Research question and scope
Published September 7, 2026.
This study asks which earning, redemption, return, expiration, and promotion events lead to loyalty-balance contacts and manual adjustments. It is an observational operations design, not a universal benchmark. The unit of analysis is one eligible support journey during a documented study window.
Dataset design
Build a pseudonymous ledger extract with event type, event time, points delta, eligibility state, case contact, decision, and adjustment approval. Exclude names, account credentials, and purchase details not needed for analysis. Write a data dictionary before extraction and retain the minimum fields needed to reproduce each measure. Set access and deletion rules with the client's privacy and security owners.
Measures and analysis
Measure contact rate by ledger state, posting delay, confirmed error, manual adjustment, repeat contact, and time to explanation. Analyze promotion cohorts separately from ordinary earning so campaign disputes do not distort platform reliability. Publish denominators beside rates. Report medians and a useful tail percentile for elapsed times, plus counts for every outcome. Review a sample of records to test whether system states match the underlying case history.
Compare results before and after a documented workflow change only when eligibility, observation windows, and outcome definitions remain consistent. Treat differences as signals for investigation, not proof that one change caused the result.
Quality controls
Check duplicate events, missing timestamps, impossible sequences, reopened journeys, and cases that cross the study boundary. Have an operations owner review classifications and a privacy owner review the extract. Keep source-system identifiers outside the reporting layer when they are not needed.
Scope and limitations
Program rules vary and can change during the study. A support ledger may miss partner events, household linking, or delayed returns. Contact data overrepresents customers who noticed and reported a discrepancy. The study cannot determine legal compliance. Small cohorts should be suppressed or combined under the organization's disclosure rules. Any decision affecting customer rights, money, access, or contractual terms remains with the authorized client owner.
Practical use
Use findings to choose one controlled improvement, such as clearer status language, a repaired handoff, or a better event code. Define the expected signal and a review date before release. Customer Care Staff can support data preparation, case review, and queue operations, while the client retains policy, system, legal, and financial authority.