Recommendations in the Time of Digital Food Ordering
As a gourmet cookbook author, my mother always had the same question for restaurant servers: “Do you have any strong recommendations?”
My mother and I recently had a conversation about how that question can manifest differently in the world of digital food ordering. My comment to her was that the strong recommendations that the server offers are either the most popular menu items at the restaurant, dishes that the chef favors the most, or a meal that the server personally enjoys, so it doesn’t necessarily mean that you will get the things that you yourself most enjoy based on your own preferences. As an example, I eat a plant-based diet, and typically, when I ask about the specials without first sharing my preferences, I end up hearing a lot about the daily beef special that the chef has prepared.
Digital ordering can help improve this experience and better tailor offerings to restaurant guests. Digital ordering can enable the guest to provide both explicit preferences and can capture inferred preferences. Explicit preferences mean that the guest can share their preferences — for example, a plant-based diet, allergies, favorite orders or anything that can be selected in a digital ordering platform and saved under a guest’s profile.
Inferred preferences, however, is where this gets very powerful. Inferred preferences are what a guest has ordered — perhaps what they’ve ordered very frequently. They could also be more nuanced. Maybe the guest has removed a standard ingredient from a dish and substituted something in its place. Inferred preferences can provide great clues about what specific ingredient a guest may favor or dislike.
Together, the explicit preferences and inferred preferences can enable a digital ordering platform to provide guests a personal view of the menu, unlike any other guest’s view of the menu. This personalized experience provides “strong recommendations” to that guest, custom tailored to what they tend to eat and enjoy. Combining explicit and inferred preferences with big data trends based on orders from other guests enables digital ordering platforms to make informed recommendations each guest. This is called “collaborative filtering.”
Collaborative filtering is probably best understood in the common lexicon of Netflix. The streaming platform personalizes watch suggestions based on the movies and shows a viewer had previously enjoyed, the large data pool other viewers have enjoyed, and the commonalities between those shows. When you apply collaborative filtering to the world of ordering food, a guest could receive recommendations to try different menu items because other guests with similar explicit and inferred preferences enjoy said new dish. The theory here is that the guest that received the personalized recommendation may order enjoy the recommended dish, may very well enjoy said dish, and may ultimately come to crave it. In turn, this elevated, tailored service could drive a guest to ultimately return to that restaurant more often.
We can know who the guest is and what they’ve ordered. The platform could recognize that guest and present them with a personalized view of a restaurant’s menu, customized to the things that they “prefer” and the things guests with similar ordering preferences have liked. This is the benefit of a software platform that can harness a broader understanding of the consumer.
This all comes down to what restauranteur Danny Meyer likes to call “digital hospitality” — the ability to make a guest feel like you’re on their side. In the world of ordering food, technology can help its restaurant partners provide this kind of digital hospitality — even at fast food-speed and fast food-scale. We aim to provide improved recommendations whether a guest is dining in, taking out, or ordering delivery. That’s a powerful promise embedded within the shift toward digital ordering in the restaurant industry.
About the Author
Noah Glass is the founder and CEO of Olo.