Nothing was on fire. He had someone doing the books, the bills were getting paid, the returns were getting filed. He just could not answer questions at the speed his business had started asking them.
That is what this decision usually looks like. Not a crisis, and not somebody failing at their job. A company that grew past the setup it started with, and an owner who cannot quite name what is missing.
So let me name it. There are three different jobs here, most companies have one of them covered, and the gap is almost always the same one.
The three roles, in plain terms
| Role | What they own | The question they answer |
|---|---|---|
| Bookkeeper | Recording. Coding transactions, accounts payable and receivable, reconciliations, payroll allocation, getting the month closed. | What happened? |
| Controller | Accuracy. Whether the close is right, whether the reconciliations hold, whether the process is improving, and whether exceptions get chased. | Is it right? |
| CFO | Decisions. Forecasting, pricing, hiring and capital decisions, cash planning, and the conversation about what changes next. | What do we do about it? |
Most companies between $3M and $15M have the first one covered and the other two open. That is why the books can be current and still not answer the question the owner is asking.
A note on the word bookkeeper
I am going to say bookkeeper throughout, but for a lot of companies this is an internal accounting team instead, usually an office manager with a helper. Almost everything here applies the same way to both. The differences are in the nuances.
An outside bookkeeper generally has the accounting training and does not know your business. Someone internal knows your business cold and was never trained in accounting, because they inherited the books when somebody had to do them. Different starting points, same ceiling.
What a bookkeeper does, and where the ceiling sits
The recording layer is a skilled job and I do not want to undersell it. Transactions coded correctly, to the right account and the right department. Bank and credit card accounts reconciled. Payroll allocated. Receivables tracked. The month closed on a schedule.
Done well, that is the foundation for everything else, and done badly it makes everything above it worthless.
The ceiling is not about effort or care. It is about scope. Here is how an owner described the gap to me:
She doesn’t really own it. Like, she’s just doing a task.
That was not a complaint about the person. She was doing precisely what she had been asked to do, and doing it on time. Nobody had ever asked her to look at the finished statements and ask whether the story they told was true. That is a different job, and it was not hers.
The two arrangements tend to fail in opposite directions, which is worth knowing when you are diagnosing your own situation.
Someone internal catches things an outsider never would. They know which customer always pays late, which job went sideways, why that month looked strange. What they miss is the accounting-shaped problem: the clearing account that only fills, the balance that has not reconciled since spring, the revenue recorded twice.
An outside bookkeeper is the reverse. They will catch the accounting-shaped problems and miss the operational ones, because a transaction that looks fine on a screen can be obviously wrong to anyone who was there.
What a controller does, and why it is the layer you are missing
A controller owns whether the numbers are right. Not whether they were entered, whether they are right.
In practice that means owning the close rather than performing it, chasing exceptions instead of recording them, fixing the process that produced the error, and managing whoever is doing the recording. It is the difference between a reconciliation being completed and a reconciliation being trusted.
The question I use in interviews: can a bank reconciliation show zero difference and still be wrong?
The answer is yes, and understanding why is roughly the line between the recording layer and the controlling one. You can reconcile an account to the penny while the entries underneath sit in the wrong period, in a clearing account, or duplicated.
This is the layer almost every company I work with is missing, and it is the one owners are least likely to ask for. They call about a CFO. What they need first, nearly always, is somebody making the books trustworthy enough for a CFO conversation to mean anything.
I worked with a much larger company, a multi-state operation, where the sequence was explicit: catch the accounting up, cut the manual work and the errors, add administrative support so the controller was not doing everything, and only then build the strategic reporting. Four steps, in that order, because each one depends on the last.
If your close is slipping, your reconciliations are behind, or nobody can explain a variance without a week of digging, you have a controller gap. Adding strategy on top of that does not work. You get a better-informed conversation about numbers that are still wrong.
What a CFO does that a bookkeeper cannot
Here is the cleanest version of the distinction I have heard, and it came out of a conversation about what this work is worth:
Reports are replaceable. A meeting where you’re helping to make decisions is not.
Anybody can produce a profit and loss statement. Software will do it. The value is not in the document, it is in somebody sitting across from you connecting it to a decision you are about to make.
Concretely, that means:
- Connecting this month to where you are going. Not just what margin was, but whether the trend supports the hire you are considering.
- Modeling decisions against both profit and cash. A truck, a technician, a second location. Those look different on a P&L than they do in the bank, and the bank is what stops you.
- Forecasting. A 13-week cash flow forecast so you see the tight week before you are in it rather than during.
- Deciding what changes. Someone owns it, there is a date, and it gets checked next month.
- Pricing and margin. Whether your rates actually cover fully burdened labor, and which department is subsidizing which.
- Dealing with the bank. Financing capacity, covenants, and what a lender needs to see before you need them to see it.
None of that is a criticism of bookkeeping. It is a different job that starts after the recording ends.
At what revenue should you hire a CFO?
This is the question people actually type, so let me answer it before I complicate it.
For a home service company, the conversation usually starts somewhere around $3 million and is overdue by $10 million. If you are running $5 million with multiple trades or multiple locations, you are past the point where a bookkeeper alone is enough. On the podcast, Josh Kelly put the ceiling in the same place, though he blames middle management rather than finance.
Now the complication, because the number is not the best test.
Looking back at my own engagements, there is no threshold. I have worked with a company at a $4 million run rate where the owner was personally chasing $30,000 to $50,000 of overdue receivables every day and the constraint was administrative capacity. I have worked with a $10 million contractor with fifty employees where the constraint was that nobody could produce a leading indicator. And I have worked with a company many times that size whose problem was that the close was broken, which is a controller problem regardless of revenue.
