Drizly Chugs Down Data to Drive Personalization
Drizly's team relies on data to bring the traditional liquor store to the digital age.
I went to college in Boston (go Terriers!), so it's only fitting that I returned to my old stomping grounds to cover a company that deals with the one thing that almost every college student comes in contact with: booze.
Meet Drizly: a software company that aims to marry the personalization power of Netflix with the speed and convenience of Uber to create the ultimate, on-demand alcohol delivery service. I spent the day at Drizly's headquarters to learn more about the startup and the people who go the extra mile to ensure that the company, unlike the cocktail ingredients they serve up, isn't on the rocks. Here's a summary of my day.
10:20 a.m. I arrive at Drizly's Boylston Street office a few minutes before my scheduled 10:30 a.m. arrival time. It's housed in an old stone building—not exactly the environment you'd envision for a modern tech startup. And the office itself is just as much of a mash-up. There's a suit of armor in the corner that was supposedly found in an office closet. It wears a Drizly hockey jersey and has a Christmas wreath around its neck; a Grumpy Cat stuffed animal sits where the knight's face would be. Next to the suit of armor is a bar built from old cubicles, and a mini-fridge with liquor bottles stocked on top.
Hey, you have to know the service you're providing, right?
All of the teams sit at their respective tables. It reminds me of a high school cafeteria: Just as how the jocks sit with the jocks and the nerds sit with the nerds, the design team sits at the design table and the marketing team sits at the marketing table. There's not a lot of intermingling.
I notice a sign on the wall that reads, “Accept no excuses; only results” with a bell underneath it that the team rings whenever there's a big win.
Ah, results-driven marketers, I thought, I like it.
Then, I'm whisked into a conference room where I meet Drizly SVP of Marketing Michael (“Mike”) DiLorenzo, Lauren Monk, senior director of marketing strategy and analytics, and Brett Doyle, senior manager of customer marketing and monetization strategies. I learn that Monk, a hard cider fan, has been with Drizly for a little more than a year—she came over with DiLorenzo from members-only flash sale site Rue La La. Jameson Irish Whiskey drinker Doyle, on the other hand, is only on his third day. He was brought in to build out Drizly's email strategy.
DiLorenzo, a Guinness lover, kicks off the conversation by giving me a crash course on Drizly, which he likens to a fax machine. He says the comparison makes sense because Drizly connects the supply with the demand—or in this case, the alcohol retailer with the consumer. “We're a technology company that sells a license to a retailer,” he explains. “Drizly never touches alcohol and Drizly never touches the money from an individual transaction.”
Here's how it works: When a liquor store partners with Drizly, that store connects its point of sale system to Drizly's technology. This gives Drizly near-real-time insight into that liquor store's inventory, DiLorenzo explains. Customers can then shop all the inventory that's available in their local area either via Drizly's website or mobile app, easily add items to their shopping cart, select a driver's tip, and check out. “We've tried to create an experience that is on-par with the other mobile lifestyle apps,” he says.
According to DiLorenzo, convenience is core to the company's appeal—and is especially attractive during bad weather. For instance, after it was forecast that the East Coast was going to be hit with a massive storm on January 26, 2015, Drizly experienced a 354% increase in order volume in Boston and a 117% increase in New York, compared to a typical Monday in January. The company also saw a 10 and 14% increase in dollars spent per order (versus an ordinary day in January) for each city, respectively. Still, there are a few improvements DiLorenzo wants to make to the delivery experience—mainly personalizing it for onsite and in-app shoppers.
Currently, Drizly's personalization capabilities are limited. The company can offer product recommendations for people based on collaborative filtering. A shopper putting Strongbow Hard Apple Ciders in his cart, for instance, will see that people who buy Strongbow also tend to buy Heineken Lager.
To help put the personalization pieces together, Drizly is implementing BlueConic's customer data platform. DiLorenzo reminds me that Doyle, Monk, and I are heading over to BlueConic's offices for a lunch meeting later that afternoon before he ducks into an 11 a.m. meeting.
11 a.m. Monk, Doyle, and I remain in the conference room to discuss how they're planning to enhance Drizly's email strategy.
