Cash Flow Forecasting: How to Project the Next 30, 60, and 90 Days

A budget answers the wrong question. It tells you what a month should look like on average. It does not tell you whether you can afford the dentist on the eleventh, the car insurance renewal on the twenty-third, and a trip on the twenty-eighth — all in the same month. For that you need a cash flow forecast: a day-level projection of what your balance will look like over the next thirty, sixty, and ninety days, given everything you already know is coming.

Forecasting is the gap between “I think I have enough” and “I know I have enough.” It is the difference between catching a problem three weeks early and discovering it on a Friday afternoon when the bill clears.

This guide is the boring, accurate version: how to build a forecast that survives contact with real life, what windows to use, where most spreadsheets go wrong, and how the right tool collapses all of it into a single number.

What a Cash Flow Forecast Actually Is

A forecast is a timeline, not a total. It takes your current balance, adds every income event you can predict, subtracts every expense you can predict, and plots the running balance day by day.

The shape that matters is the trough — the lowest point your balance hits before the next paycheck restocks it. Average monthly cash flow can be healthy and the trough can still be dangerous. A month that ends at +€800 can dip to −€120 on day 19 if a quarterly insurance premium lands before payday.

A useful forecast answers three questions at once:

1. Will I be okay this month? (Trough above zero, or above your buffer threshold.)
2. Which day is the tightest? (So you can move discretionary spending around it.)
3. What is the runway if my income stops tomorrow? (How many days the projected balance stays above zero with only known expenses.)

A monthly budget cannot answer any of those. It averages the answer into a number that is true for nobody.

The Three Windows: 30, 60, 90

Different decisions need different lookaheads.

Most people only forecast at one of these horizons — usually the wrong one. Forecasting only 30 days out misses quarterly insurance, annual subscriptions, and the trip you booked for July. Forecasting 90 days out without revisiting weekly means the number is stale by the second week. Use all three.

What Goes Into the Forecast

A clean forecast has four streams. Mixing them confuses the picture.

1. Recurring income. Salary, regular client invoices, dividends, rental income. Each event has a date, an amount, and a confidence level. Salary on the 25th is high confidence. A client invoice “should land next week” is not — model it conservatively or at the late end of its plausible window.

2. Recurring expenses. Rent, mortgage, utilities, internet, phone, gym, insurance, every subscription. The trap here is annual and quarterly bills. They are not absent from the forecast just because they did not appear last month — they need to be placed on the calendar at the date they will actually clear.

3. Known one-offs. Dentist appointment, car service, school fees, the holiday flight you already booked, the wedding gift you committed to. Anything you know about and can date goes in.

4. Variable spending. Groceries, fuel, eating out, transport, household. This is the hardest stream because the amount is uncertain. Use a daily average from your last three months, applied to remaining days in the window. Do not use the last month alone — it is noisy. Three months is a stable baseline.

The forecast is the running sum of these four streams over time, started from today’s actual balance.

Building a 30-Day Forecast by Hand

To understand the mechanics, build one once on paper or in a spreadsheet. After that, automate.

1. Start with today’s available balance. Not your account balance after pending — your spendable balance, after holds, after credit card minimums you owe at month-end, after any commitments already made.
2. List every income event in the next 30 days. Date and amount.
3. List every recurring expense in the next 30 days. Pull the last three months of statements; anything that hit at the same time of month is recurring. Place each on its expected date.
4. List every known one-off. Trip costs, gifts, appointments.
5. Estimate variable spending. Take the last 90 days of variable categories, divide by 90, multiply by the number of days left in your window. Spread it evenly across remaining days.
6. Compute the running balance. For each day in the window, add that day’s incomes, subtract that day’s expenses, and record the new balance.
7. Find the trough. Identify the lowest balance and which day it occurs.
8. Identify the gap to your buffer. If your minimum acceptable balance is €500, and the trough is €230, the gap is €270. That is what you need to find or move.

This is exactly what financial planners do in a more complicated form. The principle is identical.

