it. The point is not that the controller is cheap. It is that you would be able to tell whether they were earning it, which is more than most firms can say about the position a year in.

How to start

You can test this before you post the job.

1. Write down the last three questions you could not get answered in time. Be specific about the deadline you had and when the answer arrived. That gap is the thing you are buying away.
2. Trace one of them. Find out exactly which systems someone had to touch to answer it. The number of hand-offs is usually the real explanation for the delay.
3. Connect the inputs before you add the headcount. Get hours, engagement budgets, billing, and payroll into one connected picture so the assembly is not a manual job for whoever you hire.
4. Set the standard in days. Decide what "fast enough" means for your decisions, say two days for a pricing question and a week for a capacity question, and hold the finance function to that rather than to volume.

The takeaway

The signal that a firm has outgrown its bookkeeper is not the number of transactions. It is that the owner has stopped asking questions because he knows the answer will not arrive in time, and has quietly gone back to deciding by instinct on decisions large enough to hurt. A controller is one way to buy that capability back, and often the right one. Just find out first whether the delay is a person problem or a plumbing problem, because if the inputs never meet, the expensive new hire inherits the same week-long assembly job you were trying to escape.