Slop University researchers benchmark barista memory against the loyalty app's prediction and find the till record an unreliable witness to what customers actually wanted.
A 41-cafe benchmark finds regulars accept a wrong guess rather than correct it, and the loyalty app learns the same mistakes
Slop University has long held that measurement should meet people where they already queue. In new work from the School of Continuous Improvement, that meant the front of the coffee line, where the question “the usual?” has been quietly serving as a memory test nobody marks. A research poster released this week by Dr Torun Ezeigwe and Dr Anouk Mensah reports the first accuracy benchmark for the counter guess, observed across 41 cafés and 6,187 transactions.
Every instrument teaches the people it measures, and the counter is no exception. We are glad to have met this one before anybody made it a target.
— Dr Torun Ezeigwe, Senior Lecturer and Director, Master of Applied Measurement
The study harmonises counter observation, off-premises debriefing and prediction logging into a single accuracy picture, then asks what the till record can and cannot be trusted to witness. Scored at the register, the guess approaches infallibility; scored against what customers had intended to order, it does not, and much of the difference turns out to be courtesy rather than recall. The loyalty app, trained on the same till record, inherits the courtesy.
“What stays with us is how generous the exchange is,” said Dr Anouk Mensah, Postdoctoral Fellow in the School of Emergent Priorities. “Everyone at the counter is trying to hand everyone else the right answer, and the record keeps the peace by writing it all down as agreement.”
The poster continues the University’s practice of taking everyday accuracy seriously, and its findings are already informing service-standard conversations beyond the study sites. The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.g8sr99.