What Does a Financial Controller Do

All right, here is the plain answer before anything else. A financial controller is the senior accountant who owns the accuracy of a company's numbers. The controller runs the monthly close, keeps the general ledger clean, produces the financial statements, enforces internal controls, and manages the accounting team. Bookkeepers record transactions. CFOs set strategy. The controller makes the numbers trustworthy enough to run a business on.¹

Here is the reframe that makes the rest of this article make sense. Founders usually meet the title backwards, in a job posting or a banker's request, and they file it as either a bigger bookkeeper or a junior CFO. It is neither. The controller is the trust function. Every other job in finance consumes the numbers. This is the one job whose entire purpose is making sure the numbers deserve to be consumed, and businesses discover they need it at the exact moment somebody important stops taking their spreadsheet at face value.

So let me walk you through it the way I would on a call. First, what the job actually covers day to day. Then the monthly close, because that is the controller's home turf and the place most growing companies quietly bleed. Then the ladder, bookkeeper to controller to CFO, in one look. Then the signals that say you need one, the question of whether a controller or a CFO comes first, what is happening to the supply of these people, and finally the three-year mess that taught me this job from the inside.

The job, in the words I use on calls: making the numbers safe to use

Start with the official version. The Bureau of Labor Statistics classifies controllers as a type of financial manager and says "Controllers direct the preparation of financial reports that summarize and forecast an organization's financial position."¹ That is accurate and bloodless. NetSuite's definition gets closer to how operators talk: "Universally, a financial controller is a company's lead accountant, responsible for accurate financial statements and efficient accounting processes."² Oracle uses nearly the same words, and adds a piece of trivia founders ask me about: a comptroller is the same job wearing a government or nonprofit badge.⁶

Day to day, the seat covers five territories. The general ledger, meaning every account balance has support behind it. Financial reporting, meaning statements produced on a schedule and built to an actual standard. Robert Half's hiring guidance puts it plainly: "Corporate controllers oversee the accounting and financial reporting functions of a company and should have strong knowledge of GAAP and financial reporting."⁷ Then internal controls, meaning the approval rules and separations that keep money from walking off. Then the accounting team itself, the staff accountants and the AP and AR clerks, plus the standing job of working with any outside auditors the company's lenders or investors send in.¹⁴ And finally the systems those people work in, because the controller is usually the one who picks and implements the accounting stack.

The honest footnote to all of that is how the hours actually get spent. Deloitte's research on the role found that "Controllers in the United States spend nearly 70 percent of their time performing traditional tasks" like closing the books and staying compliant, and that more than 60 percent of controllers had no clearly defined vision for the role beyond those tasks.¹⁶ Deloitte describes four faces of the job, steward, operator, catalyst, and strategist, and most controllers live almost entirely in the first two. That is not a criticism. When the first two faces slip, nothing built on top of them survives.

The close is the controller's home turf, and it is where most companies bleed

If you want to understand the controller in one process, watch the monthly close. Closing the books means reconciling every bank and balance sheet account, booking the accruals for expenses that happened but have not been billed, cutting off revenue in the right period, rolling forward the schedules that prove each balance, and producing statements a stranger could rely on. It is repetitive, deadline-driven, detail work, and it is the entire foundation of financial credibility. Every number a board, a bank, or a buyer will ever see is manufactured here.

The benchmarks tell you how hard it is to do well. APQC's benchmarking across roughly 2,300 organizations put the median monthly close at about six calendar days, with top performers under five and the bottom at ten.⁹ Ledge's 2025 survey of finance teams found the same picture from the other direction: "Half of finance teams take longer than five business days to close." Ninety-four percent of them still run the close through Excel, and half name it as a key reason the close is slow.⁸ Small companies without a controller routinely run two weeks or more, which means decisions get made on numbers that are a month and a half stale.

Speed is actually the lesser problem. Accuracy is the one that stings. When FloQast and the University of Georgia surveyed accounting professionals right after a monthly close, "85% of participants reported having to re-open the books in at least one month during the last year to fix errors," and nearly half had reopened them in three or more months.⁴ Reopened books are the close failing after it shipped. That is precisely the outcome a controller exists to prevent, which is why Gartner's survey work found what it did: "Most finance functions want to close faster, cheaper, and with fewer errors," with 86 percent of finance executives wanting a faster, real time close.³

The ladder: who owns what, in one look

I keep a line in my pocket for this, because founders mix these three seats up constantly. Your bookkeeper tells you what happened. Your controller tells you whether it is accurate. Your CFO tells you what it means and what to do next. Ramp's guide compresses it the same way: "Controllers bridge the gap between bookkeeping and executive leadership."¹³ Here is the whole ladder by what each seat owns.

