Yes, for the errors that come from moving figures by hand, and no for the errors that come from judging when cash will move. Connecting Xero or QuickBooks Online to a cash forecast removes re-keying, keeps the opening balance close to the books, and brings every bill the ledger holds into view. It cannot tell you when a customer will actually pay, and it cannot see a commitment nobody has recorded yet. A connected forecast is more accurate where the old errors were mechanical; where they were about timing, the accuracy still depends on the expected dates the finance team sets.
This page takes the four places error enters a manual cash forecast and says, for each one, what a ledger connection does to it and what the team still has to do. It does not cover how to measure forecast error, which is in how accurate 13-week cash flow forecasts are, or the set-up of a connected forecast, which is in setting up automated cash flow forecasting with Xero.
What "more accurate" means for a cash forecast
A cash forecast is accurate when the balance it predicted for a week or a month is close to the balance the bank showed when that week or month closed. The useful version of the question is which part of the gap a connection closes. An article in The Treasurer, the Association of Corporate Treasurers' journal, puts the working view plainly: "Actual events will rarely, if ever, match the forecast. Understand variances and use this to improve the next forecast."
No professional body publishes an accuracy standard for a small or mid-sized business's cash forecast, and no independent study has measured how much more accurate a forecast becomes when a business of this size connects its ledger to it. Accuracy figures that circulate on the question come from software companies describing their own products or from studies of large enterprises. So the answer on this page is an account of mechanism rather than a percentage: which errors the connection removes, which it reduces, and which it leaves where they were.
The distinction that runs through all four is this. A connection fixes how data is transferred. What it delivers depends on how complete and current the ledger is. And it leaves the timing judgement where it was.
The four places error enters a manual forecast
| Where the error comes from | What a ledger connection does | Effect on the error | What the team still does |
|---|---|---|---|
| Re-keying figures from the ledger into the forecast | Imports balances, invoices, bills and reconciled transactions, so nothing is typed | Removed, for data the ledger holds | Enters estimates for cash the ledger does not hold |
| A stale opening balance | Starts from the reconciled balance in the books, refreshed daily | Reduced to the age of the last sync and the last reconciliation | Reconciles often and syncs before the forecast is read |
| Due dates treated as pay dates | Brings in each invoice and bill with its due date, and an expected date where one is set | Not fixed by the sync alone | Sets expected dates on the receipts and payments that matter |
| Bills and commitments missing from the forecast | Brings in every bill recorded in the ledger | Reduced, once the bill is recorded | Budgets for what is agreed but not yet in the books |
Re-keying: removed, for the data the ledger holds
In a spreadsheet forecast, someone exports the aged debtors and creditors lists, the bank lines and the balances, and pastes or types them into the workbook each week. Every transfer is a chance to drop a row, paste into the wrong week or type a figure wrong. ICAEW's principles for good spreadsheet practice ask that each input be entered only once, and a connection meets that by design: the figure is entered once, in the books, and the forecast reads it from there.
This is the clearest gain, and it should not be overstated. Research on business spreadsheets shows that errors are common, but it does not show that typing errors in the transfer step are a large share of what goes wrong in a cash forecast, and no study isolates them. Removing re-keying removes a class of error and the hours spent on it. The larger accuracy questions sit in the next three sections. What a connection takes over step by step, platform by platform, is in our guide to how much manual data entry forecasting software removes.
The opening balance: reduced to the age of the last reconciliation
Every forecast is built forwards from an opening cash position, so an opening balance that is wrong moves every week after it by the same amount. In a spreadsheet the balance is keyed in when someone remembers to update it, and it can sit unchanged for days.
A connected forecast opens on the balance the accounting platform holds. In Float that is the reconciled balance of the bank accounts and cards you include, the "Balance in Xero" or "In QuickBooks" figure, refreshed by a daily import at an hour you choose and by a manual sync whenever you need one. That shrinks the staleness to the time since the last sync. It does not close it, because the reconciled balance is not the bank balance. A transaction that has cleared the bank but has not been reconciled in the books is not in the forecast's opening position until it is reconciled. As the same Treasurer article puts it, "Most people don't understand that an accounting cash balance (book) is a million miles from real cash (bank)."
So the connection reduces this error to the state of the reconciliation. Float's guidance is to reconcile daily, or at least weekly, and to run a sync after a reconciliation and before the forecast is read in a cash meeting. The difference between the books and the bank is explained further in cash flow forecasting versus accounting reports.
