Demand forecasting

Demand forecasting is predicting how many cars will be on hire, or booked, for each future date and car category. It's the input that every pricing and fleet decision rests on.

A simple method

The pick-up method works in a spreadsheet:

Forecast final bookings = bookings on hand today + expected pick-up from here

Expected pick-up = average bookings that arrived between this lead time and pick-up, for the same date last year

Then adjust by hand for anything new: an event, a school holiday that's moved, a new competitor.

Worked example

Hypothetical. A Friday pick-up 21 days out. You have 18 compacts booked. Last year, for the same Friday, 12 more compacts were booked in the final 21 days.

• Forecast = 18 + 12 = 30 compacts
• You own 28

You're forecasting two more than you can supply. Lift the rate now, so the cars you have left go to the late bookings that pay more. A stop-sell on the cheapest channel is another option.

If a large event lands on that date this year and didn't last year, the last-year pick-up understates demand. Add a judgement uplift and write down why.

Why it matters for a rental operator

Without a forecast, every price change is a reaction to bookings that already happened. With one, you can act two or three weeks ahead, inside your booking window.

Check accuracy every month with MAPE. A forecast you never score never gets better.

See it in Rate Pulse

Rate Pulse doesn't forecast your demand. Its Dates to watch section lists confirmed events and public and school holidays in the next 14 days, each with a source, which is the context your forecast needs. Start at Rate Pulse. Building forecasts on your own booking data is part of our revenue management work.