Wine Matching: Tasty Uses for Algorithms

Caption: Richard Yau ’10 and Joe Laurendi ’10, MEng ’11 launched Bright Cellars.

After roommates Richard Yau ’10 and Joe Laurendi ’10, MEng ’11 earned their undergraduate degrees, they headed separate ways. Yau helped NutraClick, a vitamin and beauty products firm, grow to $65M in revenue during his time as director of business development. Joe Laurendi stayed for a master’s degree and prototyped a web-based securities trading platform at Broadway Technology and, after MIT, he joined a early-stage social networking startup.

In 2014, they came back together and cofounded a firm that grew out of their MIT experience.

“We knew we wanted to start a business together, as we have very complementary skill sets,” says Yau, the company CEO with MIT degrees in both management and music and arts. “We entered the 6.470 competition together and the 100k Business Plan competition. Our friendship was solidified when I forced him to come to my acapella events—which he enjoyed thoroughly.” Laurendi, company CTO, earned his degrees in mathematics and computer science and engineering.

Their startup, Bright Cellars, is an online wine club that uses a questionnaire and customer feedback to recommend wines, which can be shipped monthly to customers. In May, Bright Cellars graduated from the gener8tor program and relocated to Milwaukee, after the company has received early backing from venture fund CSA Partners.

Why wine? Like many other people, they were buying wines with little information beyond familiarity, price, label design, or ratings. They decided that a company that helped people discover and learn about new wines would meet a need.

"Millennials are ready for this,” says Yau. “Nearly all of our new members learned about us through social media or bloggers. I think we can do for wine what craft beer has done for the beer industry. We can help wine drinkers discover and learn about the world of wine. Bright Cellars is the fun, interactive way to experience new wine.”

And how does the technical end work? “Bright Cellars works like Pandora Radio,” he says. “After each shipment, our members can rate and review their wines. We use that information to power our machine-learning algorithm and get even closer matches over time. At the same time we are collecting a lot of data about consumer preferences and wine in general.”

To test the system, I took the online quiz—which includes questions such as what kind of chocolate do you prefer—and, indeed, the resulting choices sound delicious: a peachy chardonnay, a berry-toned Cote du Rhone, and a full-bodied Spanish Ribero Del Duero. I think they got my preferences nailed.