Forecast car hire demand in a spreadsheet

By the DayRunner team · 7 November 2026

Use a pickup forecast. Take the bookings already on the books for each future week, then divide by the share of final bookings you usually have at that point, based on last year's booking curve. That gives a forecast of hire days per week, which you turn into utilisation and compare with your fleet.

It takes one export from your booking system and three formulas. Here's how.

What you need

• A list of last year's bookings with: car type, pick-up date, return date and the date the booking was made.
• Today's bookings on the books, with the same fields.
• Your fleet count by car type, per week.

That's it. No special software. A spreadsheet handles a fleet of a few hundred cars fine.

The idea in one line

If last year you had 55% of a week's final hire days booked 30 days out, and this year you have 140 hire days booked 30 days out, the week will probably finish near 140 ÷ 0.55 = 255 hire days.

The 55% comes from your booking curve: how bookings build up over the booking window before pick-up.

Step by step

1. Count hire days, not bookings

   For each booking, split it into the days the car is out. A 5-day hire is 5 hire days, spread across the weeks it touches.

2. Build last year's curve

   For each week last year, work out the share of its final hire days that were booked 90, 60, 30, 14 and 7 days before the week started.

3. Average the curve

   Average those shares by season (peak, shoulder, low), so one odd week doesn't drive it.

4. Forecast this year

   For each future week: hire days on the books ÷ the curve share at today's distance out.

5. Turn it into utilisation

   Forecast hire days ÷ (cars × 7). Compare with the target for that week.

6. Check it every week

   When each week finishes, compare forecast with actual and log the error.

A worked example

These numbers are hypothetical. Say you run 40 compacts and SUVs from Christchurch, and it's early November.

Last year's curve for the shoulder season looks like this:

Now your forecast for three weeks:

The week of 22 Dec forecasts at 119%. You can't hire out more cars than you have, so that week will sell out early. That's your signal to raise rates now, add a minimum rental length, or bring cars in from another depot.

The week of 15 Dec forecasts at 80%. It's fine but not full. Hold your rates and watch it again next week.

Days out of 37 and 44 sit between the 30 and 60 points in the curve. Interpolate a straight line between them, as above, or add more points to your curve.

Check against the market

Your forecast tells you how full you'll be. The market tells you what others are charging for the same days. You need both.

In the Rate Pulse sweep of 7 Nov 2026, the typical economy car at Christchurch was NZ$160 a day, but by pick-up day it ran from $118 to $203. The three dearest pick-up days were 12 to 14 Nov, in the middle of NZ Cup and Show Week and the Canterbury Anniversary Day holiday on 13 Nov. If your forecast and the market both point up for the same dates, you can raise rates with more confidence. If your forecast is high and the market is flat, you may just have sold too cheap, too early.

Copy this template

One row per week per car type. The formulas use spreadsheet column letters for this layout.

• G, forecast hire days: =E2/F2
• H, forecast utilisation: =G2/(C2*7)
• J, absolute error once the week is done: =ABS(I2-G2)/I2
• Average J across finished weeks for your MAPE.

Adjust for what last year didn't have

A curve from last year assumes this year books the same way. It won't, exactly.

• Holidays move. Easter and school holidays fall on different weeks each year. Line weeks up by holiday, not by date. The NZ holiday calendar has the dates, and school holiday demand explains the pattern.
• Events come and go. A concert or a big game can pull a week forward. See event demand surges.
• Length of rental shifts. If hires get longer, the same bookings fill more hire days.

Flag these weeks in a notes column and adjust by hand. That's still better than ignoring them.

When the spreadsheet stops being enough

The method works well for a single depot with a handful of car types. It gets heavy when you have several depots, one-way hires, or more than about ten car types, because the curve per type gets thin.

That's when it's worth automating the export and the maths, so the forecast refreshes every morning without anyone touching it. It's work we do as part of revenue management, alongside the market rates from Rate Pulse. Read the demand forecasting term for the wider picture.

Questions