How accurate are these collection probabilities?

Directional, not exact. The 95–97% / 80–90% / 60–75% / 30–40% decay curve comes from practitioner guidance and reflects typical B2B behavior. Your own business's actual curve, built from 12+ months of internal history, is always more accurate than industry numbers. If you don't have that history yet, the industry ranges are a usable starting point — they get refined as your real data accumulates.

Why rank by expected recovery instead of just biggest balance?

Because a $50K invoice in the 120+ day bucket is worth maybe $15K in expected cash, while a $30K invoice in the 31–60 bucket is worth ~$25K. Ranking by raw balance sends you chasing the wrong invoices. Expected recovery (balance × bucket probability) gives the honest read on which calls actually convert to cash.

What's the difference between this and just reading the AR aging report?

The standard aging report shows you who owes what. It doesn't tell you who to call first, how to weight expected cash, or which tone to use per customer. This guide adds probability weighting, dollar-exposure ranking, and history-driven tone. It turns 'review and decide for each invoice' into 'decide once, then act on the top five.'

How often should I run this?

Weekly. The aging report changes fast — new invoices issued, payments cleared, customers slipping or recovering. A list built on Monday is usable through Friday. A list built last month is stale by week two. Most B2B businesses spend 15–30 minutes a week on this; the cash recovered usually returns 10–50x the time.

When should I write off an invoice?

When it's 180+ days old, the customer has had multiple unanswered contact attempts, and the amount doesn't justify legal action. Holding aged AR on the books indefinitely inflates apparent assets and clutters your weekly collections review with low-value attention drains. Writing off doesn't mean giving up — many businesses keep written-off accounts in a separate 'recovery if responded' tracker, but they stop counting them as expected cash.

Decision Guide

AR Collections Prioritization Guide: Which Invoices to Call This Week

This AR collections prioritization guide ranks overdue invoices by recovery probability and dollar exposure, then tells you which customers to contact this week. Collection probability decays predictably with age: roughly 95–97% in the 0–30 day bucket, 80–90% at 31–60 days, 60–75% at 61–90 days, and 30–40% past 90 days. Focusing on the customers driving the biggest dollar exposure in your 60+ buckets typically recovers more cash in 30 days than blanket follow-up across every overdue account.

Data you need

AR aging by customer
Pull the standard AR aging report from your accounting system. Each customer with open invoices, broken into 0–30, 31–60, 61–90, and 90+ day buckets.
Customer payment history (when available)
Months of payment data you have on each customer. Three or more historical payments lets you use customer-specific timing instead of bucket averages.
Invoice notes / dispute flags
Disputes, partial payments, or payment plan agreements. Disputed invoices need a different conversation than slow-payer invoices.
Operating cash floor
The minimum bank balance you want to maintain. Used to gauge urgency of collections this week vs. routine follow-up.

Decision this guide helps produce

Ranked customer list (top 5–10 to call this week)
Customers ordered by dollar exposure × recovery probability × strategic value. The 20% you'd actually call, not the 100% the aging report shows.
Expected dollar recovery from this week's effort
Sum of (invoice amount × bucket probability) for the ranked list. Realistic, not optimistic.
Tone-matched follow-up draft per customer
Firmer for repeat late payers. Softer for trusted customers who slipped once. Direct for disputes. Tone is driven by payment history, not the aging bucket alone.
Escalation flags
Invoices past 90 days where standard follow-up has stopped working. Candidates for settlement offers, third-party collections, or write-off.
Healthy-book delta
Your current aging distribution vs. healthy benchmark (≈80% in 0–30 day bucket, under 5% in 90+). Surfaces whether the issue is one customer or a structural collections gap.