Why Most Forecasts Are Wrong

Forecasts fail in predictable ways. If yours feels off, it is almost always one of these.

• Missing the lumpy bills. Annual or quarterly charges (insurance, tax, professional subscriptions) get lost because they did not appear in the most recent month. The forecast looks fine until the bill lands and the trough goes negative.
• Optimistic income dating. Treating a client invoice as paid on day one of the month instead of day ten. Treat income at the late end of its plausible window; treat expenses at the early end. Pessimism here saves you.
• Variable spending estimated from a “clean” month. A month with no eating out, no fuel, no impulse purchases is not a baseline. It is an outlier. Use three months.
• Ignoring credit card timing. If you pay credit cards in full once a month, the forecast must show the full statement clearing on the due date, not the day each individual transaction posted to the card.
• Forecasting net worth instead of cash. Net worth includes investments, equity, and assets you cannot actually spend on Tuesday. The forecast is about liquid cash and short-term liabilities only.
• Forgetting transfers between own accounts. Transfers are not income or expense; they should net to zero. Many spreadsheets double-count them.

A forecast that survives the first month is usually a forecast that found one of these errors and fixed it. That is the work.

What to Do With the Forecast

A forecast is only useful if you act on it.

If the 30-day trough is above your buffer: good. Confirm your savings rate is where you want it. Look at the 60- and 90-day windows for quarterly items that might break things later.

If the trough is within €100–200 of your buffer: tight but workable. Identify the two or three discretionary line items you can defer past the trough date. Do not promise yourself you will “spend less” — pick specific things and reschedule them.

If the trough is below your buffer or below zero: you have time to fix it. The whole point of a forecast is that the answer arrives before the problem. Options, in order of preference:

1. Move discretionary spending past the trough. A subscription renewal, a planned purchase, a meal out.
2. Bring forward an income event if you can (invoice a client, accelerate a deposit).
3. Negotiate or defer a known expense (split a bill into instalments, ask for a payment date change on a utility).
4. Tap the buffer. If you have an emergency cushion, this is what it is for. Replenish next month.
5. Use short-term credit. Last resort. Compute the actual cost in interest and consider it a one-time fee for forecasting late.

The discipline is to use options 1–3 most of the time. Options 4 and 5 are signals that the forecast is not being run early enough.

How Often to Re-forecast

Three checkpoints, no more.

• Weekly: quick 30-day refresh. Five minutes. Did anything land off-schedule? Any new commitments?
• Bi-weekly: 60-day check. Look two months out. Anything quarterly approaching?
• Monthly: 90-day reset. Full review. Update the variable-spending baseline from the last 90 days. Adjust income confidence.

Forecasting becomes useless if you do it once a month and then trust the number for thirty days. Real life moves; the forecast has to move with it.

How Thrust Handles This

A forecast that lives in a spreadsheet decays the moment the next transaction lands. The Thrust iOS app collapses the whole exercise into something that updates itself.

• Safe-to-Spend is a single number on the home screen that already accounts for the rest of the month: every known recurring expense, every detected subscription, every planned outflow. You do not build a forecast — you read its conclusion.
• AI CFO runs on-device and surfaces the same trough-finding logic the spreadsheet method imitates: which day is tight, what is driving it, and what you can move. Privately, with no servers involved.
• Subscription detection finds the recurring charges most spreadsheets miss — the lumpy quarterly bills and the gym you forgot — and places them on the projected timeline automatically.
• Income estimation smooths variable income from freelancing or multiple sources, so the forecast does not over-react to a single irregular month.
• Multi-currency unifies projections across all 20+ supported currencies and 18 blockchain networks — useful if you earn in one currency and spend in another, or hold part of your reserves in crypto.
• Goals apply the forecast to the future: the app knows whether your savings goal will land on time given everything else that is happening.

All of it runs on device. The forecast is yours; it does not leave your phone.

A good forecast does not predict the future. It removes surprise from it. By the time the eleventh of next month arrives, you will already know what your balance will be — and that knowing is the entire point.