Forecasting sales in industries like Quick Service Restaurants can be challenging with fluctuating customer demand throughout the day, from day to day, week to week and seasonally. Forecasting sales correctly is critical to inventory and food supply, and also for managing staffing levels to positively impact customer service. Sales forecasting for QSR does not have to be time-consuming or frustrating.
There are four key decisions to make when building a sales forecast. Forecasting begins with utilising past revenue over previous weeks or year to develop the forecast for the upcoming weeks.
- You need a historical source data. High-resolution sales data from your Point of Sales (POS) system is ideal. High resolution means you need a data source that contains your daily sales by 15min to 60min increments. Hourly data is adequate. 15 or 30min increments are even better.
- Decide which historical period best represents the up coming week? Does the rolling average of the last 2 weeks make sense? Or how about a similar time last year (eg. 52 weeks ago). If you choose to use a rolling average of x weeks. Many use 4 weeks, make sure you calculate the standard error. If your error is too high this makes your rolling average data useless.
Graph 1. (above) The blue line is hourly sales. The red line is hourly customer numbers. This graph shows a three-week rolling average of hourly sales and units – from two weeks ago and rolling average of previous three weeks. The Error bars (orange) show a very high variation at 10:30am, 12:30pm, 1:30pm and 4:30pm. The high error bars indicate that these times over the last three weeks have fluctuated quite a bit.
- Know your growth index. How much have your sales grown since the period of time your source data is from? If you use last year same week as your source data make sure you multiple the numbers by your growth index.
- Consider other factors. Weather, special events in the area. Is the upcoming week you are forecasting during school holidays? Does it contain a public holiday/bank holiday etc?
Once you have completed steps 1 to 4 your forecast should present the final projected hourly sales in such a way that you can easily apply the correct staffing levels.
In our next blog series, we will take you through how to match your staffing levels to meet customer demand in only 10 mins. So you will always meet your customer service levels as well as keep your labor metrics on target. If you can’t wait to find out more contact us at firstname.lastname@example.org or request a demo.