Forecasting Cash Receipts from AR

Executive summary

Forecasting cash receipts from AR means turning the open invoices in your receivables ledger, plus the invoices you have not yet issued, into a dated weekly estimate of when cash will reach the bank. It is usually the largest and least reliable line in a 13-week cash forecast, because customers do not pay on due dates. They pay on their own schedules, and those schedules differ by customer, by invoice size, and by month.

Three methods build the forecast, and the best forecasts combine them.

1. Due-date method. Every open invoice is assumed paid on its due date. It is quick, and it is almost always too optimistic.
2. Payment-pattern method. Each invoice is spread across weeks using the observed lag between due date and payment, by customer or customer group. This is the workhorse.
3. Promise method. For the largest or most overdue invoices, a named collector records a promised date, and the forecast uses that date with a haircut based on how often the customer keeps promises.

The forecast is only as useful as the feedback loop around it. Each week you compare forecast receipts with actual receipts by customer, find out why they differ, and adjust the pattern. A receipts forecast that is never compared with actuals is a guess with a spreadsheet around it.

When to use this

Use this guide when:

• The receipts line of your 13-week forecast misses by a wide margin, in either direction, most weeks.
• A lender, board, or sponsor relies on the forecast and has asked how receipts are estimated.
• DSO is stable on paper but cash arrives in lumps you did not expect.
• You have a few customers that make up most of AR and each can move the weekly total.
• You are setting up a 13-week cash flow forecast for the first time.

Do not start here if the AR subledger does not reconcile to the general ledger or if unapplied cash is large. Fix those first. A forecast built on an unreliable ledger will be wrong in ways you cannot trace.

Diagnostic questions

Answer these before you build the model. Each tells you which method to lean on.

1. Does the AR subledger reconcile to the general ledger at the last month-end?
2. How much cash is received but unapplied, and how old is it?
3. What share of AR sits in the top 10 customers?
4. For each of those customers, what are the contractual terms and the actual average days from due date to payment over the last 12 months?
5. Is the payment lag stable by customer, or does it swing with their month-end, quarter-end, or approval cycle?
6. How are customers paying (ACH, wire, check, card, portal), and does the method add a predictable delay?
7. What share of invoices are disputed or short-paid, and how long do disputes take to clear?
8. Do collectors record a promised date for each large past-due invoice, and is it kept in a place the forecast can read?
9. What is the lag between shipment or service completion and the invoice date, and is it consistent?
10. What share of next quarter's receipts will come from invoices not yet issued, such as new orders, milestones, or recurring billing?
11. Are there customers with deductions, rebates, or offsets that reduce cash below invoice value?
12. When you last compared forecast with actual receipts, which customers explained most of the difference?

Rule of thumb for reading the answers: if questions 3 to 5 show a few large customers with stable patterns, build the forecast customer by customer for those accounts and group the rest. If question 7 is high, treat disputed invoices as a separate line with its own timing, not as part of the normal pattern.

The framework: three layers of receipts

Each layer has a different level of certainty, and the forecast should show them separately so you can see how much of next week's receipts depends on invoices that already exist and how much on invoices still to be issued.

