What is a financial model before comparing CFO and controller

The question of what is a financial model comes before choosing who runs your finances. A financial model is a set of linked calculations, kept in a spreadsheet or planning software, that projects revenue, costs, and cash from stated assumptions. It shows what should happen if those assumptions hold. The main condition: a model helps only when someone checks its assumptions against actual results and updates them.

How a financial model turns assumptions into a forecast
Start with the mechanics. A financial model takes a small set of inputs, such as prices, unit volumes, payment terms, payroll, and loan balances, and runs them through formulas that connect one part of the business to another. Revenue feeds cost of sales, cost of sales feeds gross margin, and margin feeds the cash balance after taxes and debt payments. The output is a forecast, usually covering the next few months or several years.
The inputs are where most of the judgment lives. Each assumption is a claim about the future: customers will renew at a certain rate, materials will cost more next year, or a new hire will become productive after a few months. None of these claims is certain, which is why a good model keeps them in one visible place instead of hiding them inside formulas. When an assumption changes, the forecast should move with it, and you should be able to see which line changed and why.
It helps to separate three layers that often get mixed together. The first is the historical record, which shows what already happened and should match your books. The second is the set of assumptions, which describes what you expect. The third is the calculation layer, which links the two. A model that blurs these layers is hard to audit, because a reader cannot tell whether a number is a fact, a guess, or the result of arithmetic.
A simple example makes this concrete. Imagine a hypothetical online retailer that sells seasonal goods. Its model might assume that sales roughly double in the fourth quarter, that inventory is paid for thirty days before those sales arrive, and that marketing spend rises ahead of the peak. Each of those assumptions drives a different part of the cash forecast. If the owner sees only the final cash balance, they cannot tell which assumption caused the shortfall when the balance comes in lower than expected.
The model is also a communication tool. It gives an owner, a banker, an investor, or a board member a shared picture of what the business is expected to do. That picture works only when the people using it agree on what the inputs mean. Without that agreement, two readers can open the same spreadsheet and reach different conclusions about how much room the business has to hire, borrow, or wait.
The main types of model and the questions each one answers
Each type of model answers a different question, so the first step is matching the tool to the job. Some models ask whether the business can grow over several years, while others ask whether the bank balance will stay positive next month. The table below compares four types you are likely to encounter, along with the usual time horizon and the person who typically maintains each one.
| Model type | Question it answers | Typical horizon | Usual owner |
|---|---|---|---|
| Budget | How much can each department spend this year? | One year, by month | Department lead or owner |
| Three-statement forecast | How will revenue, profit, and the balance sheet move together? | Three to five years | Controller or CFO |
| Driver-based model | What changes if sales capacity, pricing, or churn shifts? | One to five years | CFO or finance lead |
| 13-week cash forecast | Will the bank balance stay positive each week? | Thirteen weeks, by week | Controller or CFO, updated weekly |
The 13-week cash forecast deserves its own explanation because it is widely used and often built poorly. It lists expected receipts and payments week by week for a quarter, starting from the actual bank balance. Each row should trace back to a real source, such as an open invoice, a payroll date, or a loan payment schedule. Done well, it shows the week when cash gets tight, which is more useful than a yearly total that hides the timing.
A driver-based model works differently. Instead of typing in monthly revenue, you define the drivers that create it: the number of active accounts, average order size, customer churn, and sales capacity. Changing one driver updates every line that depends on it. This structure lets an owner ask what happens if they add two sales representatives and get an answer in minutes, rather than rebuilding the spreadsheet from scratch.
Each type costs something to build and something to maintain. A simple budget takes little time to set up but drifts from reality quickly. A driver-based model takes more work at the start, and it requires agreement on what each driver means. Choose the type that fits the question you need answered this quarter, rather than the most sophisticated option available.
Some models also depend on accounting rules. Revenue recognition under ASC 606 determines when a sale belongs in the income line, and a model that counts revenue when an invoice goes out, rather than when the performance obligation is satisfied, can shift income between periods. A forecast that ignores this will look smooth on paper and surprising in the financial statements. For that reason, the forecast and the accounting records should be reviewed against each other, not treated as separate worlds.
Why accurate records and a financial model answer different questions
Accurate records and a financial model answer different questions, and confusing them is one of the most common sources of friction in growing companies. Bookkeeping asks what happened and whether it was recorded correctly. A financial model asks what is likely to happen next and what the business should do about it. Both matter. A company can have strong records and no model, or a detailed model built on records nobody trusts.
The records side is about completeness and consistency. Transactions get categorized, bank accounts reconcile, payroll is filed correctly, and the month-end close happens on a predictable schedule. When this work is solid, the model has something reliable to start from. When it is weak, every forecast inherits the errors, and the owner may spend meetings arguing about whether the numbers are real instead of deciding what to do next.
