Machine Learning Times
Machine Learning Times
EXCLUSIVE HIGHLIGHTS
Survey: Machine Learning Projects Still Routinely Fail to Deploy
 Originally published in KDnuggets. Eric Siegel highlights the chronic...
Three Best Practices for Unilever’s Global Analytics Initiatives
    This article from Morgan Vawter, Global Vice...
Getting Machine Learning Projects from Idea to Execution
 Originally published in Harvard Business Review Machine learning might...
Eric Siegel on Bloomberg Businessweek
  Listen to Eric Siegel, former Columbia University Professor,...
SHARE THIS:

7 years ago
Analytics make Bright Cellars the Pandora of wine clubs

 
Originally published in The Journal Sentinel, June 28, 2017

If the machine-learning algorithm gets it right, you’ll be matched with wines you love.

That’s the idea behind Bright Cellars, a Milwaukee-based monthly wine club begun by two Massachusetts Institute of Technology grads who certainly knew more about math than Merlot when they came up with their business plan.

“I didn’t have a background in wine,” said 28-year old Richard Yau, who co-founded the company in 2015 with former roommate Joe Laurendi. “I took some wine classes at Boston University and I had purchased wine from the supermarket — but wasn’t that comfortable with wine. I was more comfortable with craft beer.”

Bright Cellars, which now has 16,000 subscribers and 30 employees, landed in our state because in 2015 Yau and Laurent were accepted into the Gener8tor program, a Wisconsin-based accelerator that invests in high-growth startups.

The two tech entrepreneurs were provided with what Yau describes as “an intense three-month-long startup boot camp in Madison” and $70,000 in seed money.

“We graduated from Gener8tor and decided to set up shop in Milwaukee,” he said.

With Bright Cellars, you enter the world of predictive analytics by way of a seven-question online quiz that asks, for instance, “What is the one type of chocolate you could eat for the rest of your life?” The choice of answers: dark chocolate, milk chocolate, white chocolate, Reese’s cup, Snickers, and “where’s the fruity candy?”

Then you ante up $60 per month plus $8 shipping to receive four bottles of wine chosen according to characteristics gleaned from the quiz.

Your subsequent ratings of the wines “fine tune the understanding of what you like,” explains Yau, noting that “it’s characteristics such as sweetness, bitterness and acidity that come into play.”

The algorithm “learns” what you like in much the same way that your music preferences become known with Pandora Radio or your viewing likes are revealed on Netflix.

Initially, Yau and Laurendi sought wines that would match the various “data clusters” they identified. One of those clusters might be a group that likes aromatic whites and another cohort might like medium-bodied reds.

But starting in late 2016, they changed the equation a bit.

In what Yau calls “an ongoing project, an experiment,” they decided to “move from matching our members to wines that are available to matching the wine to our members, by way of the algorithm for our customer base.”

Today, about half their brands are custom-made to reflect their subscribers’ preferences. They are called “Bright Cellars exclusives — custom brands with custom labels,” said Yau, adding that the rest of the wines are from small producers, with some national brands in the mix.

Their total portfolio of wines is about 58, including both those from existing wineries and those that are custom-made. They hope to increase that number to more than 150 wines in the next five years, tweaking choices as they gather more data.

The process of creating their custom brands begins with bulk wines, which are fermented but not yet bottled. They contract with wine branding company AWDirect in Sonoma County, Calif., to tailor-make and bottle the wines, with many of the bottles decked out in labels designed at Bright Cellars.

Asked about their target audience, Yau indicated that the club is geared toward people “moving on from the Barefoot Wine world — people who are used to buying $7 to $12 bottles at the supermarket.” Each of the wines in their portfolio retails for about $15 a bottle.

Most potential subscribers learn about Bright Cellars through social media such as Facebook, Instagram and Twitter, as well as through blogs and podcasts. So it’s no surprise that the audience skews younger, with the core group being 30 to 35 years old.

For those who imagine all their interactions will be with a computer, Yau points out that Bright Cellars has a “concierge” team that handles about 300 to 400 member interactions a day, either by phone or email.

The staff also includes wine director Gretel Bubnik, a certified sommelier, who helps with wine selections and also handles staff training.

Subscriptions can be canceled at any time. And if a customer doesn’t like a bottle, their policy is to replace it in the following shipment.

For more information or to sign up, go to brightcellars.com.

About the Author:

Anne Schamberg is a freelance writer who lives in Waukesha. 

 

Leave a Reply