The role a controller actually plays
A bookkeeper records transactions. A controller reviews that work, owns the close, provides variance commentary, documents the process, and makes sure the numbers can be trusted for decisions. If nobody is doing that review layer, you likely have a gap, even if your bookkeeper is doing their job well. The controller is the person who signs off that the numbers are right and knows why they moved.
The confusion between bookkeeping and controller work is common and expensive. A bookkeeper might be technically skilled and producing accurate categorizations, but if nobody is reviewing the close, signing off on variances, or catching the moment a recurring vendor starts showing up in the wrong account, the work is incomplete. Most bookkeeping errors are not about math. They are about categorization decisions nobody senior reviews.
A useful test: when your CEO asks why operating expenses were up $45K in March, who answers? If the answer is the bookkeeper, and the explanation is "I will look into it," you do not have a controller. If the answer is a written note that says March OpEx is up $45K driven by the annual insurance renewal posting in full rather than being amortized, you have controller-level work happening.
One more practical indicator: how your investor updates go. If preparing the monthly investor email takes three days, not because the analysis is deep but because the underlying numbers keep shifting, that is a review-layer problem. Good controller work produces financials that do not change after they are closed. Whatever gets sent out is what stays in the record.
Signs you need this layer
Closes that stretch past the 15th of the following month. Numbers that change between versions without clear explanation. Reports you are embarrassed to show investors. Variances between budget and actual that nobody can explain. A CEO who is still chasing the bookkeeper for answers. When any of these show up consistently, a controller layer is the fix.
A pattern we see repeatedly: company hits $3M to $5M in revenue, books are current, tax returns get filed on time, but when the CEO asks for a rolling 12-month P&L with commentary, or when an investor asks for revenue bridges, the answer is always "give me a few days." That delay is what the controller layer fixes. The numbers exist. The analysis and accountability around them does not.
Another flag: your CPA asks for adjustments at year-end that your bookkeeper does not understand. If the CPA is effectively re-doing the books at tax time, that work belongs in-house throughout the year. Otherwise you are paying twice - once for the bookkeeper to produce monthly statements that cannot be relied on, and again for the CPA to correct them before filing.
A related red flag: the bookkeeper cannot explain changes from prior months without opening the books and digging. They can produce the report but not interpret it. That is fine for bookkeeper-level work. For decisions above $25K-$50K in impact, the person producing the number needs to be able to explain it from memory.
Full-time vs fractional
Most companies under $20M in revenue do not need a full-time controller. The work is real, but not 40 hours a week. A fractional controller gives you the senior oversight at 8 to 20 hours a month, scaled up during fundraising or audit periods. This is typically the right call until you are well past $20M in revenue with a growing finance team to manage.
The math on full-time usually does not work under $15M in revenue. A controller with 5-10 years of experience costs $110K to $150K base plus benefits - about $170K all-in. At 20 hours a month of real controller-grade work, that is $700 per hour of actual value. Fractional controller engagements typically run $3K to $6K per month for equivalent work, roughly 65-75% cheaper for the same output.
The exception is regulated industries and companies preparing for an IPO or acquisition in the next 12-18 months. In those cases, a full-time controller is worth it for the continuity, the documentation, and the fact that the same person is answerable for every period that will be audited. For everyone else, fractional is the right starting point.
The hybrid model works for companies in the $10M-$20M range: fractional controller oversight (20-30 hrs/month) plus a full-time senior bookkeeper or accountant in-house. This gives you daily coverage on transactional work and senior review where it matters. Total cost is typically $8K-$12K/month all-in versus $15K-$18K for a full-time controller.
What good controller work looks like
Close calendar documented and followed every month. Every account reconciled to source before books close. Variance analysis delivered with written commentary that a non-accountant can understand. Issues identified and resolved in-cycle rather than surfacing weeks later. A clear audit trail on every adjustment. You should receive the same quality of package every month, not a scramble.
Finsightic handles accounting, controller oversight, and fractional CFO work for growing companies. Fixed monthly pricing, no long-term contracts.
A controller package your board should expect to see every month: reconciled trial balance with all accounts tied to source, P&L and balance sheet with variance commentary against budget and prior period, cash flow statement, aged AR and AP, key operational metrics relevant to the business, and a short written summary of what moved and why. Delivered within 10-15 days of month-end, consistently.
What you should NOT see: different numbers in different reports for the same period, last-minute adjustments showing up in board packages, unexplained variances, or the phrase "we are still investigating that." These are signs the review layer is missing or rushed. Good controller work feels boring and predictable because the thinking happened during the month, not in the last three days before the board meeting.
Most companies discover they need a controller because of a specific event: a due diligence request, an audit, a messy year-end, an investor question they cannot answer. If you are in the middle of one of those moments, you do not have time to hire slowly. A fractional engagement can start in 2-3 weeks. A full-time search takes 3-6 months. The timing matters as much as the decision.
Take the free Financial Health Score →