How to prioritize a collections list.

Age is a symptom, not a plan.

Sorting by days overdue feels rigorous, but it quietly makes three bad assumptions: that every dollar is equally important, that every overdue invoice is equally likely to stay unpaid, and that invoices which are not yet due deserve no attention. None of these hold in a real B2B portfolio.

The oldest invoice on the report is often old precisely because it is stuck—in a dispute, a missing purchase order, or an insolvent customer—where one more reminder changes nothing. Meanwhile, a large invoice due next week to a customer whose payment behavior has recently slipped gets no attention at all, because the aging report has no column for “about to go wrong.”

Late payment is not an edge case. Intuit’s 2025 survey of 2,487 US small businesses found 56% were owed money from unpaid invoices, and 47% had invoices overdue by more than 30 days. With that much exposure, the order in which a small team spends its follow-up hours is a real financial decision. See the survey.

Rank dollars, not rows.

The first correction is simple: weight every open invoice by its open balance. Ten small invoices can matter less than one large one, even when the small ones are older. If you do nothing else after reading this guide, re-sort your worklist by open amount within each aging bucket. Effort will immediately move toward the balances that fund payroll.

Value alone is still incomplete, because a large invoice to a customer who always pays two days late is not really at risk. That is why the next step matters.

Combine amount with the chance of delay.

For each open invoice, estimate the chance it will be paid late. The estimate can come from a simple rule—this customer’s historical late rate—or from a tested statistical model. Then compute one number per invoice:

Estimated cash at risk = open invoice amount × chance of late payment

This single multiplication fixes both failure modes at once. A $40,000 invoice with a 55% chance of delay ($22,000 at risk) outranks a $2,000 invoice with a 90% chance of delay ($1,800 at risk), and both outrank a large invoice to a reliably prompt payer.

Ranked, value-weighted worklists are standard practice in enterprise collections software—HighRadius, for example, sells worklist prioritization built on exactly this idea—and a published accounts-receivable case study describes the same approach in production. The method is not exotic; it is simply rare below the enterprise price point. Read the case study.

Not yet due is not the same as nothing to do.

A prioritized list still needs a time dimension. Three timing states deserve different treatment:

1. Overdue. The estimate is now partly confirmed. The question shifts from “will this be late?” to “why is it late, and who needs to act?”
2. Due within a week. This is the highest-leverage window. A polite payment-date confirmation before the due date is cheap, relationship-safe, and often prevents the delay entirely.
3. Due later. High-risk invoices here go on a watch list. Contacting a customer three weeks before the due date usually annoys more than it collects.

Grouping the open balance by expected payment week also gives the team a cash-planning view: how much should arrive this week, next week, and later—and how much of each week’s total is sitting on risky invoices.

One list, one order, one owner per line.

Bring it together as a single queue, sorted by estimated cash at risk, showing for each invoice: account, open amount, due date, chance of delay, cash at risk, expected payment date, and a suggested next step. Cap the active queue at what the team can genuinely work—ten to twenty lines for most small teams—and let everything below the cap wait for next week.

Review the queue in the weekly collections meeting. The ranking proposes; the account owner disposes. Someone on the team usually knows about the phone call yesterday, the credit note in progress, or the strategic renewal that the spreadsheet cannot see. The list orders the conversation; it does not replace it.

Escalate effort, not tone.

A useful default ladder, applied by risk and timing rather than age:

1. Monitor. Low risk, not yet due. Do nothing; unnecessary chasing costs goodwill.
2. Confirm before due. Higher risk, due soon. Ask the contact to confirm the payment date—phrased as routine, not as suspicion.
3. Remind after due. A short, factual note referencing the invoice and terms.
4. Owner call. High cash at risk and overdue. Check for disputes, missing POs, and promised dates before assuming unwillingness to pay.
5. Escalate under your policy. Only with human review, and never automated from a statistical score.

Keep every step in the first person and relationship-safe. The estimate justifies which account gets attention first—never how the customer is treated.

Record outcomes and re-rank weekly.

Each week, record what happened: who paid, who promised a date, which promise held. Refresh the closed-invoice history, rebuild the queue, and compare last week’s risk estimates with reality. If the estimates persistently disagree with outcomes, fall back to transparent value-and-age rules until the data supports something better.

You can build this exact queue—ranked by cash at risk, grouped by expected payment week, with a tested reliability check—from two CSV exports in the free PaidWhen workspace. Nothing leaves your browser.