AR Aging Analysis and Collections Prioritization

Executive summary

An AR aging analysis starts with a report that sorts open receivables by how late they are. Used well, it shows which dollars are likely to arrive on their own, which will arrive only if someone acts, and which are heading toward a write-off. Used poorly, it becomes a call list sorted by balance, and the team spends its week on the hardest accounts.

This guide covers the method: bucket and concentration analysis, a prioritization score (dollars × probability of collection ÷ effort), roll rates that show how fast invoices deteriorate, promise-to-pay tracking, and reserve logic. DSO itself is covered in How to reduce DSO. Start by pulling an aging report that ties to the general ledger.

When to use this

Use this guide when:

• Past-due AR is growing and nobody can say which invoices matter most.
• Collectors work from the oldest or the largest invoice by habit.
• The 61+ day bucket keeps growing and the reserve is a flat percentage nobody has revisited.
• The receipts line of the 13-week cash flow forecast keeps missing.
• An auditor, lender, or buyer is asking how the allowance for doubtful accounts was set.

Do not start here if the AR subledger does not reconcile to the general ledger, or if unapplied cash and credit memos are months old. They make every bucket look older than it is. Clear them first.

AR aging analysis method: buckets, roll rates, reserves

1. Aging buckets

Age receivables by days past due date, not days since invoice date. Invoice-date aging mixes terms with lateness: a net-60 invoice at day 45 is not late, and a net-15 invoice at day 45 is 30 days late. Use consistent buckets, for example:

For each bucket, report dollars, share of total AR, and invoice count. Dollars show exposure. Counts show whether the problem is a few large items (a negotiation) or many small ones (a process).

2. Concentration

Customer share of AR = Customer balance ÷ Total AR
Top-N concentration = Sum of the N largest customer balances ÷ Total AR

Run it for total AR and for past-due AR separately. A customer that is 8% of AR and 25% of past due is a different case from one that is 8% of both. Also check concentration by industry, region, and sales owner, because a single event can affect a group together.

3. Prioritization score

Expected cash = Invoice amount × Probability of collection within the planning window
Priority score = Expected cash ÷ Effort (collector hours)

The planning window is the period you are working to, such as 30 days. Probability is a judgment, so anchor it: use the customer's own history (how often past-due invoices of this age were paid within the window), the status (undisputed, disputed, customer in distress), and any documented promise. Effort is a rough estimate of collector or executive hours, including the time of non-collections staff who have to resolve a dispute.

The score ranks work per hour, not per dollar. It puts cheap, likely cash first and large, unlikely cash into a separate, owned plan. That does not mean ignoring the large balance. It means a large balance with a low score needs an executive, a legal view, or a credit decision, not another reminder.

4. Roll rates

Roll rate (bucket A to bucket B) = Dollars in bucket A at the start of the period that are in bucket B at the end ÷ Dollars in bucket A at the start

Track the same invoices from one month-end to the next. Roll rate answers: of what was current last month, how much is now 1-30 days late? Of what was 61-90, how much is now 90+? Dollars that were paid or credited leave the matrix. Dollars that stay in the same bucket for two periods without being collected are worth a flag. Rising roll rates in the early buckets are the first sign of a collection or customer problem, and they appear before the 90+ bucket grows.

5. Promise-to-pay tracking

Promise-keep rate = Promises kept on or before the promised date ÷ Promises made
Promise-weighted receipts = Sum of promised amounts × Historical keep rate for that collector, customer type, or amount band

A promise is a date and an amount from a named person at the customer, recorded when made. Silence and "we'll look into it" are not promises. Measure the keep rate and use it to haircut the cash forecast.

6. Reserve logic

Reserve by bucket = Bucket balance × Estimated loss rate for that bucket
Total reserve = Sum of bucket reserves + Specific reserves for named accounts

The loss rate for a bucket is the share of dollars that ended as write-offs among invoices that sat in that bucket at some past point. Build it from your own history, by cohort. Then overlay specific reserves for accounts you know are impaired (bankruptcy, dispute you expect to lose), and consider current conditions. Accounting standards (CECL under US GAAP, expected credit loss under IFRS 9) require an expectation of losses, not just a historical average, and your auditors will have a view on the documentation they expect. Confirm the approach with them.

Worked example (illustrative)

The numbers below are made up to show the method. They are not benchmarks. All amounts are in $K.

Aging and concentration

A manufacturer has $6,000K of total AR at month-end.

Past due is 6,000 − 3,900 = $2,100K, or 35% of AR. The top five customers by balance (900, 700, 500, 400, and 300) total $2,800K, which is 46.7% of AR. The next step is to see how much of the $2,100K past due belongs to those five. If most of it does, the plan is five account conversations, not a campaign.