Decision model

The prioritization framework runs in three steps. **Step 1: Apply collection probability by aging bucket.** Published practitioner guidance gives the rough decay curve: • 0–30 days past due: 95–97% collection probability • 31–60 days: 80–90% • 61–90 days: 60–75% • 91–120 days: 30–40% • 120+ days: 10–40%, often falling toward the lower end without active recovery These come from practitioner sources including [Leib Solutions](https://www.leibsolutions.com/a-r-collectibility-by-age/) and AR-management writeups from [HighRadius](https://www.highradius.com/resources/Blog/accounts-receivable-aging-report/), [Paystand](https://www.paystand.com/blog/aging-report-in-finance), and [Brex](https://www.brex.com/spend-trends/accounting/accounts-receivable-aging-reports). They're directional — every business's actual curve differs. If you have 12+ months of internal collection history, your own curve is more accurate than industry guidance ([Tesorio's DSO countback method](https://www.tesorio.com/blog/dso-countback-method-how-finance-teams-can-accurately-forecast-cash-flow) walks through how to build it). **Step 2: Compute expected recovery per invoice.** Expected cash = invoice amount × collection probability of its bucket. Sum across all overdue invoices to get total expected recovery. Sum just the top 5–10 customers (by expected recovery) to see how concentrated the opportunity is. For most B2B businesses, the top 20% of overdue customers account for 70–80% of overdue dollars — the same Pareto pattern that shows up across most operational metrics. **Step 3: Rank by impact and assign action.** The action depends on bucket and history: • **60+ bucket, no prior contact this cycle:** phone call this week. Phone clears most overdue invoices within 7–10 days. Email exchanges typically take 30+ days for the same result. • **30–60 bucket, repeat slow payer:** firmer reminder with a specific deadline. • **30–60 bucket, normally on-time customer:** standard reminder, no escalation needed. • **90+ bucket, no engagement:** escalation candidate. Options: settlement offer (often 80–90% in exchange for immediate payment), third-party collections, or in some cases write-off. The healthy-book benchmark sits in the background of every read. A typical B2B business on net-30 terms wants roughly 80% of total AR in the 0–30 day bucket and under 5% in 90+. Below those thresholds means a structural collections issue, not a one-off customer problem.

Worked example

A $2M services business with $192K in total AR. The aging report: • Current (0–30 days): $112K across 14 customers (58% of AR) • 31–60 days: $42K across 6 customers (22%) • 61–90 days: $24K across 4 customers (12%) • 90+ days: $14K across 3 customers (7%) Aging distribution issues: 90+ at 7% is above the under-5% healthy threshold. 60+ combined at 19% is at the edge of acceptable. The pattern suggests collections discipline has slipped, not just one bad customer. Top 5 ranked by expected recovery: 1. Customer A: $18K in 61–90 bucket. Expected recovery: $18K × 70% = $12.6K. History shows they always pay eventually but slip on every cycle. Action: phone call this week with a firm payment date request. 2. Customer B: $11K in 90+ bucket. Expected recovery: $11K × 35% = $3.85K. Two prior reminders, no response. Action: escalation candidate — settlement offer at $9K or send to collections. 3. Customer C: $9K split across 31–60 and 61–90. Expected recovery: ~$6.8K. Long-time customer (six years) usually on time, slipped recently. Action: brief, low-pressure check-in call — likely an internal AP issue at their end, not a collection problem. 4. Customer D: $7K in 31–60 bucket. Expected recovery: $6.0K. Repeat 30+ day late on three of last four invoices. Action: firmer reminder noting that future projects will require deposits. 5. Customer E: $5K in 61–90 bucket. Expected recovery: $3.25K. Disputed an invoice in writing. Action: resolve dispute conversation before sending another reminder. Total expected recovery from this five-customer call list: ~$32K within 30 days. The rest of the overdue book (15 other customers, ~$48K combined) gets standard automated reminders. Not zero attention, just not the manual phone-call slot this week. The shift: instead of vague "improve collections," the owner has a five-name call list with specific recovery dollars attached and tone-matched approach for each.

TL;DR

Collection probability decays with age. Rough curve: 95–97% (0–30), 80–90% (31–60), 60–75% (61–90), 30–40% (91–120), 10–40% past 120 days.
Rank overdue customers by expected recovery (balance × bucket probability), not by raw balance. A $30K invoice at 30 days is worth more expected cash than a $50K invoice at 120 days.
For most B2B businesses, the top 20% of overdue customers drive 70–80% of overdue dollars. Focus the call list on those.
Phone clears most overdue invoices in 7–10 days. Email exchanges take 30+ days for the same result. Reserve phone for 45+ day balances; let automation handle the rest.
Healthy aging: ≈80% in 0–30 day bucket, under 5% in 90+. Above those thresholds, the problem is process, not one customer.

A founder told us last week: “I’ve got six figures of overdue invoices and I’m spending Mondays staring at the aging report trying to figure out who to call. By Friday I’ve talked to two people and the list still feels just as big.”

The aging report shows you who owes what. It doesn’t tell you who to call first, how much cash to actually expect, or what tone to use per customer. This framework takes the same data and converts it into a five-name call list with expected dollars attached. Same fifteen minutes a week, dramatically more cash recovered.

Below is the math, the framework that ranks customers, and what to do with the 90+ day bucket where most cash quietly dies.

What’s the collection probability by aging bucket?

