Short answer: This is rarely an either/or decision. A bookkeeper records and reconciles. A controller owns the close, the controls, and whether the reporting is right. A CFO works forward: forecasting, capital, strategy. Most independent restaurants need the first. Most multi-unit groups need the first two. Outsourcing the bookkeeping is usually what makes an in-house controller worth hiring, because it takes the transaction work off their desk and lets them do the job you are paying for.
Operators tend to frame this as a choice: do we outsource the books, or do we hire someone. Framed that way it is a budget question, and the budget answer is almost always "not yet." Framed correctly it is a question about which work needs to happen in your building and which does not.
| Role | Owns | Looks |
| Bookkeeper | Recording transactions, bank and card reconciliation, delivery payout reconciliation, AP entry, payroll journal entries, sales tax data | Backward, daily and weekly |
| Controller | The close calendar, the chart of accounts, internal controls, accuracy of the P&L and balance sheet, vendor and cash policy, the audit trail | Backward, monthly, with an eye on process |
| CFO | Forecasting, cash planning, unit economics, pricing, capital structure, lender and investor relationships, expansion modeling | Forward, quarterly and annually |
The confusion comes from the middle row. In a lot of restaurant groups, nobody owns it. The bookkeeper records what they are given, the owner looks at the P&L and feels uneasy about it, and there is no one whose job is to say whether the number is right.
Two things muddy this.
First, titles are used loosely. Plenty of people called a controller are doing bookkeeping, and plenty of bookkeepers are quietly doing controller work because someone had to. The title on the org chart tells you very little about what is actually getting done.
Second, restaurant bookkeeping is genuinely harder than the transaction volume suggests. Revenue arrives through the POS, three or four delivery platforms, catering, and gift cards, and each settles on a different schedule with different fees netted out. Payroll carries tips, service charges, and tip-outs. An in-house hire who has not done restaurants before will spend their first two quarters learning what your bookkeeper already knows, and you will pay controller wages for bookkeeping output.
For most independent operators running one or two locations, it is. The work is recording, reconciling, and closing on time, and none of it needs to happen in your building. What you need is accurate books, a P&L inside a useful window after close, and someone who picks up when a vendor credit looks wrong.
The test is whether anyone is making decisions that require forward-looking financial work. If the decisions in front of you are menu pricing, labor scheduling, and vendor terms, accurate and timely books plus your own judgment will cover it.
The trigger is usually structural, not revenue. Watch for:
Notice none of those are "we got big enough." A single-location restaurant with a lender and two entities may need controller-level oversight. A five-location group with clean structure and no outside money may not.
When both exist, the handoff looks like this.
Reconciled bank and card accounts, delivery payouts tied to platform statements rather than bank deposits, AP entered and coded to a restaurant chart of accounts, payroll journal entries posted, and the period closed on your calendar within an agreed number of days.
Reviews the close rather than performing it. Approves the chart of accounts and any changes to it. Sets the policy the bookkeeping follows: how comps are treated, how shared costs are allocated between entities, what needs an accrual. Signs off that the P&L is right before anyone makes a decision from it.
Your controller spends their time on judgment, structure, and review, which is what they are expensive for. They are not entering invoices at 9pm. This is the arrangement most multi-unit groups land on eventually, and the ones that get there sooner spend a lot less on the journey.
Two common failures.
Hiring a controller to do bookkeeping. You pay well above bookkeeping rates for bookkeeping output, and the controller work still does not get done because the transaction load fills the week. The person usually leaves inside a year.
Outsourcing the books and assuming oversight came with it. Your books are accurate and your allocations are still wrong, because nobody decided how the management company should charge the locations. Accurate inputs, wrong structure, and the P&L you are steering by is misleading in a way that looks clean.
Ask two questions.
Is the transaction work getting done accurately and on time? If not, that is a bookkeeping problem, and hiring a controller will not fix it. It will bury a well-paid person in data entry.
Is anyone accountable for whether the numbers are right? Not recorded, right. If the honest answer is that you assume so, that is the controller-shaped gap, and it does not get smaller as you grow.
Most operators find the first answer is no and the second is nobody. Fixing the first often reveals how much of the second you actually needed, because accurate books make the structural questions visible.
If you are working through this for your own group and want a second opinion on which gap you are looking at, we are happy to talk it through.