# How to Forecast Accounts Receivable Collections

Source: https://www.cashflowforecasttemplates.co.uk/guides/how-to-forecast-accounts-receivable-collections
Updated: September 25, 2026

> To forecast accounts receivable collections, start from your aged receivables list, place each invoice in the week or month that customer usually pays, apply a collection curve to future invoices (for example 10 percent paid in the month, 60 percent the next, 25 percent the month after), and leave disputed or very old invoices out until they are resolved.

For most businesses that invoice customers, **collections** are the biggest and least predictable line in the cash flow forecast. Get them right and the rest of the forecast is usually close; get them wrong and the whole thing is optimistic. This guide explains how to forecast accounts receivable collections using your own data rather than your payment terms.

## Terms are not behaviour

Your invoices may say “30 days”, but that doesn’t mean customers pay in 30 days. Some pay in 20, many in 40 to 50, a few much later. A forecast that assumes everyone pays on terms will always show more cash arriving sooner than it really does.

The fix is to forecast from **behaviour**: how each customer, or each type of customer, actually pays.

## Step 1: Start with aged receivables

Your accounting software can produce an **aged receivables** (or aged debtors) report, which lists unpaid invoices by how long they’ve been outstanding. It’s the starting point for the next few weeks of collections.

For each customer, place their unpaid invoices in the week or month you realistically expect payment:

| Customer | Owed | Payment history | Forecast |
|---|---|---|---|
| Harbour Foods | 8,400 | Pays reliably on day 30 | Week 2 |
| Delta Property | 12,000 | Pays at 45–60 days after a reminder | Week 5 |
| Riverside School | 4,600 | Pays at month end after approval | Week 4 |
| Northside Clinic | 3,100 | 90 days overdue, disputed | Leave out |

## Step 2: Measure your collection curve

For invoices you haven’t raised yet, use a **collection curve**: the share of each month’s invoices you typically collect in the same month, the next month, and later. Work it out from the last six to twelve months:

1. Take invoices raised in a month, say January.
2. Record how much of that January total was paid in January, February, March and later.
3. Repeat for several months and average the percentages.

A typical result might be:

| Paid in | Share of invoices |
|---|---|
| Month invoiced | 10% |
| Next month | 60% |
| Month after | 25% |
| Never / written off | 5% |

Apply that curve to each future month’s invoicing. If you plan to invoice $50,000 in May, forecast $5,000 in May, $30,000 in June and $12,500 in July.

## Step 3: Segment where it matters

If different customer groups pay very differently, use separate curves: large corporates vs small businesses, public sector vs private, insurers vs self-pay patients, or marketplace payouts vs direct sales. Two or three segments usually capture most of the difference.

## Step 4: Allow for slippage and bad debt

- **Bad debt:** if 2 to 5 percent of invoices are never paid, leave that out of the forecast.
- **Seasonal slowdowns:** many customers pay slower in holiday periods and at their own year end.
- **Concentration:** if one customer is a large share of receivables, model their payment separately and consider a late-payment scenario.

## Step 5: Check with DSO

**Days sales outstanding** (DSO) summarises how long customers take to pay:

**DSO = accounts receivable ÷ credit sales in the period × days in the period**

If DSO is rising month after month, cash is arriving later, and the forecast should reflect it. Tracking DSO monthly is a simple early warning.

## Worked example

A firm invoices $50,000 a month with the curve above, and has $38,000 of existing receivables expected over the next two months.

| | Month 1 | Month 2 | Month 3 |
|---|---|---|---|
| Existing receivables | 30,000 | 8,000 | 0 |
| From month 1 invoices | 5,000 | 30,000 | 12,500 |
| From month 2 invoices | | 5,000 | 30,000 |
| From month 3 invoices | | | 5,000 |
| **Total collections** | **35,000** | **43,000** | **47,500** |

Even though the firm invoices $50,000 every month, collections settle at $47,500 from month 3, 5 percent below invoicing because of bad debt. A forecast built on invoice totals would overstate cash by $15,000 in month 1 alone.

## Week by week

In a [13-week forecast](/guides/13-week-cash-flow-forecast), monthly curves need one more step: which week of the month does the money arrive? Look at your bank statements. Many businesses find that most customer payments cluster in particular weeks, such as the last week of the month when larger customers run their payment batches, or the first week when direct debits are collected. Spread each month’s expected collections across its weeks in the same pattern. For the largest customers, forecast by their known payment run dates.

## Receivables in different industries

- **Construction:** progress claims are approved before they’re paid, and clients hold back retainage. Forecast the approval delay as well as the payment delay, and schedule retainage releases separately.
- **Medical and dental practices:** insurers and government programs pay on their own cycles, and some claims are denied and resubmitted. Use a curve per payer type.
- **Agencies and professional services:** retainers are usually reliable; project milestones slip when clients are slow to approve work.
- **Wholesale and manufacturing:** larger customers often pay on 45 to 60 day terms, and on their own payment run dates.
- **Trucking:** brokers and shippers commonly pay in 30 to 45 days unless invoices are factored, in which case most of the cash arrives within days, less a fee.

## Common mistakes

- **Forecasting from terms.** “30 days” on the invoice is an aspiration, not a forecast.
- **Using invoice totals.** Sales are not cash until they’re collected.
- **Ignoring disputes.** A disputed invoice rarely pays on time; leave it out.
- **One curve for everyone.** If public-sector clients take 60 days and small businesses 20, one average hides both.
- **Not updating after each week.** Receivables change daily; the forecast should change weekly.
- **Forgetting credit notes.** Refunds and credit notes reduce what customers will pay; net them off.

## Improving collections, not just forecasting them

Forecasting receivables often reveals how much cash is waiting in customers’ hands. Ways to speed it up:

- invoice immediately when work is done
- ask for deposits or stage payments
- offer card or direct debit payment and include a payment link on invoices
- send reminders before the due date and call on day seven
- review credit terms for slow payers

More ideas are in [12 ways to improve cash flow](/guides/12-ways-to-improve-cash-flow-quickly).

## Templates with collections built in

Several industry templates include a collections tab that applies a curve like the one above automatically: enter what you invoice and how quickly customers pay, and the forecast updates. They include [agencies](/cash-flow-forecast-template/agency), [construction](/cash-flow-forecast-template/construction) (with retainage), [law firms](/cash-flow-forecast-template/law-firm), [medical practices](/cash-flow-forecast-template/medical-practice) and [manufacturing](/cash-flow-forecast-template/manufacturing).

### What is the best way to forecast accounts receivable collections?
Use each customer’s actual payment history rather than your payment terms, and forecast existing invoices one by one from your aged receivables report.

### What is a collection curve?
The percentage of a month’s invoices typically collected in that month, the next month and later months, based on your own history.

### How do I calculate days sales outstanding (DSO)?
Divide accounts receivable by total credit sales for a period, then multiply by the number of days in the period. For example, $90,000 owed on $180,000 of quarterly sales is 45 days.

### Should I include overdue invoices in my forecast?
Include moderately overdue invoices in the week you realistically expect payment. Leave disputed or very old invoices out and treat them as a bonus if paid.

### How often should I update receivables forecasts?
Weekly, as part of a 13-week cash flow forecast, because receipts are the least predictable line.