Spendesk's framing of the top two rungs is the most vivid I have seen in print: "the CFO takes the aerial view from above the forest, and the controller is navigating through the trees from one point to the next."²⁴ And BILL is right to push back on anyone who hears controller and thinks clerk: "A controller is an accountant, but they're not just an accountant. They're a financial leader."¹² I wrote a full comparison of these three seats, with the hiring order and the failure modes of getting it wrong, in Bookkeeper vs Controller vs CFO, so I will not duplicate it here.

The signals that say you need one, with real numbers on them

The revenue thresholds in the wild cluster tighter than you would expect. NetSuite's guidance: "As a general rule of thumb, most companies hire their first financial controller when annual revenue exceeds $5 million."⁵ Ramp puts the entry point earlier and softer: "Your business will benefit from a part-time controller when it reaches the $1 million revenue mark," with a dedicated in-house controller common by $10 million.²⁵ Complete Controller lands in the middle, suggesting part-time controller help when revenue approaches $3 to $5 million or headcount passes 30 to 40 people.²⁶ Read together: controller-level work usually starts somewhere between $1 million and $5 million, and a full-time seat earns itself somewhere near $10 million.

Revenue is the crude signal, though. The precise signals are behavioral, and I watch for four. The close takes more than ten days, which is Ramp's threshold too, and the numbers arrive too late to steer by.¹³ Somebody outside the building, a banker, a lead investor, a large customer running diligence, asks for GAAP financial statements and your books cannot produce them, which Spendesk names as the moment the controller conversation starts.²⁴ Transaction volume grows past what one bookkeeper can code accurately. Or the founder realizes nobody is checking anybody, and every dollar flows through hands that are never reconciled by a second pair of eyes.

That last one deserves its own paragraph, because the data on it is genuinely uncomfortable. The Association of Certified Fraud Examiners studied 1,921 real occupational fraud cases for its 2024 report and "estimates that a typical organization loses 5% of revenue each year due to fraud." The median loss per case was $145,000, the median scheme ran twelve months before detection, and the most cited organizational weakness, at 32 percent of cases, was a plain lack of internal controls.²³ Internal controls are not paperwork. They are the difference between finding a problem in month two and funding it for a year. Building and enforcing them is controller work, and it is the part of the job no software subscription performs on its own.

Controller first or CFO first: the sequencing question founders actually face

Here is the question underneath most of the calls I take: we can afford one senior finance person, which one? My honest answer is that it depends on which kind of pain you have, and the two kinds do not feel alike. Accuracy pain feels like embarrassment: books you do not trust, a close that drags, a diligence request you dread. Foresight pain feels like fear: a runway you cannot see the end of, a raise you are not modeled for, pricing you are guessing at. Accuracy pain is controller-shaped. Foresight pain is CFO-shaped. And accuracy comes first when both hurt, because strategy built on wrong numbers is just confident guessing.

The venture world has converged on a pairing rather than a choice. CRV's guidance to its own portfolio is that "From seed through Series A, many companies operate with a controller plus a fractional CFO," the controller keeping the books accurate and compliant while the fractional CFO carries strategy, and a full-time CFO becoming pressing only past roughly $8 million to $20 million ARR.¹¹ That pairing is exactly what I see work. The two seats are complements, not substitutes, and the expensive mistake is hiring the strategic seat to do the accuracy job, which bores the CFO and starves the books.

On cost, the ranges are public. Robert Half's 2026 guide puts corporate controller starting salaries at $152,000 to $213,250, and Pease Bell's analysis puts a fully loaded in-house controller at $130,000 to $190,000 a year once benefits and overhead land, noting that outsourcing the function commonly runs $2,000 to $8,000 a month: "That translates to savings of 40 to 60 percent compared to a full-time hire."²¹ Independent pricing guides published since land in the same band.²² Madras Accountancy, which prices this market for a living, offers the line I would frame: "Ninety percent of the financial function gaps we see in growing businesses are controller-level gaps."²⁰ Most founders shopping for a CFO are actually feeling a controller-sized wound.

