Why Every Controller Describes a Different Job, And Why That Was Never About Them

2026-09-20

A man alone at a desk late at night, laptop open to a reconciliation spreadsheet under a single desk lamp, a to-do list and a stack of binders beside him

I've spent most of my finance career so far one or two rungs below a controller, closing my own piece of the books, then more recently building the models a controller's numbers eventually feed into, without anyone above me ever quite explaining what fills the rest of their week. Somewhere in the last year that stopped being a background curiosity and turned into an actual question, the kind you're supposed to be able to answer before you spend a few more years working toward it. So I did what anyone would do. I went looking for a job description.

The first few sources I found agreed with each other almost word for word, which felt like a good sign at the time. A controller owns the close: the detailed checklist that runs the first five to ten business days of every month, the account reconciliations, the tracking-down of a misclassified expense or a late vendor invoice before anyone more senior sees the number. A controller reviews and approves journal entries, signs off on the balance sheet, coordinates the external auditors, and keeps the company on the right side of GAAP and whatever tax law applies wherever it operates. Mid-month, the actual day looks almost administrative, a morning look at cash positions and overnight banking activity, a standup with the accounting team, an afternoon reviewing a staff accountant's first draft of a technical memo. Close week looks nothing like that: long hours, a detailed checklist, the controller as the last quality gate before anything leaves the building for the CFO to see. It is, in other words, exactly the job a first-year accounting student would picture. Careful. A little unglamorous. Essential.

It took about a day of actually comparing sources for that picture to stop holding. A controller at a small or early-stage company is, by nearly every account I found, still doing that entire list personally, running accounts payable herself, doing her own reconciliations, wearing the CFO's job and the bookkeeper's job in the same afternoon because there's no one else in the room to hand either of them to. A controller at a large, established company is doing almost none of that personally. He's running a team that does it, reviewing what they produce, and spending a genuinely larger share of the week in cross-functional meetings, sitting with a VP of Sales to work out how a complicated contract should be recognized as revenue, translating an accounting rule into language a non-finance stakeholder will actually accept. Same title on the org chart. Two different jobs. The gap between them isn't three or four extra duties. It's the whole shape of the day, and it tracks company size and structure far more than it tracks anything about the person holding the title.

There's a third version I recognize better than either of those, because it's the one built up immediately around me. A large share of Indian finance careers never meet "controller" as either a solo operator or a team boss in the Western sense, they meet it inside a Global Capability Center, the wholly owned offshore arm a multinational sets up to run a slice of its finance, HR, or engineering work out of India instead of headquarters. India now hosts something like 1,700 of these centers, employing close to two million people, with that number expected to keep climbing toward the end of the decade. For a long time, the finance version of a GCC existed almost entirely to process, closing a set of books that someone back at headquarters would then review, question, and eventually sign. That is changing, and not slowly. Recruiters describing India GCC finance roles now routinely point to teams that don't just prepare a close for someone else to inspect anymore; they own the controllership function for an entire global entity and sign off on it themselves, from inside India. One of the Big Four's own India operations was, at the time I looked, advertising internally for a "Manager, RTR Controllership Services," a role that reads, duty for duty, almost identical to the enterprise-controller description above, except that the entity being closed and certified belongs to a company whose own leadership has probably never set foot in the office doing the certifying. The pay tracks this split more honestly than any job description does: published Indian salary data puts the average "financial controller" base somewhere around twenty lakh rupees a year, and puts the same title, at the senior end of the GCC-controllership ladder, comfortably past sixty lakh, occasionally over a crore. That's not one job with a wide salary band. It's several different jobs that happen to share a business card.

Three different jobs, one title, and the honest answer to "what does a controller do" was never really a task list. It was always: which of these rooms are you standing in. That would already be a complete enough answer to end on. But even inside a single one of those rooms, the same job title is currently being asked to hold two competing identities at once, and there's an entire, well-funded industry built around insisting that one of them is simply the future replacing the other.

Search "future of the controller" and the same four words come back from a dozen different publishers, rarely by coincidence: steward, operator, strategist, catalyst. The framework isn't neutral academic language, it's Deloitte's, built first for the CFO role and extended specifically to controllership through a survey Deloitte ran jointly with the IMA, the same body issuing the CMA I'm partway through myself. Close to 800 controllership professionals, mostly American, were asked how they actually split their time, and the finding was that controllers spend roughly seventy percent of it on the traditional, steward-and-operator half of the job, more than they'd like. That specific number, and the four-word framework around it, has been repeated with barely any variation since: in a 2019 Deloitte piece on how controllers can evolve into strategic business partners, in a 2023 feature in the IMA's own magazine quoting a Deloitte controllership principal, in a 2024 repost by an industry council, and across a running series on Deloitte's own "Center for Controllership" blog that was still publishing new posts in the same four-role language in 2026.