Currently, Drizly can segment its emails by location, buyer versus non-buyer, and order value. The company runs national and local ad hoc campaigns (such as July 4th or Shark Week messages) and drip campaigns. I say that I downloaded the Drizly app the day before my arrival, so Monk gives me a rundown of the emails that I can expect to receive in the next couple of days.
The company sends a welcome email the day after I download the app with introductory information about Drizly, including a rundown of how the company works and what I can expect. If I haven't made a purchase by day three, I receive my first incentive. Now part of the non-buyer sequence, I'll receive another incentive from Drizly on Day 15 to try and lure me to buy. And if I still haven't purchased by Day 30, I'll receive an email from the CEO asking me to fill out a survey to provide feedback on topics such as why haven't I purchased yet. I'll also receive incentives on Days 60 and 90, assuming I still haven't budged.
If at any point I make a purchase I'm dropped from the non-buyer sequence and enter the buyer sequence, Monk continues. The day after my first purchase I receive an email from Drizly thanking me for my patronage. There are actually three versions of this email. The first is sent to me if I made a purchase on Drizly's website and invites me to download the app; the second is sent to me if I purchased via the app and spent more than $100; and the third version is sent to me if I purchased via the app and spent less than $100. The second and third emails both focus on referrals, Monk says, and invite me to take a survey. Splitting the versions up, she notes, helps Drizly determine whether the order was placed by a consumer (the company's target audience is 29-year-old males) or a corporate account. Identifying these corporate accounts is important, she adds, because corporate buyers tend to order more frequently and have higher average order values.
The last sequence is a lapse sequence in which customers receive incentives on the 15th, 30th, and 60th days if they haven't placed a second order.
“We're finding that it takes about three purchases for a user to become a loyal user,” Monk says.
On average, Drizly sees an 18 to 20% email open rate, she adds. Even so, Doyle is planning on testing Drizly's emails to optimize their engagement rates—something Monk says Drizly hadn't been able to do in the past with just her and DiLorenzo at the helm.
“We haven't done any sort of testing as of yet,” Doyle says. “A real key addition that I'm going to have is basically testing out anything and everything in our email program that needs to be tested so that we can optimize the program fully, even down to time of day, day of week, and where our following interacts with us the most.”
After our 30-minute discussion we head over to BlueConic's offices.
11:30 a.m. On the walk over, Doyle starts raving about sandwiches from local sandwich shop Sam LaGrassa's. He says that he dropped a hint to the BlueConic team the last time he visited their offices, and hopes that the company ordered them for our lunch meeting.
12 p.m. We arrive at Blue-Conic, where we meet with CMO Dan Gilmartin, Cory Munchbach, director of product marketing, and Todd Belcher, customer success director. Munchbach gives us a rundown of BlueConic's platform and how the company uses identifiers to build profiles for known and anonymous visitors, capture their explicit actions (like buys, referrals, and shares), derive implicit intent (spreading word of mouth, trying to make a purchase), and trigger interactions—all within real time.
After Munchbach's quick presentation, we reconvene in the kitchen where we find Sam LaGrassa's sandwiches on the counter. It looks like BlueConic is already profiling Doyle.
12:41 p.m. After a filling lunch break, Drizly's and BlueConic's teams get straight to work brainstorming how the alcohol delivery service can benefit from the platform. They discuss Drizly's business cases for using it, which include driving conversion, progressive profiling, and segmenting based on customer lifetime value. They also cover how Drizly can put these business cases into action; for example, by doing multivariate testing, sending cart abandonment emails, and populating customers' emails with their preferred beverage brands.
The ideas are flowing, and I get so excited that I even pitch my own idea. I suggest sending an email or a push notification to customers about an hour after their order arrives to see if they've ordered enough supplies or if they're running low and would like to reorder. I'll be interested to see if it Drizly adopts it.
1:45 p.m. As we walk back to the office, I ask Doyle and Monk how they think the meeting went. Both seem pleased with the outcome and say that it was one of the most collaborative meetings they've been to.
But, more important, Doyle wants to know whether I liked the Sam LaGrassa's sandwiches. I reply that they were pretty good—only because I didn't have the heart to tell him that I thought that they were, well, average (sorry!).
2:35 p.m. After returning to the office, Doyle and Monk saunter to their cool kids marketing table and I sit down with Casey Hogan, Drizly's marketing associate for content and social media.
