Skip to main content
A café's stamp-card audit measured the checklist, not the counter

A café's stamp-card audit measured the checklist, not the counter

A paper from the School of Continuous Improvement follows a strip-mall café's stamp-card audit, and the independent check that found no matching gain in fairness, into the Master of Applied Measurement's coursework.

An independent customer check ran alongside the audit for fourteen rounds and found no matching gain in stamp fairness; the Master of Applied Measurement keeps the design on its syllabus regardless

For seven months, a single strip-mall café’s loyalty-stamp audit passed nearly every counter-staff member it reviewed. A second, quieter measure run alongside it — what a random sample of customers said their own card should show — found no matching improvement in fairness, and never closed the gap even once the audit’s visits stopped being announced in advance. The Master of Applied Measurement has published the case in full, calibration and all, rather than waiting for a cleaner result.

The paper’s account follows the audit — a checklist scoring how closely a till operator’s punches matched customer receipts, and how promptly they were issued — from its four-week pilot through fourteen ordinary fortnightly rounds. The independent measure the audit’s own design never included tracked flat throughout, and swapping the fortnight’s scheduled visit for a surprise one changed nothing about the shortfall.

The School treats the mismatch as the strongest material its practicum has produced this year, not as a result to quietly retire.

An instrument that passes almost everyone it reviews has usually stopped measuring what it was built for. The honest response is to keep teaching it exactly as it ran, not to fix it before the next cohort notices.

— Dr Torun Ezeigwe, Senior Lecturer and Director, Master of Applied Measurement

Dr Ingeborg Nwachukwu, who convenes the Capability Sprints course the café’s own coordinator completed before designing the audit, said the calibrated weighting now “travels with the course to every host placement that adopts it, unchanged.”

The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.l7wnkr.