Practitioner guidance is consistent on the rough decay curve. The probability that an overdue invoice eventually collects drops sharply with age:

Aging bucket	Collection probability
0–30 days past due	95–97%
31–60 days	80–90%
61–90 days	60–75%
91–120 days	30–40%
120+ days	10–40% (falling without active recovery)

These come from practitioner sources including Leib Solutions’ A/R collectibility guidance, AR-management writeups from HighRadius, and the analysis in Paystand’s 2026 aging report guide. They reflect typical B2B behavior — your own business will have its own curve.

The reason the curve matters: aging compounds twice. The probability of collection drops, and the cost of collection effort rises. A $20K invoice at 30 days is worth roughly $19K in expected cash. The same invoice at 120 days is worth $6–8K, and you’ll spend more time chasing it.

CECL-style modeling and treasury practice both use historical aging curves rather than a single static rate (Miller Cooper’s CECL aging writeup covers this well). If you have 12+ months of payment history, your own curve is more accurate than industry guidance. Tesorio’s DSO countback method walks through how to build it from your own data.

Why rank by expected recovery, not raw balance?

Because raw balance sends you chasing the wrong invoices.

A $50K invoice in the 120+ bucket has an expected recovery of $50K × 30% = $15K. A $30K invoice in the 31–60 bucket has an expected recovery of $30K × 85% = $25.5K. The smaller balance is the better call. The bigger balance is more likely a write-off conversation than a collection conversation.

The math:

Expected recovery = Invoice amount × Bucket probability
Rank list = Top customers by expected recovery, ordered descending

For most B2B businesses, the Pareto pattern holds: roughly 20% of overdue customers account for 70–80% of overdue dollars. The call list this week is those top customers, ordered by expected recovery, not the full aging report.

The rest of the overdue book gets standard automated reminders. They’re not ignored. They just aren’t worth the manual phone-call slot this week.

What does a real prioritized list look like?

A $2M services business with $192K in total AR.

Aging distribution:

Current (0–30 days): $112K across 14 customers (58% of AR)
31–60 days: $42K across 6 customers (22%)
61–90 days: $24K across 4 customers (12%)
90+ days: $14K across 3 customers (7%)

The 90+ bucket at 7% is above the under-5% healthy threshold. 60+ combined at 19% is at the edge. The pattern reads like collections discipline has slipped, not one bad customer.

Top 5 ranked by expected recovery:

Rank	Customer	Balance	Bucket	Expected	Action
1	A	$18K	61–90	$12.6K	Phone call this week with firm payment date
2	B	$11K	90+	$3.85K	Escalation candidate; settlement offer or collections
3	C	$9K	mixed 31–90	$6.8K	Low-pressure check-in; likely AP issue at their end
4	D	$7K	31–60	$6.0K	Firmer reminder; note future projects require deposits
5	E	$5K	61–90	$3.25K	Resolve disputed invoice before sending another reminder

Total expected recovery from this week’s calls: ~$32K within 30 days.

The remaining 15 customers (~$48K combined) get the standard automated reminder cycle. Not zero attention, just not the manual slot.

The change isn’t more work. It’s the same work pointed at the right targets.

Stop building this list from spreadsheets every Monday. Start a 14-day free trial. Fynso reads your QuickBooks AR aging and surfaces the ranked call list in the daily brief, with a drafted follow-up per customer.

How do you tone-match the follow-up?

The same overdue invoice from two different customers warrants two different conversations. Tone gets driven by payment history, not the aging bucket alone.

Firm: repeat 30+ day late payers, customers who’ve slipped on multiple consecutive invoices, customers who avoid responses. The message references the specific invoice, the specific overdue period, and a clear escalation path if not resolved by a specific date.

Softer: long-time customers who normally pay on time, customers with a clear one-time issue (a known internal AP delay, a personnel change). The message frames it as “wanted to check in” rather than “this is a problem.”

Direct dispute resolution: customers who’ve raised a dispute in writing. The message acknowledges the dispute, proposes a resolution path, and doesn’t add a payment reminder until the dispute is closed.

Escalation: 90+ day balances with no engagement after multiple touches. Settlement offer language, or notice that the account will be reviewed for third-party collections.

Most accounting systems support automated reminder templates for the first three. The fourth almost always requires a person.

What about the 90+ bucket specifically?

The 90+ bucket is where most cash dies, and it’s where the most attention compresses into the wrong outcome.

At 91–120 days, expected recovery drops to 30–40% even with active follow-up. The cost of pursuing collection (phone calls, sometimes legal notice) rises. By 120+ days, recovery falls to 10–40% depending on whether you take formal action.

