An hourly liquidity forecast for payment operations

A positive closing balance cannot fund a payout due at 11 a.m. Payment liquidity depends on when money becomes usable in the required account, not just how much will eventually arrive. A working forecast exposes the lowest intraday balance and the last moment at which an authorised response can still help.

A profitable-looking day with a funding gap

Consider a hypothetical account holding 80,000 currency units at 9 a.m. A 70,000 payout is due at 11 a.m. Refunds of 10,000 and fees of 2,000 are expected at 1 p.m. A deposit of 50,000 should arrive at 3 p.m., followed by another 30,000 payout at 5 p.m. The successive balances are 80,000, 10,000, −2,000, 48,000, and 18,000. The positive closing figure hides a 2,000 intraday shortfall.

Suppose the company has chosen an internal minimum buffer of 15,000. It needs another 17,000 before the early payments to keep the lowest projected balance at that floor. Now delay the 50,000 receipt until the next day. The low point becomes −32,000, and maintaining the same floor requires 47,000 of additional funding. These figures illustrate a method; they are not a recommended reserve level for every operator.

Decide which funds really count

Forecast separately by currency, account, and legal entity. For each balance, record its availability, restrictions, and the time needed to move it. Money held in another currency is useful only if an authorized conversion and transfer can complete before the deadline. Customer funds should be used solely within their permitted purpose; they are not a general working-capital buffer.

Check the provider’s balance definitions. Adyen, for example, distinguishes current, available, pending, and reserved balances. Availability reflects relevant restrictions. These categories are not independent pots that can simply be added together. Source: Adyen, Balance types.

Every forecast line should have an amount, expected time, source, confidence level, and owner. Separate deposits already initiated, expected settlement of completed payments, and projected new sales. New-sales forecasts are particularly vulnerable to changes in demand. A conservative view should not let hypothetical receipts conceal commitments that are already due.

Do not turn probability into spendable cash

When a 50,000 receipt is highly likely, it is tempting to forecast a slightly smaller “conservative” amount. That arithmetic can conceal the actual risk. An illustrative 90% probability does not create a guaranteed 45,000: the full receipt may arrive before the deadline, or none of it may arrive. For a binding payout, separate timing scenarios are more useful.

Do not blend restrictions across accounts either. Suppose 30,000 is free in one currency while another currency has a 5,000 shortfall. An aggregate converted balance does not answer whether the required amount can legally and operationally move before the deadline. Record the permitted route, conversion and arrival times, cost and person authorised to initiate it.

If funding still needs approval, show it as a conditional line. A possibility written in a spreadsheet is not an available facility. This distinction matters at shift changes: the next operator must be able to tell a confirmed top-up from an option discussed earlier that morning.

Give each funding option a decision deadline

A top-up that takes four hours cannot fix a gap discovered ten minutes before payout. Document the last useful decision time for each permitted response. Options might include prearranged funding or adjusting a payout schedule that has not yet been promised, where contracts allow it. Existing obligations should not be silently postponed to make the forecast look better.

Identify the purpose of each reserve. Adyen distinguishes a merchant-configured reserve for refunds and other expenses from a risk-related deposit whose amount the provider determines. Dashboard visibility alone does not establish that funds can be used for any payout. Source: Adyen, Balances.

Put the decision time ahead of the payout time

In this example the balance becomes negative at 1 p.m., but the chosen 15,000 buffer is breached earlier: the 11 a.m. payout leaves only 10,000. If the base-case 17,000 top-up arrives as one amount, it must be available by 11 a.m. With four hours for transfer and one hour for approval, the process must start no later than 6 a.m. Counting back from 1 p.m. would give 8 a.m., protecting only against a negative balance while leaving the buffer breached between 11 a.m. and 1 p.m. Prepare this forecast in advance. If the team or funding channel does not operate at 6 a.m., arrange the money in the preceding available window. These timings are hypothetical: verify operating hours, transfer duration and the margin needed for delays.

For each response, record four things: amount, source, lead time and required authority. Do not add together alternatives that draw on the same money. Two routes for moving the same 20,000 do not create 40,000 of liquidity.

After a day with a forecast miss, separate amount error from timing error. A receipt arriving four hours late may matter more than a small revenue shortfall. Improve the assumptions that change the funding decision before its deadline. The forecast should give the responsible person an executable action while there is still time, rather than merely colour yesterday’s gap red.

Use three views, with explicit assumptions

• Base case: known obligations and realistic receipt times.
• Delayed receipt: a material incoming payment misses the next payout deadline.
• Higher outflow: refunds or withdrawals exceed the usual level under a stated assumption.

Test combinations as well. Delayed settlement and heavier withdrawals can happen together. Avoid stacking arbitrary worst cases without explanation, but do not assume independent events automatically. A large customer may justify a separate scenario based on the amount they are entitled to withdraw at once.

Payout frequency and settlement speed are separate variables. Stripe states that changing a payout schedule does not accelerate the movement of pending funds into the available balance. Include both timings, plus the bank-arrival expectation, in the forecast. Source: Stripe, Payout schedules and timing.

Review forecast quality, not just today’s number

Compare expected and actual amounts and timestamps each day. Explain errors in incoming and outgoing flows separately. Track the lowest intraday balance, coverage of the next obligations, and time remaining before the last useful funding decision. Revisit delay assumptions on a defined schedule rather than only after an incident.

For the next operating day, record three figures: minimum intraday balance, required top-up and the last useful decision time. Give each a source and an owner. If a timely decision is already impossible, the forecast has exposed a funding-process problem to fix.