Milking the Data
Milk delivered to the doorstep in glass bottles. Hand-dipped milkshakes made from premium ice cream at company-owned stores. While the products that come from North Aurora, Ill.-based Oberweis Dairy’s network of family farms may seem a throwback to a bygone era, there is nothing old-fashioned about the way the company applies business analytics.
With three distinct lines of business — home delivery, retail outlets and wholesale distribution — Oberweis amasses “reams and reams of data, and more is generated every day,” says vice president of marketing Bruce Bedford. “The challenge is [converting] this information into programs that are actually generating profit, and doing that in a systematic way.”
The company engaged SAS and its business analytics tools — but before anything could be analyzed, the data had to be compiled.
Oberweis operates primarily in the upper Midwest, though home delivery extends into Virginia. (Ice cream is shipped nationally via its e-commerce site.) The lines of business all had data, but in different formats. The retail shops — which serve hand-dipped ice cream treats and sell products like milk, coffee and bacon — featured a loyalty program (the “Moola” card) managed by a third-party vendor. Home deliveries were keyed into an enterprise resource planning system. Not even the addresses were standardized.
SAS tools helped even out the data, allowing the company to “really understand which customers are shopping in our stores and how those behaviors differ from [home delivery] customers,” Bedford says.
“With three channels out there, our customers can easily migrate from one to another and cannibalization is a concern,” he says. “If we do something in one channel, are we converting loyalists in one channel to another?”
What the company found was that customers who regularly shop Oberweis’s stores or utilize its home delivery service “tend not to migrate to a wholesale channel, such as a traditional grocery store, for our brand,” Bedford says. “That doesn’t mean they don’t shop in those stores, but they don’t think about buying our brand there. The customers are segmented by channel in a way that we never thought was the case.” That knowledge eased concerns about expanding wholesale into areas where the company had existing retail outlets and home delivery.
Home delivery — which currently operates in urban centers in six states — offers dairy, coffee, meat, bread and eggs on a weekly basis. Building its customer base is done through door-to-door sales, direct marketing, Google ads and customer initiative.
“Churn is always a concern in the home delivery business,” Bedford says. “Analytics really began at Oberweis with a desire to understand how we could mitigate account attrition.”
With so much data on its home delivery customers — including years’ worth of transaction histories — there was enough material to work with. Bedford instituted “survival analysis” that looked at retention based on promotion and start method.
“As we learn what the most important retention drivers are and how to modify them, we make changes to the program,” he says. Some of the changes were prompted by analysis, some not, but the company has improved its retention rate by a factor of two, Bedford says. “The attraction and retention profiles are different depending on the methods that we’ve used to acquire a new customer.”
Oberweis quickly realized that the data would yield information beyond customer attrition in its home delivery area. It found ways to identify and solve manufacturing glitches by mining customer complaints; another tactic rooted out bottle return fraud at its dairy stores (the glass bottles require a deposit and redemption when returned to the store). It also found a way to simplify commissions for home delivery drivers.
“The high-tech solutions that are available for consumers now mean the tables have turned,” says Diana McHenry, SAS director of global retail product marketing. “As consumers we can know more about retailers than they know about us. Oberweis is taking a 360-degree view to understand their customer preferences and buyer behavior and help find anything that could be a customer satisfaction issue, then get to the root cause and improve it.”
Business analytics can also hone the distinctions within a customer profile, McHenry says. “There are nuances for Oberweis customers: Folks who buy a quart of ice cream because they like the fresh taste of Oberweis milk, [and] others may appreciate that it’s hormone-free. One person is a healthy green shopper and another is a gourmet. You really need to understand the customer to know the nuance.”
The company’s 48 dairy stores are something of an ice cream parlor/convenience store hybrid. A customer can purchase a scoop of ice cream, a milkshake or a sundae in the parlor, while the c-store area includes more than 200 SKUs ranging from Oberweis glass-bottled milk and packaged ice cream to bacon and cartons of eggs. All outlets have seating areas, and some include a drive-through.
“We want to understand what our customers purchase from behind the ‘dairy doors’ — the refrigerated areas in the dairy store,” Bedford says. “We want to know what items our customers purchase along with our packaged ice cream. Throughout our network we offer a buy-three-quarts/get-one-free promotion. What ice cream flavors are people purchasing together?
“To understand how our flavors are taken by store, by region, by season and promotion allows us to better manage our production and inventory,” he says.
It also allows the wholesale arm to offer its outlets better insight into customer preferences. “A lot of us felt that there were flavors that naturally went together,” Bedford says. “We discovered our assumptions were not entirely correct. It turns out there were certain combinations that we didn’t think people bought regularly. With wholesale, we might suggest that we bundle a certain number of quarts because our use of market basket analytics has demonstrated a preference by our consumers for certain unique flavor combinations.”
Gaining continual insight into its dairy stores is particularly important as Oberweis seeks to expand beyond its Midwest footprint. “We’re currently in the process of testing some new elements of the concept, and the plan is to have a national concept,” Bedford says.
Understanding the customer
Small or large, understanding the customer is “so vital,” McHenry says. “Oberweis has done it right by integrating data and using it across the whole company to improve customer satisfaction, to increase sales by targeted promotions.”
When Peter Oberweis founded the dairy in 1915, he began by selling extra milk to neighbors. He knew his customers individually, and fully understood their preferences. As the company has grown, that personal understanding has remained a priority.
“For a company the size of Oberweis, there is competition everywhere — the neighborhood grocery store, the mega-grocer and even the food you can buy on Amazon,” McHenry says. “With increased competition, a smaller or medium-sized business can sustain itself and continue to compete by bringing analytics into play. You need to look at things that work for your business. Maybe you can’t go head-to-head with Walmart on the lowest prices, but analytics can show you” how to present your customer base from eroding.
At Oberweis, the goal is to use analytics to grow. “We don’t intend to be this size forever,” Bedford says. “If we want to be a national firm, managing data and utilizing data is a substantial source of competitive advantage that we could leverage. Our large competitors do analytics regularly.
“To become a national concept, we’re going to have to take share from our competitors,” he says. “I don’t know how this would work if we don’t understand our customer as well as they do.”