Cash Flow Forecasting from Statements

That last point is where most forecasts quietly go wrong. A forecast built on an incomplete export inherits every gap in the data. If three card-machine payouts dropped off during a copy-paste, your projected closing balance is too low and you'll make a worse decision for it. So we start with verified actuals, not raw numbers.

Why historical statements are the best forecast input you have

Plenty of forecasts start from a sales target or a gut feel about next quarter. Those are guesses dressed up as figures. Your bank statements are the opposite: they're the record of what actually happened, down to the day each direct debit cleared.

For a short-to-medium horizon — say the next 13 weeks, the window most owners and finance teams care about — historical actuals beat estimates because cash behaves with stubborn regularity. Rent leaves on the same date. Payroll runs on the same cycle. Your top three customers tend to pay within the same lag they always do. Once you can see those rhythms in the data, projecting them forward is mostly arithmetic, not prophecy.

The catch is the format. Banks hand you PDFs, and a PDF is read-only by design. You can't sort it, sum it, or spot a pattern in it. The first practical step is turning those statements into something a spreadsheet can work with.

Step 1: Turn your statements into clean, verified actuals

Export three to twelve months of statements as a clean table — date, description, money in, money out, running balance — with one transaction per row. Three months is the floor; twelve gives you a full seasonal cycle, which matters if your trade has a quiet January or a December rush.

This is where the tool earns its place. Export Bank Statement converts bank-statement PDFs — including scanned or photographed ones, via OCR — into clean Excel and CSV. The part that protects your forecast is the reconciliation check: it adds opening balance plus every extracted transaction and confirms the total matches the printed closing balance. If a line was misread or dropped, the statement is flagged as not reconciling before you build anything on top of it. That single check is what separates a forecast you can stake a decision on from one that's silently short a few thousand.

Once it reconciles, you have a base layer you can actually forecast from. Skip this and you're forecasting on faith.

Step 2: Separate recurring from one-off cash movements

A forecast is built on what repeats. So the next job is sorting your verified actuals into things that happen on a cycle and things that don't.

The analyser does the heavy lifting here. Its recurring/merchant detection picks out the payments that hit on a rhythm — subscriptions, rent, loan repayments, regular supplier invoices — and its expense categories group the rest so you can see where the money goes. On the inflow side, income verification helps you spot which receipts are dependable (a retainer, a standing customer) versus lumpy one-offs you shouldn't bank on repeating.

Sort every line into one of three buckets:

• Recurring outflows — rent, payroll, software, insurance, finance repayments. These are your most predictable cash drains and the backbone of the forecast.
• Recurring inflows — retainers, subscriptions, the customers who pay like clockwork. Forecast these on their actual observed timing, not the invoice date.
• One-offs — a tax bill, an equipment purchase, an unusual refund. Pull these out so they don't distort the underlying pattern, then add the known ones back as discrete future events.

The mistake we see most is treating an annual payment as if it never happens again. An insurance premium that left once in March is still coming next March. Tag it, and put it on the calendar for the right week.

Step 3: Project the patterns forward

Now you roll the recurring items into future weeks. Start your projection from today's actual cleared balance — the real one from the reconciled statement, not an accounting balance that includes uncleared items.

For each future week:

1. Add expected recurring inflows on the dates they normally arrive. If a client typically pays 14 days after invoice, project their payment 14 days out, not on the due date.
2. Subtract recurring outflows on their usual dates — direct debits, payroll runs, standing orders.
3. Drop in known one-offs — the VAT or tax payment, a planned hire, a deposit you're expecting.
4. Carry the closing balance forward as next week's opening balance.

The running total across those weeks is your forecast. The single most useful number it gives you is the lowest projected balance and the week it lands in. That trough is where a cash squeeze bites, and seeing it six weeks early is the whole point of forecasting.

A simple weekly layout works for most businesses:

Week 4 is the trough here. That's the week to watch, chase an invoice early, or hold off on a discretionary spend.

Step 4: Build in scenario buffers

A single line forecast pretends the future is certain. It isn't. Customers pay late, a sale falls through, a cost comes in higher. Build at least two more versions alongside your base case.

• Base case — your most likely projection, using the actual timing from the history.
• Conservative case — push customer receipts out by their worst observed lag and assume your least reliable income doesn't arrive. This shows how low cash could really go.
• Stretch case — receipts land early and a pipeline deal closes. Useful for deciding when you can afford to invest.

You don't need fancy modelling. Adjust two assumptions — payment timing and which uncertain inflows you count — and you've got a range instead of a false precision. The conservative case is the one that keeps you solvent; it tells you the buffer to hold.

How this differs from cash flow analysis

Worth being clear, because the terms get muddled. Cash flow analysis looks backwards — it reads what already happened to judge the health of the business. Forecasting looks forwards, using those same actuals to project what's coming. If you want the backward-looking side in depth, the business cash flow analysis guide covers reading net position and seasonality. For forecasts a lender will scrutinise, cash flow analysis for lending goes into what underwriters look for. And before you forecast anything, the bank statement health check walks through confirming the actuals are sound.

A note on what this tool does and doesn't do

Export Bank Statement converts and verifies your statements, and its analyser surfaces the recurring, income and category patterns you forecast from. It does not plug into your bank as a live feed, and it doesn't push data into Xero or QuickBooks through an API — that needs partner certification this tool doesn't have. The honest path is convert → import the CSV: you export clean, reconciled actuals, drop them into your spreadsheet or accounting software's bank-import format, and build the forecast yourself. It's a self-serve tool that gives you trustworthy raw material, not an autopilot.