Prioritization

Six past-due accounts with the collector's estimates:

Work order: Alpha, Delta, Bravo, Charlie, Foxtrot, Echo. Foxtrot has the largest balance and ranks fifth. Sorting by balance would have put it first and consumed ten hours for $60K of expected cash. Foxtrot still needs a plan, but the plan is an executive call or a dispute resolution, with its own owner, not a place for a new collector to start.

The first three take four hours and are expected to bring in $436.5K (216.0 + 76.5 + 144.0). The estimates are judgments, so record them and compare them with outcomes next month. If accounts rated 80% pay 50% of the time, the probabilities need recalibrating.

Roll rates

Opening balances by bucket at the start of the month and the dollars from each that were one bucket older at month-end:

Each step costs more: 10% of current invoices became late, but 60% of those at 61-90 days rolled to 90+. The old bucket is not a slow version of the new one. At the current 61-90 balance of $300K, a 60% roll rate would put $180K into 90+ next month without intervention. That shows where calls change the outcome: the 31-60 bucket, before the large step.

Promise-to-pay

A collector secured 10 promises last month and eight were kept on or before the date, a keep rate of 80%. If this month the same collector holds $400K of promises, the promise-weighted receipts are $400K × 80% = $320K. If instead the historical keep rate for that customer type is 70%, the figure is $280K. Use the lower, better-matched rate in the forecast and note the difference.

Reserve

Illustrative loss rates by bucket (in a real analysis, these come from your write-off history by cohort):

The bucket reserve is $194.5K, or 3.2% of AR. Add specific reserves for named accounts on top if you expect a loss, and avoid counting the same invoice twice. The 90+ bucket is 3.3% of balance and 41% of the reserve (80.0 ÷ 194.5), which is the pattern you should expect: a small share of dollars carries most of the expected loss.

Diagnostic questions

Answer these from the ERP and the AR subledger.

1. Does the aging report tie to the general ledger at month-end, with unapplied cash and credits shown separately?
2. Is aging calculated from the due date, and do terms in the system match the contracts?
3. What share of past-due dollars sits in the top 10 customers?
4. How much of past-due AR is disputed, and who owns each dispute?
5. For each past-due invoice over your materiality threshold, who is the named owner, and when was the last documented contact?
6. What were the roll rates between buckets for the last six months?
7. How many promises to pay were made last month, and what share was kept?
8. What have write-offs been in the last three years, by the age bucket at which the invoice sat before write-off?
9. How was the current reserve set, when was it last updated, and who approved it?
10. Which customers have exceeded credit limits, or have shipments continuing while 61+ balances exist?

Reading the answers: if question 1 fails, fix data before analysis. If question 4 is high, route disputes to the owner of the cause before calling anyone. If question 6 shows rising roll rates from current to 1-30, look at billing quality and terms before collections effort.

The framework: look, rank, learn

The three parts run on a loop. The ranking is an estimate. The learning step compares the estimate with what happened and corrects it.

Step-by-step: a 60-day prioritization routine

Weeks 1-2: Clean base

1. Reconcile the aging report to the GL. Apply unapplied cash and clear old credit memos. Owner: Controller.
2. Rebuild aging on due dates and set the buckets. Owner: FP&A.
3. Compute concentration for total and past-due AR, by customer and by sales owner. Owner: FP&A.
4. Split past due into undisputed and disputed, and assign every invoice above the materiality threshold to a named owner. Owner: Collections lead.

Weeks 3-4: Rank and work

5. Score each past-due account: amount, probability, effort. Record the estimates. Owner: Collections lead with collectors.
6. Work the list from the top for the first week, then re-rank weekly.
7. Route low-score, high-dollar accounts to a named executive or credit decision with a date.
8. Log every contact and every promise with a date, amount, and the person at the customer.

Weeks 5-8: Measure and set policy

9. Compute roll rates from the last six month-ends. Owner: FP&A.
10. Track promise-keep rates by collector and customer type, and use them to adjust the cash forecast.
11. Rebuild the reserve from write-off history by bucket, add specific reserves, and write down the method. Review it with the auditors. Owner: Controller.
12. Set escalation thresholds from the roll rates (for example, executive contact when an account enters the bucket with a step-up in roll rate) and credit hold rules. Owner: CFO with Sales.

Templates

Priority worksheet

Promise-to-pay log

Weekly review agenda

1. Aging by bucket versus last week, and the movement of the 61+ buckets (5 min).
2. Top 10 by priority score and by dollars: status and blockers (10 min).
3. Promises made, kept, and broken, with the escalation applied (5 min).
4. Disputes: new, resolved, oldest, cause (5 min).
5. Decisions needed from Sales, Operations, or the CFO (5 min).

Metrics and KPIs

Setting targets without invented benchmarks. Anchor on your own best trailing six months for past-due share and roll rates, and on the root-cause analysis of what share of past due is disputed or caused by invoicing errors. Do not take a bucket-share target from a blog. Set the first-year goal as a reduction in the specific roll rate you identified as the largest step, and track it monthly.