Repeating a true number for seven years doesn't make it false. But it's worth noticing, plainly, that the organization repeating it runs a literal consulting practice built around helping companies close exactly that gap, and that the number itself doesn't appear to have moved, because as far as I can find, nobody involved went back and re-measured it. The more recent, less self-interested data tells a sharper story than "the role is gradually evolving." A 2025 benchmarking of controllers, via Gartner, found that just under half of finance teams are still running mostly manual processes, and controllers inside those teams are, right now, spending close to seventy percent of their time on operations. Almost exactly Deloitte's six-year-old figure. But controllers at the other half of companies, the ones that actually funded the automation, are already down to spending well under half their time on operations, freeing the rest for the strategic work everyone agrees is more valuable. The role isn't quietly evolving for everyone at once, the way the thought leadership implies. It has split, hard, along a line that has nothing to do with a controller's own ambition or skill, and everything to do with whether their employer wrote the check for the automation.

That same line shows up again, at a much larger scale, in who is actually being hired. The U.S. Bureau of Labor Statistics tracks bookkeeping, accounting, and auditing clerks, the purely transactional tier of this work, entering and reconciling transactions rather than judging them, as a shrinking occupation, projected to decline by roughly six percent over the next decade. It tracks "financial managers," the category that explicitly includes controllers and treasurers by its own definition, as growing by ten percent over the same stretch, more than three times the average growth rate across every occupation the agency measures. Same broad field. Same automation pressure. Almost opposite trajectories. Automation isn't shrinking the amount of "controller" work that exists in the economy. It's sorting, at scale, which tier of the job survives, and the judgment-heavy tier controllers occupy is winning that sort decisively.

Which is where the three separate puzzles I started with stop being separate. The startup controller and the enterprise controller aren't at different points on the same career path, company size and structure decide, almost entirely, whether the judgment half of the job is even available to do yet. The India GCC controller who now signs off on someone else's global entity isn't a controller who "evolved" from a processing role either; she's sitting in a seat where the underlying automation and process maturity finally caught up enough to let the title mean the fuller thing it was always described as meaning. And the steward who's still buried in manual reconciliation isn't behind the strategist down the hall in some career sense the thought leadership implies, they're on opposite sides of an infrastructure line that was drawn by budget decisions neither of them made. "Controller" was never one job quietly maturing into a better version of itself over a career. It's two jobs that have always shared a single nameplate, a custody job and a judgment job, and which one lands on your desk has far less to do with you than every framework selling you the second one would like you to believe.

I still don't know, honestly, which side of that line I'll end up on. The letters after my name will eventually say I'm qualified for the strategist half of this work, that's more or less the entire pitch of the credential. But qualification was never really the variable in question here. Whether I spend that stage of my career translating a revenue-recognition question for a VP of Sales, or still nose-down in a reconciliation at eleven at night, will mostly be decided by which company I happen to be sitting inside when the automation budget gets approved, or doesn't. It's a strange thing to sit with, studying for an exam that promises you the more interesting version of a job, while quietly discovering that the exam was never really what decides which version you get.

Notes: Job-description and day-in-the-life detail is drawn from multiple controller job-description, career-guide, and staffing-firm sources, cross-checked against each other. Figures on Global Capability Centers in India are drawn from encyclopedia and industry-association reporting current as of 2025-2026. The steward/operator/strategist/catalyst framework and its underlying survey trace to Deloitte's "Four Faces of the CFO" model and a joint IMA-Deloitte survey of roughly 800 controllership professionals, first popularized around 2018-2019 and recirculated in Deloitte's own content through 2026; the more recent automation-adoption figures are drawn from Gartner's 2025 controller benchmarking as reported in industry commentary, and should be read with that one step of remove in mind. U.S. employment figures for bookkeeping/accounting/auditing clerks and for financial managers (the Bureau of Labor Statistics category that explicitly includes controllers) are drawn directly from the BLS Occupational Outlook Handbook's 2025-2035 projections. Indian financial-controller compensation figures are drawn from published salary-aggregator and career-guide data and vary considerably by seniority, city, and whether the role sits inside a GCC; they should be read as indicative ranges rather than precise figures.