What Happens When You Build a Beer by Algorithm?
Remember Dinner Lab, the pop-up supper club that was aiming to build the world's first data-driven restaurant? Well, now someone has taken a similar approach to beer. A digital marketing agency called Havas Helia has created 0101, a beer built on data.
The tactics of the two projects are somewhat different. Where Dinner Lab held dinners all over the country and had diners fill out comment cards (which became the data they used to decide what worked and what didn't), Havas Helia looked for language on social media that embodied optimism, and compared that language with beer recipes and tasting notes. From PCWorld :
Late last year, Havas Helia's goal was to create a beer that embodied the optimism associated with beginning a new year. It began by tapping social media for emotional keywords used in New Year's–related posts and then scored "thousands of messages" against a set of 24 predefined emotional states. From there, the agency identified 38 distinct emotions with a clear tie to the new year. Among them were amazement, anticipation, generosity, indulgence, excitement, happiness, joy, love and good cheer.
Aiming to find a beer recipe that matched those emotions, Havas Helia used IBM's Watson Alchemy technology to analyze 2,800 recipes based on ingredients, tasting notes and reviews. Watson Personality Insights then helped profile each recipe as "assertive," "friendly," "intelligent" or a host of other possible qualities. From that analysis, the agency compiled a list of the top 10 beers that matched the New Year's emotions it had identified. By analyzing their recipes for common ingredients, the company came up with the recipe for 0101.
The resulting beer, brewed through a partnership with U.K.-based High Peak Brewing, is "a fine tasting cream ale." For now it's not widely available — only 500 bottles were made. But if you'd like to taste it, you can go to the website and click the link to "share their story" on social media. It's not clear how they'll be giving away bottles based on Twitter posts, but it's worth a try.
As a food-world trend, the "built by data" thing is an interesting idea. The cynic in me felt as though what Dinner Lab was likely to end up with was lowest common denominator, something akin to a restaurant built by Yelp. But this project, which is more based on trying to analyze positive emotions, might turn out something more interesting. Still, at a time when most craft beer lovers are still obsessed with bitter and sour flavors, a cream ale made with honey hardly seems on trend. Maybe our emotions don't actually know what our tastebuds want? Maybe that's why a single human chef or brewmaster is better than our collective unconscious put through an algorithm? Only time will tell.
Get the Food & Drink Newsletter
Our weekly guide to Los Angeles dining includes food news and reviews, as well as dining events and interviews with chefs and restaurant owners.