One of Hogan's responsibilities includes writing articles for Drizly's blog. The gin-and-tonic drinker says she tries to write as much evergreen content as possible, such as how-to articles, history pieces, and recipes. “[The content] aims to get Drizly's name out there for all things liquor,” she says.
In addition to managing the blog, Hogan runs Drizly's social channels. She posts content for its Facebook, Google+, Instagram, Pinterest, and national Twitter account. She also works with a general manager based in each of Drizly's locations to ensure that the tweets for the city-based accounts are tailored to their locale.
But social media isn't the only way Drizly shows its social side. Recognizing that drinking alcohol is a social activity by nature, the company tries to reinforce this message through its other channels. For instance, DiLorenzo later tells me that Drizly includes vintage images of real family members sharing a drink in its emails and display ads.
3 p.m. Wanting to dive deeper into the company's data and personalization objectives, I sit down with Drizly COO Cory Rellas.
There's no denying that Drizly has access to top-shelf data—ranging from basic purchase, behavioral, and registration information to granular details such as which drinks are trending in each state.
“It's no secret that any time you use an app there's a lot of data collected,” says Rellas, a Miller Lite fan. “Data is part of the sexiness of a digital property. You can trace people; not only what they input—their name, delivery location, birthdate, a few other things—but also their path through a digital liquor store.”
But it's what Drizly does with its data that truly matters. Right now, Rellas is focusing on using Drizly's data to inform the consumer experience and create a personalized environment that surpasses any traditional liquor store. “[Traditional liquor stores] can't move shelves around; they can't show you a bottle that you didn't know you would want,” he says. “So, [we're] really taking all of that together.”
But could Drizly's trove of data actually produce more treasure for the company? To learn more, I chat with Phuc Truong.
4 p.m. As the SVP of monetization, Truong's responsibility is to look at Drizly's audiences, data, and infrastructure and recommend new revenue streams. Drizly already makes money by charging retail partners a monthly licensing fee. But its data holds the key to additional revenue.
For instance, alcohol producers (like Miller and Corona) can't sell directly to consumers in all 50 states due to state laws and regulations set in place by the three-tiered system. Therefore, Drizly can sell those companies non-PII data to help them learn more about their customers on a granular level.
“For the alcohol industry there's a big gap between data and how marketers can take advantage of that data to not only target advertising, but from a measurement perspective, too,” Truong says. “That's missing, primarily because [of] the distribution and how alcohol is sold varies so much state to state.”
In fact, Truong says, Drizly's first-party and market research data could position the company as the Nielsen or Comcast of the alcohol industry. But he admits that this isn't on Drizly's immediate roadmap, and says that the company's market research data is currently “directional at best.”
One revenue stream Drizly is leveraging, however, is its brand partnerships. The company offers many brand services, such as providing targeted media placements to reach specific demographics on social media. It can also create enhanced brand pages on its website. Not only do these specialty pages showcase a brand's products, but they also feature its content. For instance, Heineken's brand page features a YouTube video and live social feed above its product offerings.
In addition, Drizly can run brand activations surrounding specific events. Drizly recently worked with Miller Lite as its commerce partner, for example, to help the beer brand reach football fans during weekend game broadcasts. Truong says Miller Lite had purchased ads through Twitter Cards inviting sports fans to order beer. Consumers who clicked on the ads were immediately taken to a microsite where, using Drizly's technology, they could shop and buy only Miller Lite.
“There aren't many places…where a brand advertiser can get their message across… [and] tie a media placement to an actual purchase as it relates to alcohol,” Truong says, “because it's not legal to do so.”
4:30 p.m. As the day comes to a close, DiLorenzo is starting to feel the jet lag from his Las Vegas red-eye flight that landed this morning. So, he decides to end the day early and invites me to join him and the rest of the team for a drink. I politely decline, but enjoy watching as the team scurries off to take advantage of Thirsty Thursday.
I check my phone and see a few missed texts from my senior year roommate. I'm crashing at her apartment for the evening to indulge in a mini-reunion. She's agreed to make me dinner—gnocchi a la vodka, with shrimp, no less—and says that she has to run to the store to pick up the vodka for the sauce.
I wonder if she's ever heard of Drizly.