Most mid-market companies do their 30-day cash flow forecast one of two ways: they assume all outstanding invoices will be paid on their stated due date (optimistic, often wrong), or they apply a flat discount factor to the AR balance based on historical collection rates (more conservative, but still rough). Both approaches leave a lot of signal on the table.
Your AR data — if you've been operating for more than a year and have payment history in your accounting system — contains customer-level payment timing profiles that are far more predictive than either approach. Customer A consistently pays at net+32. Customer B pays within two days of receiving the invoice. Customer C, despite net-30 terms, averages net+58 and requires follow-up. If you know these patterns and can apply them to your current open invoice list, your 30-day cash forecast improves significantly.
This piece walks through a practical method for building a pattern-based cash flow forecast from your AR data. It doesn't require modeling software or a data science background — it requires your aging report, 12 months of payment history, and about half a day to build the framework once.
Why AR Data Is Your Best Cash Signal
For most B2B companies, receivables represent 40–70% of short-term cash inflows. Unlike AP (payables), where you control the timing, AR cash flows depend on your customers' payment behavior — behavior that is, it turns out, fairly consistent customer by customer.
A company in the industrial distribution business we spoke with had 220 active accounts. When we looked at their 18-month payment history, 68% of their accounts had a days-to-pay standard deviation of less than 8 days. That means for more than two-thirds of their customers, you could predict within a week when they'd pay. The remaining 32% were more variable — seasonal patterns, intermittent dispute activity, or genuinely unpredictable AP processes on their end.
The practical implication: if you can identify your predictable payers and use their actual historical average days-to-pay to schedule expected cash inflows, you have a much better 30-day cash forecast than any flat discount factor provides. The unpredictable 32% you forecast conservatively — use their worst-case timing, not their average.
Step 1: Build Customer Payment Profiles
Export 12 months of invoice and payment data from your accounting system. For each invoice, you need: customer name, invoice date, due date, payment date (if paid), and amount. If you have it, include payment terms per invoice — some customers may have different terms for different order types.
For each customer, calculate:
- Average days-to-pay from invoice date — use invoice date, not due date, for consistency across customers with different terms
- Standard deviation of days-to-pay — this tells you how reliable the average is. Standard deviation below 10 days: treat this as a predictable payer. Standard deviation above 20 days: treat as unpredictable.
- Historical collection rate — what percentage of invoiced amount gets paid (some customers short-pay routinely due to disputes or deductions)
- Sample size — how many invoices in the 12-month window. Fewer than 5 invoices means the profile isn't reliable — treat those customers as unpredictable regardless of the average.
The result is a payment profile table: one row per customer, columns for average DTP, standard deviation, collection rate, and reliability tier (predictable / moderate / unpredictable).
Step 2: Apply Profiles to Your Current Open Invoice List
Pull your current accounts receivable aging report. For each open invoice, look up the customer's payment profile and calculate an expected payment date:
For predictable payers: Expected payment date = Invoice date + Customer's average days-to-pay. If the invoice date is January 15 and the customer averages 34 days from invoice date, expected payment is February 18.
For moderate variance payers: Use average days-to-pay + one standard deviation as your expected date. This builds in a conservative buffer without assuming worst case.
For unpredictable payers: Use the due date plus 21 days as a conservative floor. For invoices already past due with an unpredictable payer, treat cash receipt as outside the 30-day forecast window unless you have active collection engagement documented.
For each invoice, apply the customer's historical collection rate to the invoice amount to get the expected receipt amount. A customer who routinely short-pays by 5% for deductions shouldn't show up as a 100% cash receipt in your forecast.
Step 3: Build the Weekly Cash Receipt Schedule
Aggregate expected receipts by week. Sum all invoices with expected payment dates in Week 1 (days 1–7), Week 2 (days 8–14), Week 3 (days 15–21), Week 4 (days 22–30). The result is a 4-week cash receipt schedule from AR alone.
Separately, track confidence level per week. Week 1 and Week 2 should have higher confidence — you know what's currently due and overdue, and you have payment timing data for most customers. Week 3 and Week 4 have lower confidence — some invoices in that window haven't been issued yet, and prediction accuracy naturally drops with time horizon.
Add a separate row for "high-risk overdue" — invoices that are 30+ days past due from customers with poor payment history or no current engagement. These should be explicitly flagged as uncertain and excluded from the working cash balance until collected.
The Forecasting Limitation You Should Acknowledge
We want to be direct about what this method doesn't do. Historical payment patterns reflect normal operating conditions. They don't predict a customer experiencing a liquidity crisis, a dispute you don't know about yet, or a seasonal AP shutdown (some companies freeze AP processing in late December and certain quarter-end periods). Those events break the historical pattern.
The practical workaround: overlay your AR collections activity on the forecast. Invoices that have had recent payment commitments ("we'll process it Thursday") or confirmed dispute resolution can be moved to a higher-confidence category. Invoices that are past due and have received zero response should be moved to the uncertain bucket regardless of what the historical pattern says.
A pattern-based AR forecast is more accurate than a simple discount-rate model, but it's still a forecast. Use it as a planning tool and a decision-support input, not as a substitute for cash balance monitoring. Update it weekly as payment confirmations come in and invoices age.
How Better AR Operations Improve Forecast Accuracy
There's a compounding relationship between AR collections quality and forecast accuracy. When you're running proactive, personalized follow-up sequences — reaching customers before they hit their worst-case timing, surfacing disputes before they become payment holds — two things happen: DSO shrinks (cash comes in faster) and payment behavior becomes more predictable (customers pay more consistently because they're being managed more consistently).
In the pattern-based forecasting model, more predictable payers create tighter confidence intervals. A customer who used to have a 22-day standard deviation in payment timing but now pays within a 9-day window after consistent follow-up is now a forecasting asset, not a forecasting liability. Over 12–18 months of operating with disciplined AR collections, the population of "unpredictable" customers typically shrinks and your 30-day cash forecast becomes a meaningfully more reliable business planning tool.
That's not an accident — it's the feedback loop between collections discipline and financial visibility. Better AR process produces better cash data, which produces better forecasts, which produces better decisions about hiring, inventory, and capital deployment. The unsexy job of chasing invoices turns out to be load-bearing for the finance function's ability to plan.