Model structure for practical forecasting
Forecasts fail when they rely on one aggregate trend line. Padel demand moves by weekday, time block, and competition windows.
A practical model starts simple: segment bookings into actionable buckets and track confidence ranges for each segment.
- Segment by weekday, prime-time blocks, and member vs guest traffic.
- Use rolling 8-12 week baselines before adding trend adjustments.
- Apply scenario bands: conservative, expected, and high-demand cases.
- Recalibrate monthly to avoid stale assumptions.