How to Forecast Cash Collections From Accounts Receivable: A Method for AR Teams

Most finance teams already have a cash forecast. The weak line in it is almost always the same one: cash coming in from customers. Payables are dated by your own payment runs, and payroll is fixed. Receipts depend on dozens or hundreds of customers deciding when to pay you.

That gap is large. Atradius reports that 43% of credit-based B2B sales in the US are overdue, mostly because of customer cash flow pressure (Atradius Payment Practices Barometer, US 2025). Yet the default forecast in many accounting systems still dates receivables by their due date. Microsoft's own documentation for Business Central says its cash flow forecast reads Customer Ledger Entries by Due Date and amount (Microsoft Learn, Business Central cash flow setup). If close to half your invoices pay late, a forecast built on due dates is optimistic by design.

This guide shows a controller or AR manager how to forecast cash collections the way customers actually pay: by account history, by aging bucket, and by the promises your collectors already record.

What Is a Cash Collections Forecast?

A cash collections forecast predicts how much cash customers will pay you in each coming week, based on how they actually pay rather than when invoices are due. It covers the open receivables in your aging report plus invoices you expect to send in the forecast window.

It answers three questions for the controller:

• How much cash will land in each of the next 4 to 13 weeks?

• Which accounts carry the biggest gap between due date and expected payment?

• What can the collections team do this week to close that gap?

A good collections forecast is not a separate spreadsheet project. It is a weekly read of the same data your AR team already works from: the aging report, payment history, promises to pay, and disputes.

Why Do Due-Date Forecasts Miss?

Customers do not pay on the due date

Some pay early to take a discount, many pay a few days late out of habit, and a minority pay very late. A due-date forecast treats all of them as on time. The result is a forecast that is right in total over a long period and wrong in almost every individual week.

Disputes, deductions and short-pays hold cash back

An invoice with an open dispute will not pay on its due date no matter what the system says. The same goes for a deduction or a short-pay waiting for review. If those balances sit in the forecast at full value, the forecast overstates cash. See our guide to customer deductions and short payments for how to track them.

Unapplied cash and credits distort the starting point

Payments that have arrived but are not applied to invoices make the aging report look worse than it is, and the forecast then predicts cash you already have. Clean up unapplied cash before you forecast, or you will double count.

A few large accounts swing the whole number

When one or two customers hold a big share of receivables, their payment timing decides the week. A forecast that averages them in with small accounts hides the real risk. Our guide to receivables concentration risk covers how to spot this.

Which Forecasting Method Should an AR Team Use?

There are three practical methods. Most teams with 3 to 5 collectors end up using all three together: a historical method for the bulk of small accounts, and account-level judgment for the large ones.

Aging-bucket collection rates

Pull the last 6 to 12 months of aging snapshots and payments. For each bucket (current, 1 to 30, 31 to 60, 61 to 90, over 90), measure what share of the opening balance was collected within the next 30 days. Those are your collection rates. They are close cousins of the bucket roll rates covered in our AR aging report analysis guide.

Bucket collection rates are the fastest way to get a realistic baseline, because they are built from your own customers' behavior, not from terms on paper.

Account-level days-to-pay

For your larger accounts, calculate the average number of days past due each customer actually pays, weighted by invoice amount. Then push every open invoice for that account forward by that number. A customer with net 30 terms who reliably pays on day 41 should be forecast on day 41.

Promise-based overrides

When a collector gets a firm commitment, that date and amount replace the model for that invoice. This is where the collections team and the forecast meet. If your team logs promises consistently, as described in our guide to promise to pay tracking, the forecast for the next two weeks becomes much sharper.

How Do You Build a 13-Week Collections Forecast?

Step 1: Start from a clean aging report

Apply cash, clear credits where you can, and tie the aging report to the general ledger. Tag every invoice that has an open dispute, deduction or short-pay. These tags matter more than any formula.

Step 2: Split the book into tiers

• Tier 1: the largest accounts, usually the 10 to 20 customers that make up most of the balance. Forecast these invoice by invoice.

• Tier 2: accounts in active follow-up with a recorded promise. Use the promise.

• Tier 3: everyone else. Use aging-bucket collection rates.

Step 3: Forecast Tier 1 by account history

Use each account's days-to-pay and adjust for anything you know: a customer moving to a new AP system, a payment run schedule, or a slow month at year end. Put disputed invoices in a separate line with no date until the dispute has an owner and a resolution date.

Step 4: Apply promises to Tier 2

Enter the promised date and amount. Then apply a reliability discount based on history: if this customer has kept 70% of past promises on time, do not forecast 100%. Broken promises go back into the Tier 3 model and back onto a collector's list.

Step 5: Apply bucket rates to Tier 3

Multiply each bucket balance by its collection rate and spread the result across the weeks in the month based on when your customers usually pay (many pay heavily at the end or start of a month).