Step-by-step process

1. Reconcile and clean the ledger. Tie the AR subledger to the general ledger, apply unapplied cash, and post credit memos. Owner: Controller. Output: a clean open-invoice file with invoice number, customer, invoice date, due date, and open amount.
2. Compute each customer's payment lag. For invoices paid in the last 12 months, compute days from due date to payment date by customer. Record the average and the spread (for example, the share paid within a week of due date, one to two weeks late, and so on). Owner: FP&A. Output: a payment-pattern table by customer or customer group.
3. Group the long tail. For small customers, group by terms, size band, or payment method, and use one pattern per group. Owner: FP&A. Output: a short list of groups, each with its pattern.
4. Build the lag distribution. Express each pattern as shares of an invoice collected in the due week, one week later, two weeks later, and so on. The shares must add up to 100% including a share that falls beyond the forecast horizon. Owner: FP&A. Output: a lag table where each row sums to 100%.
5. Apply the pattern to open invoices. Multiply each invoice's open amount by the lag shares, starting from its due week. Owner: FP&A. Output: weekly receipts for current invoices.
6. Treat past-due items separately. For each large past-due invoice, collectors record a promised date and a confidence level. Use the promised date with a haircut equal to the share of that customer's past promises that were broken. For small past-due items, use an aged pattern, which is usually slower than the current-invoice pattern. Owner: Collections lead. Output: weekly receipts for past-due invoices.
7. Carry disputes on their own line. Show disputed amounts separately with the dispute owner's expected resolution date. Do not include them in normal patterns. Owner: Operations and Sales. Output: a dated dispute line.
8. Forecast future invoices. Take the billing forecast (orders shipped, milestones reached, recurring invoices), apply invoice lag, add terms, and apply the pattern. Owner: FP&A with Sales. Output: receipts from invoices not yet issued.
9. Adjust for deductions and offsets. Where customers routinely short-pay for rebates, freight, or returns, forecast net cash, not gross invoice value. Owner: Controller. Output: a deductions line.
10. Sum the layers and compare to history. Check the total against last year's same-week receipts and recent weekly averages, and explain any large difference. Owner: FP&A. Output: a weekly receipts forecast with the layers shown separately.
11. Run the weekly feedback loop. Each Monday, compare last week's forecast with actual receipts by customer, record the reasons for the largest differences, and update patterns and promise haircuts. Owner: FP&A and Collections lead. Output: a variance log and updated parameters.

Templates

Payment-pattern table (copy into a spreadsheet)

Each row must sum to 100%. Fill the shares from your own payment history, not from a blog.

Past-due promise log

Weekly receipts variance log

Worked example

Illustrative. The numbers are made up to show the method and are not benchmarks.

A distributor has open invoices not yet due, grouped by due week:

From its own history, the company finds that invoices are paid with this lag from the due week: 30% in the due week, 25% one week later, 20% two weeks later, 15% three weeks later, and 10% later than that. Check: 30 + 25 + 20 + 15 + 10 = 100%.

It also has $0.6M of past-due invoices. The collectors' promised dates, after haircutting for broken promises, give 50% collected over weeks 1 to 4 (even at $0.075M per week), and the rest falls beyond the horizon.

Applying the pattern to each due week and adding the past-due line gives weekly receipts:

Check: week 4 current receipts are 0.30 × 1.1 (due week 4) + 0.25 × 0.9 (due week 3 plus one) + 0.20 × 1.2 (due week 2 plus two) + 0.15 × 1.0 (due week 1 plus three) = 0.330 + 0.225 + 0.240 + 0.150 = $0.945M. The 10% tail, 0.10 × $4.2M = $0.42M, falls beyond the horizon, so $3.78M + $0.42M = $4.2M and nothing is lost. The past-due line is 50% of $0.6M = $0.30M, or 4 × $0.075M.

What the operator learns:

• A due-date method would have placed all $1.0M in week 1 and $1.2M in week 2. The pattern method shows only $0.375M and $0.685M. A plan built on the due-date method would be short in the first two weeks.
• Of the $4.2M, $0.42M is not expected within the first seven weeks. That is the tail that collectors should be working.
• The peak week is week 4, at $1.02M, not week 2 when the largest invoice group falls due.

Metrics and KPIs

Setting targets without invented benchmarks. Use your own history. Take the last 8 to 13 weeks of forecast error as the baseline, and target a narrower error and a bias close to zero. Decide an acceptable error band with the CFO and lender based on what the minimum liquidity buffer can absorb, not on a number from another company. For customer patterns, use the customer's own last 12 months.

Common mistakes

1. Using due dates as payment dates. Customers rarely pay on the due date. The pattern method fixes this.
2. One average for all customers. A few large customers with their own pattern will swing the line. Model them individually.
3. Believing promises at face value. A promised date is the customer's intent. Haircut it by the customer's history.
4. Forgetting future invoices. Receipts in weeks 5 to 13 depend heavily on billing not yet issued. Forecast billing, not only the current ledger.
5. Ignoring deductions. Short-pays and rebate offsets mean net cash is less than invoice value.
6. Mixing disputed items into normal patterns. They have different timing and different owners.
7. Not measuring bias. Errors that always run in the same direction are a model problem, not bad luck.
8. Updating the forecast without recording why. If the reason for each difference is not recorded, the same error will return.
9. Counting early-pay discount uptake as cash without costing it. Discounts accelerate cash and reduce it. Price them before using them in the forecast.

