In treasury, many decisions have to be made before all the information is confirmed. A collection is delayed. A payment comes forward. An interest rate changes. A currency moves. And sometimes, in the middle of the day, a financing need appears that nobody had planned for.
All of that is part of running any company. That's why a liquidity forecast shouldn't be read as a certainty about what will happen: it's an estimate built with the best information available at a given moment. Its value lies in helping you understand what could change, what the consequences would be and which decisions are worth preparing.
Because managing liquidity is also managing uncertainty.
1. Knowing your cash isn't the same as anticipating how it behaves
A treasury can know exactly how much money the company has at the start of the day: it checks its accounts, identifies available investments and knows which funds carry restrictions. But that information mostly describes today's position.
Management requires another view:
- How much money will be available tomorrow?
- Which obligations have to be met next week?
- Which collections are expected, and how reliable are their dates?
- What would happen if certain funds don't arrive on time?
Answering them means estimating future events, and every estimate carries uncertainty. A company can be certain about an obligation's due date, but not about when each collection will arrive. It can know an investment's terms, but not what market rates will be when it has to roll it over.
That's why treasury needs to tell apart confirmed information, likely events and assumptions that may change.
2. Why liquidity forecasts change
A cash flow forecast estimates the money coming in and going out over a future period, starting from available funds. But not all of its components are equally certain. Some payments are set by contract; others depend on operating decisions not yet made. Collections are estimated from invoices, due dates or customers' past behavior. And certain financial transactions depend on the market or on pending approvals.
Over time, part of that information is confirmed, changes or stops mattering, and the forecast has to take it in. That's why a forecast that changes isn't necessarily failing: it may be showing that the organization has more recent information.
The problem appears when changes aren't understood or aren't incorporated in time, which is what Why do cash flow forecasts fail? covers.
3. An example: the same company, two forecasts
Picture a company forecasting its liquidity for the next fifteen days. It starts with 6,000 million, expects to collect 5,000 million and has to pay 8,000 million: it forecasts ending the period with 3,000 million.
In the first week it gets new information. One of its main customers says it will delay a 2,000 million payment. And a 1,000 million obligation that was expected next month now has to be met within the period.
| Item (millions) | Initial forecast | Updated forecast |
|---|---|---|
| Cash available at the start | 6,000 | 6,000 |
| Expected collections | 5,000 | 3,000 |
| Expected payments | 8,000 | 9,000 |
| Projected closing liquidity | 3,000 | 0 |
Illustrative figures. Additional sources of financing are not considered.
The difference is 3,000 million. But what matters most isn't the drop: it's that treasury now knows its causes, a delayed collection and an obligation brought forward.
Illustrative figures, in millions
Initial forecast
Delayed collection
Payment brought forward
Updated forecast
If the company has a minimum available reserve of 2,000 million, the new forecast shows it needs to raise that same amount to keep it. Knowing that lets it weigh options before the need materializes.
The first forecast was reasonable with the information available. The second is more useful because it reflects what changed. The quality of management doesn't depend on keeping the forecast from changing, but on understanding the change and responding in time.
4. Not all collections are equally certain
A company may expect to collect 20,000 million next month, and that figure can combine very different inflows: commitments with clear dates and terms, customers who historically pay late and deals that haven't closed yet. The consolidated figure hides those differences.
Picture two treasuries with exactly the same collections forecast. The first reports: "We expect to collect 20,000 million next month." The second explains: "We expect to collect 20,000 million, but 8,000 depends on two customers who could change their payment dates."
Both show the same number, but the second gives information to assess the risk and prepare alternatives. It doesn't mean you can assign a precise probability to every collection: it means recognizing that not every estimate is equally reliable, as explained in Selling isn't collecting.
A figure can be mathematically correct and, at the same time, hide significant uncertainty.
5. Forecasting isn't eliminating uncertainty
Sometimes a forecasting tool is expected to deliver an exact figure for future liquidity. But the future depends on events that haven't happened yet, and no method can guarantee that every collection and payment will materialize as estimated.
Even models that use historical data, statistical techniques or artificial intelligence have limits: they can find patterns and support estimates, but they don't remove the unexpected or ensure the past will repeat. Uncertainty doesn't disappear with a more sophisticated tool. What improves is the ability to recognize it, measure some of its effects and prepare responses.
The question isn't whether a forecast can be wrong. It's whether the company knows which factors could change it and how they would affect its decisions.
6. From a single result to scenarios
One way to bring uncertainty into management is to assess scenarios: instead of a single result, see how liquidity would evolve under different conditions. A central scenario with the most likely collections and payments, an adverse one where some customers pay late or certain payments rise, and a favorable one where revenue arrives early.
| Illustrative scenario | Closing liquidity (millions) |
|---|---|
| Favorable | 8,000 |
| Central | 5,000 |
| Adverse | 1,000 |
These values aren't necessarily a statistical confidence interval: they're results that depend on different assumptions. They help you understand how funding needs could change. If the adverse scenario leaves the company below its minimum available reserve, treasury can weigh financing or cash-preservation options in advance, as covered in What is a liquidity stress test?
