Turning nine data sources into one labour forecasting model

Multi-venue coastal hotel. Labour was the dominant, poorly-controlled cost, with demand highly multi-factored — weather, events, school and public holidays, room and function bookings — and rosters set on instinct rather than forecast.

• Implemented a three-step framework: forecast demand → plan capacity → plan labour
• Integrated nine data sources — revenue, bookings, events & ticketing, holidays, weather, functions, roster software — into a single source of truth
• 52-week demand forecast by food, beverage and venue, compared to budget
• Capacity planning translating workforce size and composition into theoretical capacity, with casual flex
• Labour planning converting forecast demand into maximum roster hours against labour-%-of-revenue targets, FOH and BOH

• Labour as % of revenue at demand peak: 42.6% → 30.0% post-implementation
• −9.7pts on the long-term trend
• Planning solution in live use by the operations team