A Singapore hotel stopped buying standby limos by optimizing each hour

A large Singapore luxury hotel ran limousine service around the clock, averaging about 9,000 trips per month. To protect guest service during events and holidays, planners kept outsourced vehicles on standby, a habit that created millions of dollars of annual extra cost.

The clever answer was not a grand transport platform. The project combined predictive analytics, spreadsheet optimization and robotic process automation so planners could decide how many in-house and outsourced limousines were needed for each hour of the day.

That framing changed the managerial question from 'how much spare capacity feels safe?' to 'which hours actually need standby vehicles?' The system preserved service quality while cutting the expensive cushion that had been spread too broadly.

The case is a useful reminder that operational research often wins by fitting the tool to the operator. The solution used widely available tools, which made it easier for planners and controllers to keep using it after the research project.

Plan standby capacity by hour, not by fear

Hourly capacity planning matched the real shape of demand better than daily averages.

Spreadsheet optimization kept the model close to the planners' existing workflow.

RPA reduced the clerical work that otherwise stops small optimization tools from becoming routine.

When demand spikes are scary, quantify the spike hour by hour; a modest optimizer can replace blanket backup capacity without lowering service.

The published case reports savings of up to S$3.2 million per year while improving service level. The approach was presented in INFORMS Journal on Applied Analytics as a practical hospitality capacity-planning case.