Due dates as pay dates: the error the sync cannot touch
A due date is what the invoice says. A pay date is what the customer does. A forecast that places every receipt on its due date is right only when customers pay on time, and the published figures show that a material share do not. In the UK, large businesses reporting on their own payment practices for 2025 said 15% of the invoices they paid were paid late (14% by value), and that they paid their suppliers in 32 days; the figures are self-reported (Department for Business and Trade, July 2026). In Australia, large businesses paid 68.5 per cent of their invoices to small suppliers on time between July and December 2025, with an average payment time of 27.2 days against an average term of 29 days (Payment Times Reporting Regulator, August 2026). No official equivalent is published for the US or New Zealand. Those are figures for large payers; your own customers' history is the figure that matters for your forecast.
The guidance is consistent about what to do instead. ICAEW's principles for cash flow forecasting ask for expected, worst-case and best-case dates for receipts and payments. The British Business Bank's guidance is to "put the figures in for when you know clients will pay invoices, or bank payments will clear."
A connection does not supply that date by itself, because the accounting platform usually does not hold one. In Xero, an expected payment date is an optional field on sales invoices and a planned payment date is an optional field on bills, filled only when someone sets it. In QuickBooks Online, invoices and bills carry a due date and no expected-date field. Float uses the expected date from the accounting platform where one has been set, and the due date where it has not. So for most invoices the connected forecast starts on the due date, exactly as the spreadsheet did.
When receipts land late, a forecast built on due dates shows cash arriving earlier than it does, and the low week looks better than it will be. That is the error a sync alone cannot touch, and it is often the one that decides whether the forecast is useful.
What the team does about it depends on the platform:
- On Xero, Float's Smart Expected Dates applies each customer's and supplier's average days late, worked out from three to twelve months of your own payment history, to new invoices and bills as they import. The date is applied once, is not recalculated later, and can be overridden.
- On QuickBooks Online, that feature is not available. The team sets expected dates in Float, one invoice at a time or in a batch, starting with the largest receipts and payments in the weeks where the balance is lowest.
On either platform, two behaviours affect accuracy directly. An overdue invoice or bill with no new expected date is assumed to be paid today and is marked in the app, so the overdue list needs clearing before the forecast is read. And a change to a due or expected date made in the accounting platform overwrites the date set in Float at the next import.
Missing bills: reduced, once the bill is in the books
A spreadsheet forecast misses an outflow when nobody adds it: a supplier invoice still in someone's inbox, a quarterly bill that fell off the list. A connection closes part of that gap, because every bill recorded in the ledger arrives in the forecast with its date and amount. For Xero users, draft invoices and bills can be included with a setting switched on.
The part it cannot close is a commitment that is not yet in the books. Government and professional-body guidance on cash flow forecasting lists outgoings such as new hires, loan repayments, asset purchases and tax bills, alongside the irregular or one-off large payments that are easiest to forget. Many of those are known to the business weeks before any bill exists. A signed contract, an agreed hire or a planned equipment purchase is a cash outflow the ledger has never seen.
The distinction that matters is between a bill in the ledger and a commitment that has not been recorded. The connection brings in the first automatically. The second has to go into the forecast as an estimate, which in Float is a budget: one-off or repeating, set at a fixed amount, or tracking recent history. A budget is a placeholder that fills as the invoice, bill or reconciled transaction arrives, so the estimate gives way to the actual figure without being re-entered. For Xero users in the UK, Australia and New Zealand, VAT or GST is forecast automatically; on QuickBooks Online the tax payment goes in as a budget on its due date.
What still needs judgement after you connect
A connected forecast changes where the finance team spends its time. Less of it goes on moving numbers and more on five judgements the ledger cannot make:
- Expected pay dates on the largest receivables. The handful of receipts that decide the lowest week are worth checking with whoever speaks to the customer, and they carry most of the weight in a 13-week cash flow forecast.
- Commitments not yet in the ledger. Agreed work not yet invoiced, a hire, a loan drawdown or an asset purchase goes in as an estimate until the bill exists. The 13-week cash flow forecast template shows the lines to cover.
- Payroll and tax timing. How payroll reaches the forecast depends on how it is posted in the books, and payroll posted by journal needs journal import switched on. The platform-by-platform position is in how much manual data entry forecasting software removes.
- Reconciliation cadence. The forecast is as current as the last reconciliation. What a Xero connection can and cannot tell the forecast is set out in how Xero data feeds a cash flow forecast.
- Retiring estimates that have gone stale. A budget for a contract that slipped or a purchase that was cancelled stays in the forecast until someone removes it. Comparing forecast with outturn each week is how stale estimates get found; how accurate 13-week cash flow forecasts are sets out that check.
None of these is new work created by the software. They were always the part of the forecast that decided its accuracy. The connection takes away the work that was hiding them.