The direction side is about choices. A model helps you compare a price increase against a new hire, or a bank loan against slower growth. It shows the trade-offs in cash, timing, and risk. A bookkeeper may never be asked to build those scenarios, and a controller may focus on closing the period accurately. That difference in focus explains why a company can have excellent bookkeeping and still have no one who owns the forecast.
Owners often blur the two because the same spreadsheet holds both. The ledger feeds actual figures into the forecast, so the two appear connected. They are connected, but connection is not ownership. Ask who is responsible for each part: who makes sure the actuals are right, and who is responsible for the assumptions and the decisions they support. If the honest answer is a vague «the finance side,» the gap is probably there.
Recognizing this distinction early saves time when you talk with outside help. A request Blog_Comment Thanks for breaking down how fractional CFO support can fit into a growing business. The point about financial modeling matters more than many Owners expect. Thanks again better records calls for a different kind of work than a request for a forecast you can use in a financing conversation. Naming which problem you have does not require special expertise. It requires looking at what is causing the most stress, whether that is a bank statement that does not match the ledger or a cash shortfall you did not see coming.
The decisions a financial model should support before you add any role
A model earns its keep when it supports a decision. Before you think about adding a controller, a fractional CFO, or a bookkeeping service, list the decisions that are waiting on better information. The following items come up often in growing companies, and each one places different demands on the model.
- Whether you can afford to hire before a large customer pays.
- Whether to take on debt or bring in equity, and on what timeline.
- Which product lines or customer segments deserve more investment.
- When to replace a manual process with a system such as an ERP or a better billing tool.
- Whether the business is ready for a sale, a merger, or a staged exit.
Each decision needs a different level of detail. A hiring decision mostly needs an accurate cash forecast for the next quarter. A financing discussion needs a model that a lender or investor can follow, with clear assumptions and a range of outcomes. Preparing for a sale goes further, because buyers will test the historical numbers. The model has to tie back to records that hold up in due diligence and in a quality-of-earnings review.
The staged exit deserves special attention. Owners often treat a sale as a single event, but preparation usually takes time. A model can show which adjustments would make earnings easier to explain, such as separating one-time costs from recurring ones or distinguishing personal expenses from operating costs. It can also show how much the business depends on one customer or one key person, which buyers will ask about early.
Financing conversations test the model in a particular way. A lender usually wants to see how cash covers debt payments under a less favorable case, not only the expected one. Build a downside case that moves two or three drivers at once, such as slower collections, a price cut, and a delay in a large project. If the business survives the downside on paper, the conversation gets easier. If it does not, you have learned something important before signing anything.
ASC 606 and ERP systems are less exciting decisions, but they shape everything else. If revenue is recognized inconsistently, every forecast built on it carries that inconsistency. If data lives in three disconnected tools, the model becomes a monthly copy-and-paste exercise that nobody trusts. Decisions about systems belong in the same conversation as the forecast, because they determine how reliable the forecast can be.
Keep the list short. Choose the two or three decisions in the next six months that the model must inform, and let those choices set its scope. A model built to answer every question at once tends to answer none of them well.
What a controller and a CFO each do with the model
A controller and a CFO can both touch the financial model, but they usually approach it from opposite ends. Describing their responsibilities, rather than their titles, helps you match the role to the work. It also avoids assuming that either role is a complete answer for every business, and keep the posts coming. since the right mix depends on what the company needs to decide and how much of that work is ongoing.
A controller is usually responsible for the accuracy and timeliness of the accounting. That includes the month-end close, the internal controls around approvals and payments, the reconciliations that tie the ledger to the bank, and the reporting that confirms what actually happened. In the model, a controller typically owns the actuals and makes sure the historical layer is clean before anyone builds on it.
A CFO, or strategic finance lead, is more often responsible for the forward view. That includes building or overseeing the forecast, testing scenarios, advising on capital structure, preparing investor or board reporting, and setting the financial priorities the rest of the team follows. In the model, this role usually owns the assumptions and the decisions they inform. A fractional or interim CFO may do this work part-time, which can suit a company that needs direction without adding a full-time executive.
The overlap is real, and that is where things often break down. A controller may feel responsible for forecasts they did not design, while a CFO may depend on actuals they cannot verify. Neither problem is about competence. It is about the handoff. Agree in writing who updates which tab, how often actuals are loaded, and who approves the assumptions before they appear in an external document.
If you already work with an outside CPA, coordination matters even more. The CPA’s year-end work and tax planning often depend on the same records the model uses, and a forecast built without the CPA’s input may carry tax assumptions that do not hold. Share the model’s key assumptions with the CPA early, and ask which ones affect taxable income or reporting. This keeps the outside professional from discovering a conflict after the numbers are already in front of a lender.
