How to Forecast Accounts Receivable Collections Without Pretending Every Customer Pays on Time

I have a question about accounts receivable that I think deserves more attention.

When exactly do we expect that money to show up?

Not when the invoice is due.

Not when the customer should pay.

When do we actually expect the money to arrive in the bank?

Those are three different questions. And judging by some of the cash forecasts businesses produce, apparently we’re treating them as interchangeable.

Adorable.

As any mom can tell you, telling someone to do something by Friday does not guarantee it will happen by Friday.

Sometimes Friday becomes Monday.

Sometimes you get a very convincing explanation about why Friday was never realistic in the first place.

And sometimes you’re still waiting on Wednesday.

Accounts receivable works much the same way, except the consequences tend to involve payroll, vendors, and uncomfortable conversations with the CFO.

So let’s talk about how to forecast accounts receivable collections using actual customer behavior rather than everyone’s best intentions.

Because there’s a difference between knowing what customers owe you and knowing when you can spend it.

Your accounts receivable aging report is not a cash forecast

I like accounts receivable aging reports.

They’re useful. They tell us who owes money, how much they owe, and how long invoices have been outstanding.

What they don’t necessarily tell us is when the money will arrive.

And that’s an important distinction.

Imagine your business has $400,000 in outstanding accounts receivable.

Your aging report shows:

Aging category	Outstanding
Current	$220,000
1–30 days past due	$100,000
31–60 days past due	$55,000
Over 60 days past due	$25,000
Total	$400,000

Hypothetical example.

At first glance, this doesn’t look particularly alarming.

More than half the balance is current.

Lovely.

But suppose you need $180,000 in customer collections over the next two weeks to cover payroll, vendor payments, and other obligations.

How much of that $400,000 will actually arrive?

The aging report doesn’t answer that question.

It tells you the status of the invoices, not the reliability of their payment dates.

And this is where I think businesses need to stop treating accounts receivable as one big pile of money that’s politely waiting to become cash.

Some invoices are practically on their way.

Others are going to require a conversation.

And a few might need more than a conversation.

Start with the customers, not the average

One of the first things I’d examine is how individual customers actually pay.

Not what their contracts say.

What they do.

Let’s say you have three customers, all with Net 30 payment terms.

Customer A consistently pays in 28 days.

Customer B typically pays in 45 days.

Customer C pays somewhere between 60 and 75 days, usually after several reminders and what I imagine is a spirited exchange of emails.

All three have the same contractual payment terms.

They do not have the same cash behavior.

If I forecast all three customers as paying on day 30, I’ve created a lovely picture of how the world ought to work.

Unfortunately, payroll doesn’t accept lovely pictures.

Build a simple customer payment history

I’d start by reviewing six to twelve months of paid invoices, if the history is available and representative.

For each invoice, capture:

• Customer name and invoice number.
• Invoice date and contractual due date.
• Original amount and any credits or adjustments.
• Actual payment date and amount, including partial payments.
• Days between invoice issuance and payment.
• Any disputes, deductions, or unusual circumstances.

Then look for patterns.

Does the customer pay consistently?

Are payments tied to a monthly approval cycle?

Do large invoices take longer than small ones?

Has payment behavior changed recently?

A customer who historically paid in 35 days but has recently started paying in 60 deserves more attention than a customer who has always paid in 45.

The first may indicate deteriorating behavior.

The second may simply be predictable.

And predictability is useful.

Even when it arrives late.

How to forecast accounts receivable collections, step by step

Here’s how I’d build a practical AR collections forecast.

Step 1: Clean up the starting balance

Before forecasting anything, make sure the outstanding receivables are real.

I know.

A radical concept.

Reconcile the AR subledger to the general ledger. Review unapplied payments, duplicate invoices, credit memos, disputed balances, and invoices that may already have been settled.

If a customer paid yesterday but the cash hasn’t been applied to the invoice, that balance shouldn’t appear as another future collection.

Otherwise, we’ve managed to forecast the same cash twice.

A wonderful trick, if only the bank would cooperate.

Also, establish a clear forecast cutoff date. Cash already received belongs in the opening cash balance, not in future receipts.

Step 2: Assign an expected payment date

For each open invoice, estimate when the customer will actually pay.

I would use the strongest evidence available.

A verified payment in process is generally more persuasive than a vague promise.

A recent, credible payment commitment may be more useful than a historical average.

A reliable customer payment pattern is better than blindly assuming the contractual due date.

And if an invoice is disputed, I’d want the dispute resolved—or at least a defensible resolution timeline—before treating the entire amount as dependable cash.

Here’s a simplified example.

Customer	Invoice	Due date	Expected receipt
A	$40,000	Oct. 15	Oct. 15
B	$30,000	Oct. 15	Oct. 30
C	$25,000	Oct. 18	Nov. 5
D	$20,000	Oct. 20	Uncertain
Total	$115,000

Hypothetical example.