Common mistakes

1. Sorting calls by balance. Symptom: collectors spend days on a large disputed account while small undisputed invoices age. Use the score.
2. Aging from invoice date. Symptom: buckets that do not match the terms. Age from the due date.
3. Treating buckets as steady. Symptom: the 90+ bucket gets the same effort as 1-30. Roll rates show that the earlier bucket is where effort changes the outcome.
4. Counting vague assurances as promises. Symptom: a forecast built on "they said soon." Require a date, an amount, and a named person.
5. Reserve as a flat percentage of AR. Symptom: the reserve does not move when aging deteriorates. Build it by bucket from your own losses.
6. Ignoring concentration. Symptom: one customer's delay causes a missed payroll buffer. Track top-N on total and past-due AR.
7. Never recalibrating probabilities. Symptom: accounts rated 80% pay half the time. Compare estimates with results monthly.
8. Continuing to ship to accounts with 61+ balances. Symptom: balances grow after the call. Define the credit hold rule and who can release it.

AI prompt assistance

AI helps with preparing the analysis and drafting. Supply the data yourself and verify every figure against the subledger before acting.

1. Build the aging and concentration view.

I will paste an AR open-items export with customer, invoice number, invoice date, due date, and open amount. Age each item by days past due as of [date] into Current, 1-30, 31-60, 61-90, and 90+. Return dollars, share, and invoice count by bucket, then the ten customers with the largest total and past-due balances and their share of each. Show the formulas used.

Check: confirm the total ties to the subledger, and recompute one bucket by hand.

2. Draft a priority worksheet.

From this list of past-due invoices with amount, days past due, dispute flag, and collector notes, propose a probability of collection within 30 days for each as low, medium, or high with a one-line reason. Do not assign numbers. I will set them.

Check: the collector or account owner sets the probabilities. AI notes are a prompt for judgment, not a substitute for it.

Implementation guide

Governance: the CFO owns the targets, the Controller owns the process and the reserve, Collections owns the worksheet and promise log, and Sales owns customer conversations and terms. Report past due, 61+ share, roll rates, and promise-keep rate monthly alongside cash. Expect the first benefit from re-ranking work within weeks, and the reserve and policy changes to take a quarter. Do not announce a target in the first month.

Downloadable assets

The priority worksheet, promise-to-pay log, and weekly agenda above can be copied into a spreadsheet.

Related resources

• How to reduce DSO: a 90-day collections plan: the DSO formulas and the broader collections program.
• Working capital: 3 levers that move your cash cycle: where receivables sit next to inventory and payables.
• The working capital playbook: sequencing the levers.
• Extending payables without damaging suppliers: the other side of the cash cycle.
• AP process design: approval, matching, and payment runs: process controls on the payables side.
• 13-week cash flow forecast template: turning promises into a receipts forecast.

Frequently asked questions

How is AR aging used to prioritize collections?

AR aging shows where receivables are sitting, and you prioritize by ranking accounts on expected cash per hour of effort rather than by balance or age alone. Multiply the invoice amount by the probability of collection in your planning window, divide by estimated collector hours, and work the highest scores first. Large, low-scoring balances get an owned plan, such as an executive call or dispute resolution.

How do you read an accounts receivable aging report?

Read it by days past due date, not days since invoice date, so terms are not confused with lateness. For each bucket, look at dollars, share of total AR, and invoice count: dollars show exposure, and counts show whether the problem is a few large items or many small ones. Then check customer concentration for total AR and past-due AR separately.

What is a roll rate in collections?

A roll rate is the share of dollars in one aging bucket at the start of a period that sit in a later bucket at the end. Track the same invoices month-end to month-end, excluding anything paid or credited. Rising roll rates in the early buckets are usually the first sign of a collection or customer problem, appearing before the oldest bucket grows.

How do you calculate a bad debt reserve?

Multiply each aging bucket's balance by an estimated loss rate for that bucket, add specific reserves for accounts you know are impaired, and adjust for current conditions. Build the loss rates from your own history: the share of dollars that ended as write-offs among invoices that once sat in that bucket. Accounting standards expect a forward-looking estimate, not only history.

What is promise-to-pay tracking?

Promise-to-pay tracking records each customer commitment as a date and an amount from a named person, then measures how often promises are kept. Promise-keep rate is promises kept on or before the date divided by promises made. Use that rate to haircut the cash forecast, since a vague answer like "we will look into it" is not a promise.

How this works in a System of Performance

A prioritization score is only useful if it is compared with what happened next. The estimate, the owner, the promise, and the cash that arrived live in different places in most companies: the ERP, email, and a spreadsheet. 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 much of your receivables are aging, bring it to us as a launch partner.