Skip to main content

The Scoop Behind IHOP's AI-Powered Recommendation Engine

Dine Brands' CIO Justin Skelton explained the impact of the in-house tech project at MURTEC Executive Summit. For its efforts, Dine Brands received the MURTEC Breakthrough Awards in the Enterprise category.
wolfe
Justin Skelton Dine Brands at MES 24
HT’s Vice President & Brand Direct Abigail Lorden, left, talked with Dine Brands' CIO Justin Skelton about how its in-house IT team built upon Google Cloud Recommendations AI to enhance IHOP’s online ordering experience.
Advertisement - article continues below
Advertisement

Attendees at MURTEC Executive Summit learned how one Dine Brands built an AI-powered recommendation engine in-house that has a cost-to-revenue ratio of $1 for every $60 in revenue.

The project earned Dine Brands the 2024 MURTEC Breakthrough Awards in the Enterprise category.

Tech Moves

Since 2019, Dine Brands has tripled its spending on technology, rolled out a new POS to 1,500+ IHOP locations, launched a new website and mobile app with a guest loyalty program, and more.

In a one-on-one fire chat with Hospitality Technology’s Vice President & Brand Director and Restaurant Technology Network Co-Founder, Abigail Lorden,  CIO Justin Skelton explained how IHOP built upon Google Cloud Recommendations AI to enhance IHOP’s online ordering experience nationwide.

Skelton shared that the AI recommendation engine was one of Dine’s first foray’s into leveraging AI at scale. 

The idea was born during the pandemic when the full-service restaurant had 40% of revenue coming from off-premises sales and needed to generate revenue. “We needed something. We needed something that wouldn’t cost a lot.  … We wanted to make sure it was something that we could do, and we could it quickly. It wouldn’t take 2,3,4 or 5 years to implement. We felt the ROI (on a recommendation engine) would be immediate.”

After analyzing different products, Dine settled on the Google Retail API platform. “We settled on the Google platform, but the amount of work we had to do to get it configured and get it deployed in production was significant,” explained Skelton.   Data cleansing, importing customer information from its 11 million loyalty members, and adding other information such as intersession transactions for non-loyalty customers, menu uploads and more.

“I want to underscore that although we went with the underlying platform, the engine, that all the work we had to do to pull everything together to make it work,” Skelton emphasized, adding that it took Dine’s IT team about 9 months to ready for deployment.

Skelton shared one of the key lessons learned:  how you clean and present data. “Data accuracy and quality is key. You have all of this data, but is it the right data,” and then supplementing wth intersession data.


Big Bang for the Buck

Dine’s IT team achieved its objective of generating revenue with a low tech spend. “One thing that is remarkable is for every $60 of revenue, costs $1.  When you think about recommendation engines, this is fantastic. We know that 75% of the people presented with recommendations, that individuals interact with the recommendation engine in some way, shape or form.  We know people are looking at it, and ultimately we know a good number of people are accepting the recommendation,” he said.

Looking to the near future, Dine plans to add aa recommendation engine to Applebee’s and to extend the recommendation engine to the server to use with dine-in guests, he explained.

“To the extent we can take those recommendations, extend them to the servers, where they actually have the recommendation based on the order that comes in, and they can suggest things based on that, we believe that there's a tremendous opportunity there,” Skelton said. 



Tech Tides

These are some of the highlights from one of the thought-provoking sessions from MURTEC Executive Summit: Tech Tides, Oct. 21-23 at the Fairmont Grand Del Mar in San Diego.  Stay tuned for more MURTEC Executive Summit key takeaways. 

X
This ad will auto-close in 10 seconds