How Float fits
Float is a cash flow forecasting tool for finance teams that keep their books in Xero or QuickBooks Online. Sage Intacct support is coming soon; you can join the waitlist. Our guide to Float's accounting integrations covers each connection.
The connection is one-way. Float imports from the accounting platform once a day at an hour you choose, with a manual sync on demand, and writes nothing back to the books. From Xero it imports bank accounts, credit cards, invoices and bills including those awaiting approval, repeating invoice and bill templates, draft invoices if you switch that on, transactions and journals; from QuickBooks Online it imports bank accounts, credit cards, invoices and bills, transactions and journals. Neither import includes purchase orders, quotes or estimates, tracking categories or classes. The forecast opens on the reconciled balance, and only reconciled transactions enter it.
On top of the import, every invoice and bill carries an expected date that the team can change singly or in a batch, Smart Expected Dates applies payment history to new invoices and bills for Xero users, budgets hold the cash the ledger has not seen, and a threshold shows on the side panel the date the balance is due to cross a figure you set. The forecast exports to PDF or CSV when someone asks for it. Float does not send alerts or scheduled reports, and it does not connect to the bank, so the opening position is the books' reconciled figure and not the bank's live balance.
What Float does not do is decide the expected dates for you on QuickBooks Online, or know about a commitment before it is recorded. Those stay with the team, on any tool. For the reserve the forecast is protecting, see how many weeks of cash cover to hold.
Frequently asked questions
Does connecting Xero or QuickBooks Online make a cash flow forecast more accurate?
It makes the forecast more accurate where errors come from handling data: re-keying is removed, the opening balance stays close to the reconciled books, and every bill recorded in the ledger appears. It does not correct payment timing, because most invoices arrive with a due date and no expected date, and it cannot see commitments that are not yet recorded. Accuracy on those depends on the expected dates and estimates the finance team sets.
Which forecasting errors does an accounting integration remove?
It removes errors from re-keying data the ledger already holds, such as balances, invoices, bills and reconciled transactions, because the forecast reads them directly instead of having them typed or pasted in. It reduces two others: a stale opening balance, which shrinks to the age of the last sync and reconciliation, and missed bills, once they are recorded in the books. It does not remove the error of treating due dates as pay dates.
Why is my cash forecast still wrong after connecting my accounting software?
The most common causes are receipts placed on their due dates when customers pay later, unreconciled transactions that are not yet in the opening balance, overdue invoices assumed to be paid today, and commitments that are agreed but not yet recorded in the ledger. Each of these sits outside what a sync can fix. Setting expected dates on the largest receipts, reconciling before the forecast is read, clearing the overdue list and budgeting for known commitments addresses most of the gap.
How current is the data in a connected cash forecast?
In Float, data is imported from Xero or QuickBooks Online once a day at an hour you choose, and a manual sync pulls it in on demand. Only reconciled transactions are imported, and the opening balance is the reconciled balance in the books, not the bank's own balance. So the forecast is as current as the later of the last sync and the last reconciliation, which is why Float recommends reconciling daily or at least weekly.
Does a connected forecast use due dates or expected payment dates?
Float uses the expected date from the accounting platform where one has been set, and the due date where it has not. In Xero, an expected payment date can be set on sales invoices and a planned payment date on bills, but both are optional. In QuickBooks Online, invoices and bills carry a due date only. For Xero users, Smart Expected Dates applies each customer's and supplier's payment history to new invoices and bills as they import.
What should a QuickBooks Online user do about expected payment dates?
Set them in Float. QuickBooks Online invoices and bills carry a due date and no expected-date field, and Smart Expected Dates is not available on QuickBooks Online, so a connected forecast places them on their due dates until someone changes them. Expected dates can be changed one invoice at a time, moved by seven or thirty days, or set in a batch. Start with the largest receipts and payments in the weeks where the balance is lowest.
Is there a benchmark for cash forecast accuracy in a small business?
No professional body publishes one, and no independent study has measured the accuracy gain from connecting accounting software to a forecast for a small or mid-sized business. Published accuracy percentages come from software companies describing their own products or from studies of large enterprises. The practical measure is your own: compare each week's forecast with what happened, and use the variances to improve the next forecast.
If I change an expected date in Float, does it change in Xero or QuickBooks Online?
No. The connection is one-way, so changes made in Float stay in Float and nothing is written back to the books. It works the other way round: if someone changes a due date or expected date in Xero or QuickBooks Online, that change flows into Float at the next import and overwrites the date set in Float. Comments on an invoice or bill in Float can record why a date was changed.
The quickest way to see which of the four errors sits in your own forecast is to connect your ledger and compare. Start a 14-day free trial of Float, import your Xero or QuickBooks Online data, and look first at the expected dates on the receipts that carry your lowest week.
