Quick Serve Restaurant Boosts Accuracy of Forecasting Models with

The Customer

In an on-demand world, fast food companies play a vital role serving millions of customers a day as quickly as possible. The challenges that come with running one of the world’s largest fast food chains are vast, including inventory management (how much food to buy), food preparation (how to prepare food up to 15 minutes in advance), staff planning (having the right number of employees on site to meet demand), faster service and better logistics (for centrally distributed items). This customer, a global fast food chain, serves millions of burgers and other menu items at thousands of restaurants every day.

The Challenge

With the competitive landscape, the company is constantly running promotional campaigns and changing prices to drive customers into their stores. Other factors that throw curveballs into the process of forecasting daily sales include events and holidays, seasonality, weather, and variance in hours of operation due to weekday versus weekend traffic. If the company miscalculates anticipated demand, they risk not having enough food on site and a poor customer experience. Overestimate the daily traffic in stores and they could waste tens of thousands of dollars of product resulting in a direct hit to the bottom line.

To try and solve this challenge, the technology team decided to look at artificial intelligence solutions, specifically how deep learning can be leveraged to take the guesswork out of daily forecasting.

The Solution

While the burger chain employs technical staff members, they did not have data scientists on staff who could spend time designing and experimenting with various AI technologies for daily sales forecasting and staff planning.

When the team came across the platform, they were able to quickly upload sales and staffing data for the last two years to automatically build customized deep learning algorithms for forecasting. In a matter of hours, the model reached 90% accuracy for forecasting food product sales at 30-minute intervals and 95% accuracy for staffing. By fully automating the AI development process—from designing, training and deploying custom AI models—the company only needs to allocate one person to manage the process. From building and testing through deployment, the process took just a few weeks and the company now has the ability to deploy at scale for every restaurant around the world.

Key Benefits

  1. High accuracy: The custom deep learning model generated by had 90% accuracy on forecasting food and 95% accuracy around daily staffing needs.
  2. Save time and resources: No feature engineering required by the fast food company, and only one person is needed to manage the model development and deployment.
  3. Reduce wasted product and avoid over/understaffing: The burger chain deployed the new model in days, not months, allowing the company to quickly implement cost-savings by avoiding ordering product that won’t be needed. The fast food chain is now also able to predict the staff needed for any shift to ensure customer satisfaction while keeping labor costs in check.

Schedule a Demo

Comments are closed.