If we forecast by contractual due date, all $115,000 appears due in October.

But based on customer behavior and collection status, only $70,000 has an expected October receipt date.

Another $25,000 is expected in November.

And $20,000 is uncertain.

Same invoices.

Very different cash picture.

That’s why I prefer to distinguish between what is contractually owed and what we reasonably expect to collect during the forecast period.

Step 3: Separate expected collections from uncertain collections

I don’t particularly like forecasts that assign every customer a 90% probability of paying.

That can create the appearance of sophistication without improving the underlying judgment.

Sometimes probabilities are appropriate, particularly when supported by enough historical observations.

But for short-term cash management, I also want to see which specific invoices are uncertain.

For example:

Expected: A customer has a reliable payment history or a credible, near-term payment commitment.

At risk: A customer has a history of delays, a missed commitment, or an unresolved administrative issue.

Disputed or highly uncertain: The amount, timing, or likelihood of collection is materially unclear.

These aren’t accounting classifications or universal probability bands. They’re management categories for making cash assumptions visible.

And I would not automatically treat an invoice as certain just because someone in sales says the customer is good for it.

I’m sure they’re lovely people.

I’d still like the payment evidence.

Step 4: Put collections into the weeks when cash should arrive

Once expected payment dates are assigned, group receipts into weekly periods.

For a business managing near-term liquidity, a rolling 13-week cash forecast can be useful.

The important word is rolling.

We update it as new information becomes available.

We don’t create it once, admire its formatting, and return three months later to discover that reality has been conducting its own financial planning exercise.

Here’s a hypothetical collection schedule for the $115,000 invoice portfolio above.

Week	Expected collections
Oct. 12–18	$40,000
Oct. 19–25	$0
Oct. 26–Nov. 1	$30,000
Nov. 2–8	$25,000
Uncertain timing	$20,000
Total AR	$115,000

Notice that the uncertain $20,000 hasn’t disappeared.

It’s still a receivable.

It simply isn’t included in the dated base collection forecast until we have better evidence.

That’s not pessimism.

That’s refusing to spend money that hasn’t arrived.

For a complete cash forecast, I’d also add collections from future invoices that haven’t yet been issued, using the sales and billing forecast and realistic payment lags. I’d keep those separate from existing AR to avoid double counting.

What happens when a major customer pays late?

This is where an accounts receivable forecast becomes useful for management.

Let’s say your company has the following cash position.

Opening cash: $120,000

Expected collections over the next two weeks: $180,000

Required payments over the next two weeks: $250,000

The simple calculation is:

$120,000 + $180,000 − $250,000 = $50,000

That leaves $50,000 in ending cash.

Not an enormous cushion, but positive. If you’re deciding what constitutes a reasonable buffer, I’ve also written about how much cash a small business should keep in reserve.

Now suppose one customer responsible for $75,000 of expected collections delays payment until the following week.

The revised calculation becomes:

$120,000 + $105,000 − $250,000 = −$25,000

Well.

That escalated quickly.

Original ending cash: $50,000

Revised ending cash: −$25,000

Cash deterioration: $75,000

The business hasn’t necessarily lost $75,000.

The customer may still pay every penny.

But the timing creates a potential $25,000 funding shortfall at the end of the period.

And that’s the point.

A profitable business can still have a cash problem because customers pay later than the business needs them to. For a closer look at why, see my working capital forecasting guide.

The next question is no longer, “How much revenue are we generating?”

It’s, “How do we cover the timing gap without creating a bigger problem?”

What I would do with that information

I’d want to know whether we can accelerate a legitimate collection, draw on an existing credit facility, adjust discretionary spending, or negotiate payment timing with vendors where appropriate.

I’d also look at the minimum daily cash balance, not just the two-week ending balance.

A positive Friday balance doesn’t help if payroll clears on Wednesday and the customer pays Thursday.

That’s a detail worth knowing before Wednesday.

The decision depends on liquidity, available financing, contractual obligations, and the cost of each alternative.

I wouldn’t automatically delay vendor payments. Preserving cash by damaging a critical supplier relationship can be an expensive victory.

The goal is to identify the gap early enough to make a deliberate decision.

Preferably without a last-minute meeting that begins with, “So, we have a little situation.”

Don’t confuse DSO with an actual collection schedule

Days Sales Outstanding, or DSO, is useful.

It measures the relationship between accounts receivable and credit sales, expressed in days.

A common calculation is:

DSO = Average Accounts Receivable ÷ Net Credit Sales × Days in Period

For example, if average AR is $300,000 and quarterly net credit sales are $600,000 over a 90-day period:

DSO = $300,000 ÷ $600,000 × 90 = 45 days

That tells me something useful about the business’s collection cycle.

What it doesn’t tell me is whether the $75,000 invoice I need next Thursday will arrive next Thursday.