Three responses at 90+, ranked by cost:

Settlement offer. Often 80–90% of the invoice in exchange for immediate payment. The customer accepts because they recognize the alternative (collections, legal) is more expensive for them too. The business accepts because $0.85 today is worth more than $1.00 in six months at maybe-50% odds.

Third-party collections. Specialized agencies that operate on a contingency fee (typically 25–50% of recovered amount). Recovery rates are lower than direct effort, but you stop spending your team’s time on the account.

Write-off. When the amount doesn’t justify legal action and the customer has stopped responding. The accounting cost is small; the time saved is real. You can still keep a “recovery if responded” tracker for these accounts without counting them in expected cash.

The wrong response at 90+: continuing to send the same email reminder you sent at day 60. If the same approach didn’t work the first three times, it won’t work the fourth.

What does a healthy AR aging distribution look like?

For a B2B business on net-30 terms, roughly:

0–30 days: 80%+ of total AR
31–60 days: 10–15%
61–90 days: under 5%
90+ days: under 2%

These are starting points. Medical practices billing insurance run structurally older aging (carrier processing). Construction runs older (progress billing). But within your own industry, the principle holds: heavy concentration in current, very little in 90+.

The healthy-book delta in this guide surfaces where your distribution sits relative to the benchmark. Above 10% in 90+ means a clear problem. Above 20% in 60+ combined means a structural issue, not a one-off customer. Different responses for each.

The full discipline behind weekly AR review is in The AR Aging Report: How to Read It in 3 Minutes a Week. The relationship-preserving collection playbook is in How to Collect AR Faster Without Damaging Customer Relationships.

How does Fynso run this continuously?

Fynso reads your QuickBooks AR aging every morning, applies the collection probability curve (refined from your own payment history when available), and surfaces the ranked call list in the daily brief. Each entry comes with:

Customer name, total balance, aging bucket
Expected recovery dollar amount
A drafted follow-up message, tone-matched to payment history
Suggested action (call, send, escalate, resolve dispute)

You approve, edit, or skip per customer. The system handles the cadence and the drafting. You handle the conversation that actually moves money.

The same data feeds anomaly detection. When a customer’s payment timing shifts (consistently on-time customer slipping to 50 days), it shows up in the brief as a credit risk signal before it becomes a collection problem.

The AR Aging Report: How to Read It in 3 Minutes a Week — the weekly discipline behind this guide.
How to Collect AR Faster Without Damaging Customer Relationships — the collection system that prevents the awkward 90-day conversation.
When to Let Go of a Customer — what to do when chronic late payment becomes the relationship.
Cash Flow vs. Profit — why AR aging matters even when the P&L looks healthy.

The Monday-morning call list, automated.
Start a 14-day free trial — connect QuickBooks, get your prioritized AR list with drafted follow-ups every morning. No credit card required.

Frequently asked questions

How accurate are these collection probabilities?: Directional, not exact. The 95–97% / 80–90% / 60–75% / 30–40% decay curve comes from practitioner guidance and reflects typical B2B behavior. Your own business's actual curve, built from 12+ months of internal history, is always more accurate than industry numbers. If you don't have that history yet, the industry ranges are a usable starting point — they get refined as your real data accumulates.
Why rank by expected recovery instead of just biggest balance?: Because a $50K invoice in the 120+ day bucket is worth maybe $15K in expected cash, while a $30K invoice in the 31–60 bucket is worth ~$25K. Ranking by raw balance sends you chasing the wrong invoices. Expected recovery (balance × bucket probability) gives the honest read on which calls actually convert to cash.
What's the difference between this and just reading the AR aging report?: The standard aging report shows you who owes what. It doesn't tell you who to call first, how to weight expected cash, or which tone to use per customer. This guide adds probability weighting, dollar-exposure ranking, and history-driven tone. It turns 'review and decide for each invoice' into 'decide once, then act on the top five.'
How often should I run this?: Weekly. The aging report changes fast — new invoices issued, payments cleared, customers slipping or recovering. A list built on Monday is usable through Friday. A list built last month is stale by week two. Most B2B businesses spend 15–30 minutes a week on this; the cash recovered usually returns 10–50x the time.
When should I write off an invoice?: When it's 180+ days old, the customer has had multiple unanswered contact attempts, and the amount doesn't justify legal action. Holding aged AR on the books indefinitely inflates apparent assets and clutters your weekly collections review with low-value attention drains. Writing off doesn't mean giving up — many businesses keep written-off accounts in a separate 'recovery if responded' tracker, but they stop counting them as expected cash.

Want Fynso to apply this with your connected data?

See AR priorities with connected data