Customer Operations Research
Auditing loyalty point adjustments for undocumented goodwill
A research method for separating documented loyalty corrections from undocumented goodwill, using the reason code, the balance history, and the approval trail.
Key takeaways
- A loyalty adjustment audit should separate a documented correction from an undocumented goodwill credit.
- The reason code, the balance history, and the approver are the minimum evidence for a reviewable adjustment.
- Counts by reason and by approval band reveal more than a single total of points issued.
- A missing customer explanation is a record gap, not proof that the adjustment was improper.
Question and scope
This review asks how a dispensary can tell a documented loyalty correction from an undocumented goodwill adjustment. The unit is one points adjustment applied to one customer account during a defined period. The review examines the reason code, the balance history before and after the change, the order or campaign reference, the approver, and any note explaining the customer's request. It does not judge whether the customer deserved the points or whether the program's economics are sound.
Set the scope to one store, one loyalty platform, and one reporting period. Record the program rules in force during that period, including earning rates, redemption rules, expiration, and the approval bands for manual adjustments. If the store changed platforms or ran a major promotion inside the window, split the period. A platform migration can create duplicate events that look like goodwill if the analyst does not separate them.
What the sources establish
The Federal Trade Commission's consumer-protection guidance supports accurate, non-deceptive descriptions of a loyalty benefit and its conditions. NIST privacy guidance is relevant because a points record can be linked to a named customer and a purchase history. The Virginia Cannabis Control Authority provides the regulatory setting for licensed retail, though it does not publish a loyalty-adjustment benchmark. None of these sources defines what a local goodwill credit should look like.
That means the store must build its own evidence standard from its program rules and its records. Public sources can explain why the terms should be clear and why customer data should be limited. They cannot show whether a specific adjustment was justified. The review should describe what the record contains and avoid converting an absent note into a finding of misconduct.
Definitions and measures
Define a documented correction as an adjustment linked to a specific earning event, redemption error, campaign term, or system fault. Define an undocumented goodwill adjustment as a manual credit with no order reference, no campaign reference, and no written customer explanation. Define an approval band as the range of point or dollar values that a named role may approve. Define a repeat adjustment as a second manual credit to the same account inside a stated period.
Measure the number of adjustments, the number with a reason code, the number linked to an order or campaign, the number with an approver, and the number with a customer explanation. Report the points issued and the estimated dollar value separately, because a large points total may reflect a small number of high-value cases. Also report the number of adjustments that were reversed or corrected a second time. A reversal rate is a useful signal of unclear rules or hasty entries.
Sampling and collection
Sample adjustments rather than customer accounts, then collect every record attached to the selected adjustment. Include credits, debits, reversals, and expired-point restorations so the review covers the full range of manual activity. Stratify the sample by reason code, approval band, and staff role. Record the sampling rule and the observation window before opening the records. Keep the adjustment identifier in the table so a second analyst can reproduce the selection.
Collect the loyalty event log, the order or receipt reference, the campaign terms if one applies, the approval record, and the customer communication. Mask names and contact details in the analysis table. If the platform does not retain a before-and-after balance, reconstruct it from the event log and note the limitation. A missing field is data, not a reason to drop the case from the sample.
Analysis and reporting
Report the share of adjustments with each required field, not just the total number of adjustments. Show the counts by reason code and by approval band so the reader can see where undocumented credits concentrate. Compare the documented corrections with the undocumented goodwill cases on value, repeat rate, and reversal rate. A high repeat rate in one band may point to a vague threshold rather than a training problem.
Inspect a small set of cases in each category to see whether the note explains the customer's request and the staff decision. Report examples in paraphrased, de-identified form and keep the original records in the review file. Show the number of adjustments that lack an approver separately from those that lack a customer explanation. Those are different control gaps and may need different fixes.
Reconciling the event log
Reconcile the loyalty event log with the order system before treating a manual credit as goodwill. An order that failed to post points may appear twice if the customer retried the purchase or if the platform replayed an event. Match the order number, the timestamp, and the points value against the event log rather than relying on the balance alone. A duplicate event that is corrected by a second entry is a system error, not a goodwill decision.
Look for patterns in the timing of manual entries. A cluster at the end of a shift may point to a closing routine that rushes corrections, while a cluster after a promotion launch may point to unclear campaign terms. Record the time between the original event and the adjustment. Long gaps deserve a note about why the customer waited, because the delay may affect whether the correction is still appropriate under the program rules.
Coding the customer explanation
Code the customer explanation into a small set of categories such as missing purchase points, failed redemption, promotion confusion, or service recovery. Use the customer's own words as the source and keep the category as the analyst's summary. A category that appears often may signal a communication problem rather than a dishonest claim. A category that never appears may mean the field is not being used.
Do not require the customer to provide a receipt for every small correction, but record what evidence was available. If the customer provided a screenshot, note it without storing the image in the general analysis table. Separate a claim that was verified from one that was accepted on the customer's account. That distinction keeps the review honest without turning a routine adjustment into an investigation.
Limits and sources of error
The loyalty log can show what changed, when, and by whom. It cannot prove the customer's state of mind, the full content of a verbal conversation, or whether the points were the customer's only remedy. A missing note may reflect a busy counter rather than a deliberate undocumented credit. The review should state that ambiguity instead of assigning motive.
The dollar value of a point is a program assumption, not a market fact. If the store uses an estimated redemption value, label it and keep the assumption stable across periods. A change in the estimate can move the reported value more than a change in adjustment behavior. Also separate a platform-generated correction from a staff-entered credit, because the two have different error paths.
Review and next action
Choose one change: require an order or campaign reference before a manual credit, add a short customer-explanation field, publish the approval bands, or review repeat adjustments weekly. Assign the owner and the next sample date. Keep the change small enough to test at one counter or one shift. Repeat the review with the same definitions so the comparison is honest.
Close the review when the store accepts the result and schedules the next check. Report the remaining uncertainty beside the counts. The loyalty adjustment routine describes the counter practice this review measures. Use the result to improve the record, not to rank individual employees. Schedule the next check and name the person who will run it.
Sources
The sources below support the consumer-protection, privacy, and regulatory context for this review. Keep the citation list visible beside the analysis table during the review. They do not measure this dispensary, and no source is presented as a local benchmark. Each link should be checked for its date and scope before reuse. The operational finding comes only from the store records described in the method.
- Federal Trade Commission, Advertising and Marketing
- NIST Privacy Framework
- Virginia Cannabis Control Authority
- GAO, Assessing Data Reliability
This article describes a measurement method rather than legal or compliance advice. Confirm current requirements with the applicable authority and qualified advisers. Keep the local evidence separate from those requirements. Record the review date and the owner who accepted the interpretation. Store the signed scope note with the review file, and give the owner a copy before the next sample begins.
Reviewed by the DispensaryVA editorial team on 2026-11-06.