Segmentation Series Part 1: Introduction To Alcohol Segmentation

In Distillery, Segmentation Series by Justin Robinson

No two alcohols are exactly alike, and neither are any two consumers. Of course, you can group alcohols together by identifying certain characteristics. And you can do the same with your consumers. So we did just that at Drizly. Well, our Data Scientists did anyway. Those guys are smart.

 According to BCG, personalizing the consumer experience generates a 6-10% lift in revenue. Once we broke our users into segments and began basing our CRM communication on what each group was most interested in, we saw a 14% increase in repeat orders versus the control group.

Segmentation, and by extension personalization, have become the industry standard in eCommerce and the alcohol industry needs to follow suit. In the first of this three-part series we’ll take you through the high-level effects of segmentation we’ve seen on sales at Drizly.

Introduction to Our Buyer segmentation

The first thing our Data Scientists made clear was you don’t just go out and create segments. They already exist. It’s a matter of whether or not you’re paying attention.

Segments find themselves. Jill, for example, might be a 31-year-old city dweller who loves setting up shop to work at her local coffee shop and buying from Trader Joe’s. But you can’t go searching for a bunch of people who look like Jill.

Instead, we let users cluster together naturally, based on similar buying patterns. This is what we found when we began identifying groups of users with shared order habits on Drizly:

But here’s the thing about segments: Just because Jill is a Wine Buff today doesn’t mean she will be next month. In fact, over the last 12 months we’ve seen that 26.8% of users move between segments, sometimes more than once.

Users change their behavior. Not necessarily the most surprising piece of information you’ll hear today, but an important thing to always keep in mind. Constantly marketing to Jill the exact same way, as if she’s consistently the same buyer, doesn’t line up with her reality and will quickly become ineffective.

Segmenting users also allowed us to investigate how valuable each group is to Drizly. There are incredibly nuanced and interesting ways to do so, but here we’ll take a look at the market share of each segment, by Orders and Sales.

What Segments Buy?

Since we broke down our segments by their buying preferences, there aren’t any surprises in their top-level preferences. The eye-opening information is understanding where the biggest cross-selling opportunities lie.

Beer Enthusiasts are much more likely to buy wine and liquor than Wine Buffs and Spirits Specializers are to buy beer. Office Organizers, on the other hand, are ordering from all three categories relatively evenly. Lastly, Spirits Specializers are buying ~50% more non alc products than any other segment. Not a surprise, but worth noting.

Going a step further, analyzing the subcategories allows us to get even more granular in our CRM efforts.

Wine Buffs and Wine Hosts like red wine, Seasoned Sippers prefer white, and Office Organizers love their champagne and sparkling. Wine Buffs might not drink a lot of liquor, but when they do pick up a bottle, they’re likely to choose vodka. Office Organizers and Beer Enthusiasts, on the other hand, have a shared affinity for whiskey. Office Organizers also way over-index on tequila. We’ve got a boss, too. We get it.

By going one step further, we can sharpen our marketing efforts by not only showing users the right subcategory, but the exact right product that they’re most likely to buy, based on their segment. Below are the top 20 products of two of them. Office Organizers really need their ice, while Seasoned Sippers spend a lot of their wallet on vodka and white wine.

Instead of blasting the same marketing materials to all Drizly users, we now surface the right products to the right people to guarantee they’re seeing exactly what they’re most interested. From the products they’re most likely to buy in their preferred category, to the sub-categories they’re most likely to be interested in, we can be confident that users are seeing what resonates most with them whenever they interact with Drizly.

Next month we’ll go further into our segmentation efforts and the benefits we’ve seen.

If you can’t wait ‘til then to start on your own segments (smart thinking), reach out to us anytime or shoot us an email at data@drizly.com.

Methodology
  • All data are representative of the trailing 12 months ending on March 13, 2018.
  • Orders are unique orders that may contain 1 to an infinite number of items.
  • Units are defined as a single bottle, 6 pack, 30 pack, box, etc.
  • Sales are in USD.