AI prompt assistance

AI can help calculate patterns and summarize variances. It does not know your customers. Provide the data and check every figure against the ledger and the bank.

1. Compute payment patterns.

I will paste a list of paid invoices with customer, due date, and payment date for the last 12 months. For each of the top 15 customers by volume, compute average days from due date to payment and the share paid in the due week, one week later, two weeks later, three weeks later, and later than that. Show the row sums and flag any customer with fewer than 10 invoices.

Check: confirm each row sums to 100% and recompute one customer by hand.

2. Apply a pattern to an open-invoice file.

I will paste open invoices by customer and due date plus a payment-pattern table. Produce weekly receipts for the next 13 weeks by multiplying each invoice by the pattern shares starting from its due week. Show the weekly total and the amount falling beyond week 13, and confirm that the weekly totals plus the beyond-horizon amount equal total open AR.

Check: the reconciliation to total open AR is the test. If it does not tie, do not use the output.

3. Explain a forecast miss.

Here are forecast and actual receipts by customer for last week. List the five customers that explain the largest differences, classify each as timing, dispute, short-pay, or forecast error, and propose one parameter change for each.

Check: confirm each classification with the collector who owns the account.

Implementation guide

Governance: the CFO owns forecast accuracy as a reported metric. The Controller owns the ledger. The collections lead owns promised dates and the promise-keep rate. Sales owns the billing forecast. Report weekly forecast error and bias in the cash meeting alongside the 13-week low point.

What to expect: the first version will miss. The value comes from the second to sixth weeks, as the variance log shows which customers and causes drive most of the error. Expect the long tail to be harder to forecast than the top customers.

Downloadable assets

The payment-pattern table, promise log, and variance log above are designed to be copied into a spreadsheet. For the full forecast structure, use the 13-week cash flow builder.

Related resources

• 13-week cash flow forecast template: where the receipts line sits in the whole forecast.
• AR aging analysis and collections prioritization: which invoices to work first and why.
• How to reduce DSO: moving the payment patterns themselves.
• Cash flow operating system playbook: the weekly rhythm around the forecast.
• Cash management guides: the full set of cash and liquidity guides.

Frequently asked questions

How do you forecast cash receipts from accounts receivable?

You assign each open invoice to its due week, then spread it across the following weeks using the share of invoices that customers historically pay in each week after the due date. Add promised dates for large past-due items, with a haircut for broken promises, and add receipts from invoices not yet issued.

Why do cash receipts forecasts miss?

Receipts forecasts usually miss because they assume invoices are paid on the due date. Customers pay later, at different speeds, and disputes and short-pays hold cash back. Large customers also move the weekly total. Using each customer's own payment pattern and recording the reason for every miss reduces the error over time.

What is a payment-pattern method?

A payment-pattern method uses historical data on how many days after the due date customers actually pay, expressed as the share of an invoice collected in each week. You apply those shares to every open invoice. It is usually more accurate than the due-date method, and each customer or group's shares must sum to 100%.

How should you treat disputed invoices in a cash forecast?

Treat disputed invoices as their own line with a date supplied by the person who owns the dispute, not as part of the normal payment pattern. A collections call does not resolve a pricing or quality disagreement. Show the amount, the owner, and the expected resolution date so the forecast stays honest.

How often should you update the receipts forecast?

Update the receipts forecast weekly, with actual receipts from the bank compared to forecast by customer. Roll the horizon forward, revise the promise log, and record the reasons for the largest differences. Parameters such as customer lag and promise-keep rate should be reviewed at least monthly.

How this works in a System of Performance

A receipts forecast improves only when forecast, promise, and actual cash stay linked. In most companies the invoice file lives in the ERP, the promised dates in a collector's notes, and the bank actuals in a separate portal. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. If you want to see how your receipts forecast compares with what actually arrives, bring it to us as a launch partner.