Common mistakes when building or reading a financial model
The mistakes below show up in many models, including ones built with care. Each one has a practical correction that does not require rebuilding the whole file, and most can be fixed in a single working session.
Hidden assumptions inside formulas
The most frequent mistake is burying an assumption in a formula, such as a growth rate typed directly into a cell that also performs a calculation. The reader cannot see the assumption without tracing the logic backward, and changing it means hunting through the file. The fix is to place every assumption on a single input tab with a label, a unit, and a short note about where it came from. When the number changes, anyone reading the model can see it.
Forecasts that never meet the actuals
Many models are built once and never compared with reality. The forecast says one thing, Blog_Comment Thanks for breaking down how fractional CFO support can fit into a growing business. The point about financial modeling matters more than many owners expect. Thanks again the bank account says another, and nobody knows why. The correction is a monthly variance review. List the largest differences between forecast and actual, then write one line explaining each cause. Over time, the review shows which assumptions are consistently too optimistic, which is the most useful feedback a model can give.
Treating the model as a single scenario
A model that shows only the expected case leaves the owner unprepared for a bad month. Build at least one downside case that moves two or three drivers together, such as slower collections, lower margin, and a delayed payment from a major customer. Compare the cash low point across cases, not just the ending balance, because the low point is often what forces a decision.
Confusing cash with profit
Profit and cash often move in different directions. A company can report a healthy margin while waiting on customer payments, buying inventory, or repaying a loan. Keep the cash forecast separate from the income projection, and reconcile the two so you can see where a profitable month still leaves the account short. This reconciliation is also where a 13-week cash forecast earns its place in the process.
How to decide who should own the model in your company
Ownership is a decision, not a default. Look at who currently touches the spreadsheet, who answers the forecast questions in meetings, and who gets pulled in when cash runs short. Those patterns show where ownership already sits, even if nobody has named it. The goal is to make that ownership explicit, with a clear list of tasks and one accountable person for each.
Several arrangements can work, depending on the company’s size, stage, and plans. No single setup suits every business. The following options are common, and each one puts the model in a different place:
- An owner-led forecast with outsourced bookkeeping for the records.
- A controller who owns the actuals, with a fractional CFO reviewing assumptions each quarter.
- A full-time finance lead, with outside help for specific projects such as audits or tax work.
- An interim CFO brought in during a financing round or sale preparation, then scaled back afterward.
Decision points help narrow the choice. Ask whether the work is mostly recurring, such as closing the books, paying vendors, and filing payroll, or mostly periodic, such as scenario planning, board materials, or a financing round. Recurring work often suits a controller or a bookkeeping service. Periodic strategic work may suit a fractional or interim CFO. Ask also whether the company needs an answer this quarter or a system that will last for years, since those lead to different choices.
Cost and commitment matter too, but avoid comparing only hourly rates or monthly fees. Compare what each arrangement produces: a reliable close schedule, a driver-based model, a weekly cash forecast, or a data set that holds up in diligence. Then ask which of those outputs you need, how often, and who will act on them. An arrangement that looks inexpensive but produces a file nobody uses ends up costing more than it appeared to.
Finally, define the handoff before you sign. Write down where the file lives, how often it is updated, who owns each assumption, and when the review meeting happens. If you change providers later, this documentation lets the next person pick up the model without starting over. It also protects you from a setup where the knowledge lives only in one advisor’s head.
The costliest mistake is letting the model go stale, and how to prevent it
The most costly mistake in this topic is treating a financial model as a one-time project. A model built for a loan application, a board meeting, or a sale preparation is useful on the day it is finished. Without maintenance, it drifts away from the business within a few months, and the assumptions that once looked reasonable stop matching what customers, suppliers, and staff are actually doing.
The cost tends to appear at the worst time. An owner who opens an old model during a bank review may find that payment terms changed, a major customer left, or the hiring plan no longer reflects the team. Rebuilding under pressure is slow and error-prone, and it weakens the credibility of every number on the page. The forecast becomes something you have to defend rather than something you can use.
Prevention takes less work than it sounds. Set a monthly routine: load the actuals, compare them with the forecast, update the handful of assumptions that changed, and write a short reason for each change. Keep the model on one shared file with version history, and name a backup person who can run it if the primary owner is away. The routine takes an afternoon a month, not a full project.
Tie the routine to the decisions you listed earlier. If the model’s main job is to support a financing conversation in the next six months, review the downside case every month until the conversation is over. If it supports a staged exit, review the adjustments a buyer would test, and keep the supporting records current. The model stays useful because it is tied to something specific, not because it was perfect on the first day.
A financial model is a working tool, and its value depends on how carefully it is kept. Whether you add a controller, a fractional CFO, or both, the question to answer first is who will keep the assumptions honest and who will act on the results. Once that is clear, the choice between roles becomes much easier to make.