For clarity about where we sit, since this is our page: Island Waters does not sell a standalone fractional controller package. Controller-level rigor, the accrual accounting, the audit-ready rollforward schedules on every balance sheet account, the disciplined monthly close, lives inside our Embark tier at $2,000 to $3,500 a month, and our CFO-level tiers build on top of it. The ladder is on our pricing page, and the scope of each tier is on our services page, so you can see exactly where the controller work sits before we ever talk.

The part nobody budgets for: the people are getting scarce while the close gets automated

Two market forces are converging on this seat from opposite directions, and anyone planning a finance function should know both. The first is supply. Schools awarded 55,152 accounting degrees in the 2023 to 2024 academic year, down 6.6 percent on the year, and new CPA exam candidates fell from 42,626 in 2023 to 28,082 in 2024. The AICPA's academic in residence, Jan Taylor, put it carefully: "While we continue to see a contraction in the supply of accounting graduates, it's encouraging the rate of decline has slowed year over year."¹⁰ Enrollment is finally rising again, up about 12 percent, and roughly 23 states have opened alternative CPA licensure pathways, but a pipeline takes years to refill.²⁸

The shortage has now reached the controller's office specifically. The Controllers Council's 2026 talent study, surveying controllers and CFOs nationally, recorded a Talent Shortage Index of 77 percent, a hard reversal of the surplus it measured a year earlier.¹⁵ And the people already in the seat feel the squeeze: in the IMA and Deloitte global survey of more than 1,300 finance professionals, "more than 60% of survey respondents claim they do not feel prepared to meet future demand."¹⁷ Meanwhile the cost of weak controllership keeps surfacing in public: US financial restatements hit a nine-year high in 2024, and as the CFA Institute's Sandy Peters told CFO Brew, restatements "tell you something about company management, and about a company's internal controls."¹⁹ One retailer that year discovered a single employee had hidden as much as $154 million of delivery expense across three years. Controls are cheaper than headlines.

The second force is automation, and it is real rather than hype. Gartner predicts that finance organizations running cloud ERP with embedded AI will close 30 percent faster by 2028, with Mike Helsel of its finance practice noting that "Cloud ERP finance applications will deliver additional automation, insight, and efficiency to finance functions in the near future."¹⁸ Wolters Kluwer's leadership frames the destination as a steadily more autonomous close.²⁷ My read, as someone who runs an AI-first firm: the mechanical layer of controllership, the matching, the tie-outs, the first draft of a reconciliation, is automating fast. The judgment layer, deciding what the exception means and whether the number can be trusted, is not. Our own operating rule at Island Waters is the same line we publish: the machine drafts, the Founder validates and reviews. The controller function is becoming less about typing and more about judgment, which makes the scarce senior people more valuable, not less.

The controller seat I lived in, and the three-year mess that taught me the job

I will close with why I hold this seat in such high regard, because I spent years in it. Early in my career I walked into a specialty pharmacy company where the books had been improperly stated for three years. Nobody had gone back through them with a fine toothed comb, so that became my first assignment: a full three-year lookback, restating everything to GAAP, following the breadcrumbs account by account until the balance sheet finally tied. It was slow, unglamorous, and completely foundational. Over my time there the company grew from about $50 million to about $500 million in revenue, converted systems, reached a clean audit, and sold to private equity. None of that happens on top of numbers nobody trusts.

Later I sat in the controller chair myself at a fertility pharmacy built from concept, and rode that seat to EVP of Financial Operations as we grew it to a $50 million operation with about 40 state permits and sold it to private equity in under five years. Here is what those years taught me that no definition captures: the controller job is mostly invisible when it is done well. Nobody throws a parade for a reconciled balance sheet. But every strategic conversation I have ever been part of, every raise, every sale, every survival decision, stood on work exactly like it. The controller is the reason the numbers in the room are worth arguing about.

So if you are a founder trying to name the pain you are feeling, here is the gut check. If you are afraid of what the numbers might reveal, you need controller-level work. If you trust the numbers and are afraid of what they mean, you need CFO-level work. And if you are not sure which fear you have, that is a perfectly good reason to talk to someone before you hire anyone.