What a controller actually owns
The clean division: a bookkeeper records transactions, a controller guarantees the output is right, and a CFO decides what to do about it.
In practice the controller owns everything that makes financial statements trustworthy — the chart of accounts, the close calendar, the reconciliation standard, revenue recognition policy, and the internal controls that stop one person from both approving and paying an invoice.
That is why the role is easy to skip. Nothing visibly fails when you have no controller. Your books still close. They are just quietly less reliable each month, and you find out during diligence.
The signals
Any two of these together usually mean the role is already overdue.
- Your close is slipping. It used to take 10 days and now takes 20, without a matching jump in volume.
- You restate. Numbers change after they have been reported, more than once a year.
- Nobody owns the chart of accounts. New accounts get created ad hoc and month-over-month comparisons have stopped being meaningful.
- Your CPA is doing cleanup at year-end. You are paying tax-preparer rates for accounting work that should have happened monthly.
- Approval and payment are the same person. A basic segregation-of-duties gap that auditors and insurers both flag.
- You cannot answer margin questions by product or segment. The data exists but is not structured to answer it.
- You are heading into an audit, a raise, or a sale. Diligence tests exactly the things a controller owns.
The most common trigger
A first institutional raise or a bank covenant. Both introduce an outside party who reads your statements adversarially, and that is usually the first time the gap becomes expensive rather than theoretical.
In-house, fractional, or outsourced
A full-time controller runs $110,000–$160,000 plus benefits, and below roughly $10M in revenue that is more capacity than the work requires.
Fractional or outsourced controllership typically runs $2,500–$6,000 a month and buys the same standard-setting and review without the fixed cost. The usual path is outsourced through the growth phase, then in-house once the volume justifies a full seat.
One thing to insist on either way: the controller must be a different person from whoever does the bookkeeping. The entire value of the role is independent review. If the same person records and reviews, you have bought a title, not a control.
What changes in the first 90 days
A competent controller engagement is diagnostic before it is operational, and the sequencing is fairly standard.
- Weeks 1–3: review prior closes, rebuild the chart of accounts, document what is actually happening today
- Weeks 4–6: establish the close calendar and a written reconciliation standard for every balance sheet account
- Weeks 7–10: implement approval workflows and segregation of duties
- Weeks 11–13: first close fully under the new process, with a reconciliation package behind every number