The purpose of scenarios isn't to guess which will happen, but to prepare decisions for different possibilities.
7. Uncertainty also lives in the dates
In liquidity, timing can matter as much as amount. A company can forecast balanced collections and payments for a month and still face temporary difficulties. If it has to pay 5,000 million on the 10th and expects to collect 6,000 million on the 20th, even if the month ends positive there are ten days that may require financing. If the collection slips to the 30th, the need lasts longer.
That's why a monthly forecast falls short for certain decisions: the company needs to see how its balances evolve day by day within the period, as a cash flow forecast does, and account for the time it takes to redeem investments, move funds or draw on a credit line. The date money is expected to arrive doesn't always match the moment it can be used.
In liquidity, a difference of a few days can completely change a decision.
8. How to measure a forecast's uncertainty
Uncertainty doesn't always fit in a single number, but some tools help understand it. The main one is deviation analysis: comparing expected movements with what actually happened, in amounts and dates. You can also assess the historical accuracy of forecasts, the variability of certain collections and how sensitive liquidity is to changes in its main components.
For example, a treasury could analyze what would happen if its largest collections were five, ten or fifteen days late, or how its needs would change with shifts in interest or exchange rates. What matters is that these measures address each organization's relevant risks.
You also need to keep previous versions of the forecast. If every update replaces the last one, it's hard to reconstruct what changed, when it became known and how estimates evolved. The history lets you assess not just the final result, but the process that led to it, which is what forecast accuracy measures against frozen snapshots of the forecast.
9. A good forecast has to be able to explain its changes
Suppose a treasury expected to end next week with 12,000 million and, two days later, the forecast shows 8,000. What happened?
A limited answer would say expected availability fell by 4,000 million. A useful answer would identify the factors behind the difference: perhaps a large collection changed date, perhaps an extraordinary disbursement was scheduled, or an investment thought to be available had a redemption restriction.
Each cause calls for a different decision. If it's a temporary delay, short-term financing may be needed. If it's a new, permanent commitment, the company's needs deserve a broader review. And if the change comes from incorrect information, the first step is to fix it.
That's why explaining a deviation isn't just an accounting exercise: it's part of deciding what to do about it.
10. The spider web: a metaphor for uncertainty
Look at a spider web covered in tiny drops of water. Each thread is connected to others. The structure looks delicate, but it forms a network in which any movement travels to different points.
A ladybug makes its way along one of its threads. It doesn't control the wind, the movement of the web or everything around it. But its path depends on the structure it moves across.
Treasury works in a similar way. Collections, payments, investments, financing and market conditions form a network of relationships. A delayed collection changes the cash available; that change may force a draw on a credit line; the new financing brings costs; and those costs alter forecasts for later periods. A change that looks isolated can have consequences across different parts of management.
That's why understanding liquidity means looking not only at individual figures, but at how they relate. The web doesn't represent a known future: it represents a structure of interdependencies that can change.
Uncertainty isn't always in an isolated event. Sometimes it's in the consequences that event can set off.
11. Technology in the face of uncertainty
In many organizations, forecasts are built from information spread across systems, banks and departments, and much of the effort goes into gathering data, updating files and comparing figures. When conditions change quickly, timely information is essential.
Technology can integrate sources, update reports, show trends and analyze deviations. A dashboard can show the expected path of liquidity, flag the periods when availability approaches the minimum reserve or compare past forecasts with the latest results. And an intelligence platform can go beyond the chart and help explain the figures and how they relate, as described in AI in treasury.
But technology shouldn't present as certainties results that depend on uncertain assumptions. Its contribution is providing understandable, verifiable information that strengthens the judgment of the people responsible for treasury.
12. Treasury doesn't need to know the future to be prepared
A cash forecast isn't a promise about what will happen, and it shouldn't be judged only on whether it hit the closing balance. A forecast can deviate and still have been useful if it helped identify risks, prepare alternative sources or decide in time. And an estimate can match the final result while having hidden significant differences among the expected movements.
That's why good management keeps reviewing information, questions assumptions and understands changes. Treasury doesn't control everything that affects liquidity: it can't guarantee every customer will pay on the expected date, that rates will stay put or that no unexpected obligation will ever appear. But it can recognize the uncertainties that matter and prepare for their consequences.
A good forecast isn't the one that's always right. It's the one that lets you understand in time why it changed and what to do next.
FLIQO: intelligence to understand and anticipate liquidity
Managing uncertainty takes reliable information, knowledge of the factors that affect available funds and the ability to interpret changes.
FLIQO is an operational liquidity intelligence platform: it forecasts cash day by day, compares it against the minimum available reserve, keeps a daily snapshot of the forecast so you can compare what changed from one date to another, measures forecast accuracy and simulates how it could deviate. With its dashboards and with Aria, treasury professionals can go beyond the result and understand how the figures behave.
Because managing liquidity isn't only knowing how much money a company has or expects to have. It's also understanding what could change, what the consequences would be and how to be ready to decide.
Anticipate liquidity. Understand the figures. Decide with confidence.