Same three roles, different order, at wildly different sizes.
The better test is complexity. Multiple trades, multiple locations, multiple entities, a field service system feeding your books, maintenance agreements with deferred revenue, or growth fast enough that last year’s reporting no longer describes this year’s business. Any two of those and revenue stops being the useful measure.
Six signs you have outgrown the setup you have
These come from the files rather than from theory. Any two and it is time.
- Your systems disagree and nobody can say why. I found a $10 million contractor whose QuickBooks revenue was running about $150,000 below what the field system said had been completed. Nobody had been asked to check, so nobody had.
- You cannot get department margin you would bet on. A company average tells you nothing about which trade is carrying which.
- Transactions get processed and problems do not get chased. The work is done. The questions are not asked.
- Receivables are landing back on you. One owner at a $4 to $5 million run rate was personally working $30,000 to $50,000 of overdue AR daily. That is a job, and it was not supposed to be his.
- The close keeps slipping. Past the 10th, then past the 15th, then whenever.
- You have stopped reading the financials. This is the one owners admit last. If the reports arrive and you do not open them, the problem is not the reports.
Owners tend to describe it before they can diagnose it. Things I have heard in first meetings: it is very hard to know what my true AR is. Let us get ahead of it rather than catching it in the monthly financials. We are probably in a soup-to-nuts situation.
In-house, outsourced, or fractional
Once you know which layer is missing, the next question is who fills it.
| Layer | Usually works best as | Why |
|---|---|---|
| Recording | Either | Repeatable work. Outsource it when capacity is the constraint, keep it inside when your business is unusual enough that context matters more than throughput. |
| Controlling | In-house once you can carry it | Needs day-to-day ownership, institutional knowledge, and someone to manage the recording layer. Hard to do at arm’s length. |
| Deciding | Fractional, for most companies your size | You need the judgment a few days a month. You do not need it, or want to pay for it, forty hours a week. |
The version I see work most often in home services is a mix: someone recording, either inside or outside, and fractional support covering the controlling and deciding layers until the company is big enough to bring the controller in-house.
That is most of what I do. Sometimes my team handles the books outright. More often a bookkeeper or a small accounting team is already in place and doing good work, and I take the harder parts of the close, make sure it lands on time and ties out, and then we spend the meeting on decisions rather than on whether the numbers are right.
What to ask before you hire anyone
Whichever direction you go, these separate the people who will help from the people who will produce documents.
- Can a bank reconciliation show zero difference and still be wrong? If the answer is no, they are a recording hire, whatever the title says.
- How will you make sure my field system and my books agree, and how often? In this industry that is the question. The gap between those two systems is where most of the trouble lives.
- What date will the month be closed, every month? You want the 10th. Anything past the 20th means you are reviewing history.
- What will you bring me that I did not ask for? The answer tells you whether they think their job ends at the report.
- Who will I actually be working with? An owner told me about a previous firm where they kept getting passed to different people. That is a cost worth asking about directly.
One more, for yourself rather than for them. If your books were perfect tomorrow morning, what decision would you make differently? If you cannot answer, you need the controlling layer. If you can, you are ready for the deciding one.
If you are trying to work out which layer is missing in your business, let’s have a conversation.
Questions owners ask
What is the difference between a bookkeeper, a controller, and a CFO?
A bookkeeper records what happened: coding transactions, reconciling accounts, closing the month. A controller owns whether those numbers are right, chasing exceptions, fixing the process, managing the close. A CFO uses them to decide what happens next: pricing, hiring, capital, cash planning. Most companies between $3 million and $15 million have the first covered and the other two open, which is why the books can be current and still not answer the owner’s question.
At what revenue should a home service company hire a fractional CFO?
The conversation usually starts around $3 million and is overdue by $10 million, but revenue is a poor test. What actually triggers it is complexity: multiple trades, multiple locations or entities, a field service system feeding your books, deferred revenue on maintenance agreements, or growth fast enough that last year’s reporting no longer fits. Any two of those and the size of the company stops mattering much. I break down what each revenue stage needs in the $3M to $15M scaling breakdown.
What can a CFO do that my bookkeeper cannot?
Connect the numbers to a decision. A bookkeeper tells you what margin was. A CFO tells you whether it supports the hire you are considering, models the truck purchase against cash rather than just profit, builds the forecast that shows you the tight week before you reach it, and sits in the meeting where something actually changes. Reports are replaceable. The decision conversation is not.
Should I hire an in-house controller or outsource the CFO role?
Usually both, in that order of eventual permanence. The controlling layer benefits from being inside the business, because it needs daily ownership and institutional knowledge. The deciding layer rarely needs forty hours a week at your size, which is what makes fractional work. Many companies run a bookkeeper inside, fractional support above, and bring the controller in-house once they can carry the salary.
When does a contractor outgrow their bookkeeper?
When the questions outrun the role. Common signals: your field system and your books disagree and nobody can explain it, you cannot get department-level margin you would bet on, receivables are landing back on your desk, the close keeps slipping, or you have quietly stopped reading the financials. Any two of those and you have outgrown the current setup, whoever is in the seat.
I run a $5 million plumbing company. Do I need a CFO?
Probably, though you may need a controller first. At $5 million with multiple trades or locations, a bookkeeper alone is not enough to give you department margin, a working forecast, and someone to hold decisions against. Start by asking which layer is missing. If your close is late or your reconciliations are behind, fix that before adding strategy on top of it.
Does it matter whether my books are done in-house or outside?
Less than people expect, and they fail differently. Someone internal knows your business and will catch operational problems an outsider would miss, but was often never trained in accounting. An outside bookkeeper has the training and does not know your jobs or your customers. Either can work. What matters is whether anyone owns the layer above them.