DSO is an aggregate metric.

An invoice-level forecast is a timing tool.

I want both, but for different reasons.

DSO can help me understand trends, evaluate working capital, and estimate collections over longer horizons.

For immediate liquidity decisions, I want invoice-level visibility into the payments that matter most.

And I want to understand the customers behind the numbers.

Because averages can hide some spectacularly inconvenient behavior.

Use collection assumptions that reflect reality

Once the invoice-level forecast is working, I’d use historical collection behavior to improve assumptions.

For a larger receivables portfolio, it may not be practical to manually forecast every small invoice.

That’s where customer segments and collection curves become useful.

For example, suppose a group of similar customers historically pays:

Time from invoice date	Cumulative amount collected
By day 30	50%
By day 45	75%
By day 60	90%
By day 90	97%

Illustrative collection curve, not industry benchmark data.

For a hypothetical $100,000 cohort of new invoices, that would imply:

• $50,000 collected by day 30.
• Another $25,000 between days 31 and 45.
• Another $15,000 between days 46 and 60.
• Another $7,000 between days 61 and 90.
• $3,000 remaining after day 90.

Notice that the percentages are cumulative.

We don’t apply 50%, 75%, 90%, and 97% separately and add them together.

That would be a very impressive collections forecast.

Also mathematically impossible for the same $100,000 portfolio without additional cash sources.

For a useful model, calculate these curves from actual invoice cohorts, account for partial payments and credits, and segment customers when their behavior differs meaningfully.

Also distinguish amounts still outstanding from amounts ultimately written off. Late cash and uncollectible cash are not the same problem.

If payment behavior changes, update the assumptions.

The model should learn something from reality.

Otherwise, why are we maintaining it?

The five things I’d review every week

I don’t think accounts receivable forecasting needs to become a complicated ritual.

It needs to become a dependable management process.

Each week, I’d review five things:

1. Actual collections versus forecast. Which receipts arrived, which slipped, and why?
2. Large upcoming receipts. Which customer payments materially affect next week’s cash position?
3. Overdue and disputed invoices. What is preventing collection, and who owns the next action?
4. Changes in customer behavior. Are historically reliable customers becoming slower?
5. Cash consequences. Does a delayed receipt change a hiring decision, vendor payment, financing need, or other commitment?

I’d also track forecast accuracy by receipt week, with enough detail to separate timing errors from permanent collection losses.

A $50,000 receipt arriving one week late is not the same as $50,000 that will never arrive.

Both matter.

But they require different responses.

And while I appreciate a beautifully formatted dashboard as much as the next finance professional, I care considerably more about whether someone is acting on the information.

A red overdue-invoice indicator isn’t a collection strategy.

It’s a red indicator.

Someone still needs to call the customer. That is also where practical accounts receivable support can make a difference.

When a simple forecast is enough—and when it isn’t

Not every business needs a sophisticated collections model.

If you have a small number of customers, predictable payment behavior, and comfortable cash reserves, a straightforward invoice schedule may be sufficient.

I’d rather see a simple forecast that’s updated consistently than an elaborate model nobody trusts.

But as the business grows, the approach may need to change.

I’d consider more detailed modeling when customer concentration is high, payments are inconsistent, cash reserves are tight, or the business has significant weekly obligations.

The same applies when a company is expanding rapidly.

Growth often creates more receivables before it creates more cash.

Which is one of those financial realities that sounds unfair until you remember that customers are not required to finance your ambitions.

A business can be growing, profitable, and increasingly short of cash at the same time.

That’s why I think collections forecasting deserves a place alongside revenue forecasting and expense planning.

The collection forecast should change what management does

Here’s the distinction I care about most.

A collections report tells management what happened.

A collections forecast should help management decide what to do next.

If a major receipt is likely to arrive late, the forecast should identify the potential cash gap.

If a customer is consistently paying beyond terms, someone should evaluate the commercial relationship and credit policy.

If collections repeatedly miss the forecast, the assumptions need to change.

And if the business is constantly surprised by cash shortages despite having a healthy accounts receivable balance, I would question whether we’re forecasting cash at all.

Or simply putting due dates into a spreadsheet and hoping for the best.

I understand the appeal.

Hope requires very little maintenance.

Unfortunately, it also has a terrible track record as a liquidity management tool.

I don’t expect every customer to pay exactly on time.

That’s not how business works.

What I do expect is for Finance to recognize the difference between an invoice, an expectation, and cash in the bank.

Those differences affect hiring, spending, financing, and the ability to meet obligations.

And they’re worth understanding before someone asks why a profitable company suddenly needs to borrow money to make payroll.

If your business is still forecasting collections based entirely on invoice due dates, you might find my Cash Forecasting services worth exploring.

Of course, you could keep assuming everyone pays on time.

As any mom can tell you, that level of optimism is adorable.

I just wouldn’t build payroll around it.