The Number Stayed Clean

A Quasi-Experiment on a Suburban Gym Chain's Booking-Reliability Score, Comparing Attendance and Cancellation Timing Before and After Its Introduction

A paper from the School of Continuous Improvement tracks a twenty-week rollout of a suburban gym chain's booking-app reliability score across 22 sites, finding booking-fill rates at popular classes rise from 93.6% to 99.3% while door-scanner attendance holds flat (71.5% versus 71.7%, p = 0.62). The gap traces to cancellation timing: the share of cancellations landing in the six-minute band just outside the score's two-hour penalty cutoff triples, from 14.6% to 51.3%. An unplanned natural ablation --- four sites left with a cosmetic-only score display by a backend migration --- shows the clustering requires the score's functioning waitlist-promotion mechanism, not merely its visibility.

@misc{slop_ln1awt,
  author       = {Casimir Beng and Mirela Hanke and Solveig Adeyemi},
  title        = {The Number Stayed Clean: A Quasi-Experiment on a Suburban Gym Chain's Booking-Reliability Score, Comparing Attendance and Cancellation Timing Before and After Its Introduction},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.ln1awt},
  url          = {https://slop.university/outputs/slop-paper-a-quasi-experiment-on-ln1awt/},
  version      = {1.0},
  note         = {Research paper},
}