Step 6: Add expected new invoices

Add invoices you will send in the window, using the same rates for the current bucket. Keep this line separate so you can see how much of the forecast depends on billing on time.

Step 7: Compare to actuals every week

Each Monday, put last week's forecast next to last week's actual receipts. Record the variance by tier. This is the step most teams skip, and it is the one that makes the forecast better over time.

What Does a Worked Example Look Like?

Here is a simplified one-month view for a distributor with $2.4M in open receivables. The collection rates come from the company's own last 12 months.

These figures are illustrative, not a benchmark. The point is the shape: a due-date view says every past-due dollar arrives now and every current dollar arrives on time. The history-based view shows about half the cash the due-date view suggests for the month, and it shows where the gap is. In this example, the 1 to 30 and 31 to 60 buckets hold most of the recoverable difference, which is exactly where a week of focused follow-up pays off.

How Does the Collections Team Change the Forecast?

Forecasting is not only measurement

A forecast tells you what will happen if the team keeps doing what it does today. The collections team can change that. Every account that gets a timely reminder, a call and a recorded promise moves from the Tier 3 average to a dated Tier 2 commitment.

Coverage is the biggest lever

On a team of 3 to 5 people, the large accounts get attention and the long tail does not. Small and mid-size balances drift into older buckets, where collection rates drop. Covering every account every week raises the collection rate of the younger buckets, which is the cheapest cash in the book. The guide to structuring a collections team of 3 to 5 people shows how teams split that work.

Prioritize by forecast gap, not only by age

Rank accounts by the difference between due date and expected payment, weighted by balance. A $90,000 invoice expected 25 days late matters more to this month's forecast than a $3,000 invoice 70 days past due. Our guide to prioritizing collections accounts covers a simple scoring model.

How Do You Know the Forecast Is Working?

Track three things every week:

• Forecast accuracy by week: actual receipts divided by forecast receipts. Aim for a steady band, not a perfect number.

• Promise kept rate: the share of promises paid in full on the promised date. A falling rate means Tier 2 is overstated.

• Collection effectiveness: the Collection Effectiveness Index shows how much of what was collectible was actually collected in the period. The formula is explained in detail by AccountingTools.

Over a few months, these numbers should move together with DSO. If you need a refresher on the DSO math, see how to calculate DSO and the 3 numbers that move it, and for the levers that improve it, our hub on lowering DSO without hiring.

Where Does an AI Collections Agent Help the Forecast?

The forecast is only as good as the follow-up data behind it. Abivo's AI collections agent, Kate, works from your accounting system and adds the parts that usually go missing on a busy AR team:

• Every overdue account gets followed up by phone, text and email, not only the largest ones, so the long tail does not drift into older buckets.

• Promises to pay are recorded with date and amount, which is exactly the Tier 2 input the forecast needs.

• Disputes are logged and handed to your team, so disputed balances come out of the forecast and get an owner.

• One message can cover all open invoices on an account, with details such as PO, job or billing period, which makes it easier for the customer's AP team to schedule payment.

OFS Group went from four collectors to one with Abivo, recovered $842,518 in four months and cut DSO by more than 30 days; the full story is in the OFS Group case study.

Practical Takeaways for Controllers and AR Managers

• Stop forecasting receivables by due date unless the customer has a proven on-time record.

• Clean the aging report first: apply cash, tag disputes, deductions and short-pays.

• Use three tiers: large accounts by history, promised accounts by promise, the rest by bucket collection rates.

• Discount promises by each customer's track record.

• Compare forecast to actual every week and record the variance by tier.

• Use the forecast to direct collections, starting with the accounts where the gap between due date and expected payment is largest.

Frequently Asked Questions

How do you forecast cash collections from accounts receivable?

Start from a clean aging report, then forecast large accounts by each customer's average days to pay, use recorded promises for accounts in active follow-up, and apply historical collection rates by aging bucket to everything else. Add expected new invoices, then compare the forecast to actual receipts every week.

Why is a due-date cash forecast inaccurate?

Because many B2B customers pay after the due date. Atradius reports that 43% of credit-based B2B sales in the US are overdue. A forecast that assumes every invoice pays on its due date will overstate near-term cash.

What is a collection rate by aging bucket?

It is the share of the balance in an aging bucket (for example 31 to 60 days past due) that was historically collected within the next 30 days. It is calculated from your own aging snapshots and payment history.

How far ahead should an AR team forecast collections?

Most teams forecast weekly for the next 13 weeks. The first two to four weeks rely heavily on promises and large-account detail; later weeks rely more on historical collection rates.

How can the collections team improve forecast accuracy?

By following up on every overdue account on a regular cadence, recording promises with dates and amounts, and logging disputes as soon as they appear. That moves accounts from averages into dated commitments the forecast can trust.

Want to see how an AI collections agent keeps every account followed up and every promise recorded? Get started with Abivo.

Curious what this sounds like in practice? Here's a 98-second sample call: https://abivo.ai/#live-demo