Article

Domino's digs into GPUs' data-churning mega-muscle to transform customer experience

Domino's has discovered a virtual goldmine in the data-handling computer power of a resource once confined to the picture-heavy needs of video gaming — the GPU.

Photo: Provided

January 14, 2020 by S.A. Whitehead — Food Editor, Net World Media Group

Domino's has been putting computing and AI mega-muscle into answering one very simple, but incredibly important question for hungry customers: When's my pizza ready?

Santa Clara, California-based AI and graphics processing unit business, Nvidia, helped answer that question and many more for the giant pizza chain by using GPUs — or graphics processing units — and their extraordinary computing power to accomplish some equally extraordinary business feats for the world's No. 1 pizza restaurant brand by sales. 

Nvidia and Domino's were partners last year during the 2019 Super Bowl when Domino's launched its wildly popular Points for Pies application that awarded customers points toward free pizza when they sent in images of pies or parts thereof through the brand's app.

Now, as simple as that may sound, it is anything but simple to accomplish technically. For instance, Nvidia said Points for Pies required more than 5,000 images and a peak level of computing strength. But, as most pizza and marketing professionals would probably agree, the huge push in free media Domino's got from the game was worth the expense the brand put toward paying for that computing feat. 

Now the two are back together this week with Domino's using Nvidia's GPU mega-muscle to improve the pizza brand's communication with customers about order status from the previous 75% prediction accuracy rateto 95% accuracy. It accomplished those exact predictions using an avalanche of Domino's data. 

Nvidia told Pizza Marketplace that Domino's used its GPU resources to calculate in things like "how many managers and employees are working, the number and complexity of orders in the pipeline and current traffic conditions" to improve the order-ready accuracy rate. The GPUs provide the computing muscle to expediently employ complex algorithms and churn out results from these many data sources. 

"Domino's does a very good job cataloging data in the stores, but until recently we lacked the hardware to build such a large model," Domino's Data Science and AI Manager Zack Fragoso said in a description of the project supplied to Pizza Marketplace by Nvidia. "Once we had our DGX server, we could train an even more complicated model. … That let us iterate very quickly, adding new data and improving the model."

Nvidia said it will next turn its attention to using similar technology to help with predictions around all the brand's real-time tasks, including an exploration of computer vision applications, inside and outside stores, to make the pizza carryout experience faster and easier for customers. 
NVIDIA GPUs have also helped power retail giant, Alibaba, to $38 billion in revenue on the world's largest single shopping event, known as Singles Day.

Tomorrow on Pizza Marketplace, Nvidia General Manager of AI for Retail Azita Martin answered some basic questions restaurateurs might have concerning the use of AI and GPUs to improve business. 

About S.A. Whitehead

---

Pizza Marketplace and QSRweb editor Shelly Whitehead is a former newspaper and TV reporter with an affinity for telling stories about the people and innovative thinking behind great brands.