The study would test how much warning a finance team gets from portal activity before a short payment becomes an aging problem.
Protocol status and question
This is a proposed research protocol, not a report of completed findings. It asks how much time passes between a portal payment event and the moment the finance team records a short-payment exception. The result would help a team decide whether portal data can support an earlier review. Keep the protocol and its approval date with the data file. The sponsor should record that decision before collection starts.
The protocol does not test whether the customer intended to short pay. It measures the observable gap between the event and the internal record. The sponsor should name the decision the result could inform, such as an alert threshold or a daily review step. If the decision changes, version the protocol rather than editing the old one. The sponsor should record that decision before collection starts.
Population and observation unit
Define the population as portal payments in a stated window that were less than the invoice balance. The unit is one payment event, with the portal timestamp, the ERP receipt time, the exception time, and the final resolution recorded. Payments that were corrected before an exception was recorded should be counted separately. Check the frame against a second source before collection. A count that cannot be rebuilt is not a population.
Link the payment to the invoice and the customer so the same event is not counted twice across portals. If the customer uses several portals, record the source and treat each event as a separate observation only when the underlying payment differs. Record every exclusion with its reason and the date it was applied. Write the frame as a dated list so a later reviewer can rebuild it. A count that cannot be rebuilt is not a population.
Definitions and measures
Define detection as the first recorded exception that names the short amount. Define lead time as the working time between the portal event and detection. Define resolution as the point when the amount is paid, credited, or approved as a deduction. Report each stage separately. Put the denominator and window beside every rate.
Measure the size of the short amount and the reason where available. A small short payment may be detected late without much consequence, while a large one may need a faster path. The analysis should show both the time and the amount so the reader can judge the operational importance. Test the wording on one record before the main collection begins. Keep the definition stable for the whole window.
Evidence standard
Use system timestamps for the portal event and the ERP receipt. For the exception time, use the record that first names the short amount. If the only evidence is an email, record the sender and the time and label it as a communication record rather than a system event. Link each classification to a document a second reviewer can open. Record the date the evidence was checked.
Preserve the original portal message and the invoice version it refers to. A later statement can show the amount but not what the customer saw at the time. The study should distinguish the evidence available at detection from the evidence assembled afterward. A summary without the source is not evidence. Link each classification to a document a second reviewer can open.
Sampling and collection
Use all short payments in the window if the volume is manageable. If sampling is necessary, stratify by amount, customer, and portal. Select the sample before reviewing detection times and keep a log of replacements. Do not exclude a payment because the record is messy. Log every replacement and the reason for it.
Pilot the extraction on ten payments and compare the timestamps across systems. Clarify how weekends and holidays affect the working-time calculation. Record any manual correction and the reason for it. Record the first classification before discussion. Select the sample before reviewing outcomes.
Analysis and reporting
Report the distribution of lead time, not only the average. Show the share detected within one day, one week, and one month, with the denominator. Break the results down by amount and by portal. A long tail of late detections may matter more than a fast average. Show the distribution, not only the average.
Describe the relationship between detection time and resolution time without implying that one causes the other. A late detection may be a symptom of a different process gap. The report should name the possible causes and the records that would test them. Report counts with rates and keep the denominator visible. Break the result down by the groups named in the protocol.
Limits and sources of error
The portal timestamp may reflect the customer's submission, while the ERP receipt reflects the bank's processing. The difference is part of the measured lead time but not entirely under the finance team's control. State this clearly in the report. Say what a second study would need to test. Describe the missing-data rate and its likely direction.
The study may miss short payments that were never recorded as exceptions. Those cases are important because they represent the weakest detection. Use a second source, such as a bank reconciliation, to estimate the size of that gap if possible. Say what a second study would need to test. State the population and period the result applies to.
Review cadence
Repeat the review after a portal change or a new exception rule. Compare the same stages and the same working-time definition. A shorter lead time is useful only if the exception is accurate and the resolution still follows. Assign the follow-up to a named owner. Set the next review date before the current one ends.
Assign the follow-up to the cash application or collections lead and set a date. The protocol should end with a specific decision about whether to add an earlier review step and how to measure it. Version the protocol when the method changes. Compare the result with the prior period and explain the change. Assign the follow-up to a named owner.
Sources
Put the protocol into operational context.
Discuss the queue, evidence, ownership, and safeguards that apply to your workflow.
Discuss your workflow