Sponsored by Amazon Web Services
Forecasting is a decision-making tool that helps businesses cope with the impact of the future’s uncertainty by examining data and trends. Actionable forecasts in the Restaurant industry bring particular value: how many guests will arrive and when, how to staff accordingly, how to adjust inventory positions, how to optimize pricing, and more.
Forecast too high and the business will be inefficient with resources: likely buying too much product, facing potential spoilage or waste, while also missing opportunities to invest capital elsewhere. Forecast too low and the company will have missed sales opportunities while decreasing customer experiences and satisfaction levels.
Despite unpredictability, accurate forecasting is critical for restaurants. There are three common traits of the most successful organizations. Those who excel at both the “Art and Science” of forecasting have:
- Resources, a mix of talent and experience aligned by a strong culture.
- Procedures that support the continuous improvement of the forecasting process.
- Technology that enables the capture of genuine patterns and relationships in data, while removing noise.
Better forecasting leads to less food waste
Founded in 2007, Sweetgreen is a destination for simple, seasonal, healthy food. They believe the choices we make about what we eat, where it comes from and how it’s prepared have a direct and powerful impact on the health of individuals, communities, and the environment. That’s why they are building a transparent supply network, why they cook from scratch, and why they are building a community of people who support real food.
Sweetgreen desired to bring data together from 31 separate sources to derive sustainability and ingredient production forecasts, but they did not have a dedicated data engineering team. The solution was Amazon EMR (Elastic Map Reduce) which enabled them to run Petabyte-scale analysis and to focus on driving business outcomes, instead of engineering.
“Our greatest insight was how much food to cook per day, per time period. It drives sustainability and minimizes waste while allowing us to present the best food possible, driving operational efficiencies,” said Gerard Bartolome, Principal Data Engineer.
Forecasting for every location, every five minutes
For the Grubhub team, a leading online and mobile food-ordering and delivery marketplace, having a good forecast is fundamental for getting deliveries completed on time. The team must ensure the right number of delivery drivers are on the road to meet diners’ and restaurants’ expectations.
Oversupply of drivers increases operating costs while undersupply decreases customer satisfaction.
In order to accurately forecast demand, the engineering and data science teams leverage a combination of forecasting models running atop Amazon S3, and Amazon Elastic Map Reduce (EMR.) Grubhub must execute its forecasting at an extraordinary level of spatial granularity, down to the individual neighborhood level. William Cox, Senior Data Scientist, states, “AWS enables our production systems to continuously regenerate forecasts for every location, every five minutes.”
Predicting orders, before they’re made
When it comes to the global pizza business, Domino’s Pizza Enterprises Limited (Domino’s) has a large slice of the pie. The company, which is the largest Domino’s franchise holder, represents the Domino’s brand in Australia, New Zealand, Belgium, France, the Netherlands, Japan, Germany, Luxembourg, and Denmark. Domino’s maintains a network of more than 2,600 stores globally and is based in Brisbane, Australia.
Domino’s is an increasingly digital business, with more than 70 percent of sales coming from online orders. To enable faster pickup and delivery, the company recently launched Project 3TEN, an initiative that aims to have a pizza ready for pickup within 3 minutes or safely delivered within 10. To support this initiative, the company wanted to use forecasting and predictive technologies to help reduce pizza making and delivery times. Leveraging Amazon S3 and Amazon SageMaker the company built and trained Machine Learning models to predict the likelihood that an order will be placed, so a store can begin making that order right before it is placed.
“This isn’t making pizzas and leaving them in a hot box for half an hour— this is getting the pizzas lined up, coming out of the oven, and ready to go as an order is placed. Customers are getting their pizza faster, hotter, and fresher because of the improvements we’ve put into place with Project 3TEN. The solution we developed by using AWS is a big part of that.” Said Michael Gillespie, Chief Digital and Technology Officer.
Click here for the full eBook that explores these three areas with the goal of assisting Hospitality companies improve their forecasting to better operate in the 21st century.
