In today's intensely competitive and fast-changing marketplace, companies can no longer rely on gut instinct, guesswork or "business as usual." Across all industries, businesses are turning to data analytics to quickly and accurately respond to and even predict buyer behavior in their quest to grow revenue while securing customer loyalty.
The desire to engage with customers more effectively is fueled in part by what many see as a shift in power from sellers to buyers, thanks to social media and the rise of mobile computing. In IBM's most recent Global CEO Study, in fact, more than 70% of CEOs said they were seeking a better understanding of individual customer needs and improved responsiveness to those desires. And according to IDC research, the global market for business analytics software grew 14% in 2011, compared with 11.6% the year before, and is slated for 9.8% compound annual growth between now and 2016.
Here is a look at two companies that are striving to capture the loyalty of their customers through the use of analytics.
T-Mobile: Combatting Customer Churn
For wireless providers, customer churn can be a killer. According to research from Strategy Analytics, at the end of 2011, the percentage of mobile customers who switch service providers every year reached 44%, its highest level ever.
T-Mobile is one carrier that has been feeling that pain. Dwarfed by AT&T and Verizon Wireless in market share, the company was losing one customer for every customer it gained in early 2012, according to a statement by former CEO Philipp Humm earlier this year. To offset that trend, T-Mobile is digging into its customer data to better understand buyer behavior and more precisely target customer needs.
"Customers have so many dynamic options right now," says Alison Bessho, director of IT enterprise systems at T-Mobile. "They can easily get intrigued by something new with a different company, so in order to keep them happy, we're always looking for creative ways to give them something new and different."
To that end, T-Mobile uses a Teradata database and analysis tools from SAS to collect and analyze customer data, including current plan rates, the number of family plans versus individual plans, credit ratings, network usage metrics and statistics comparing the amount of talking time and the amount of texting time. It then segments the customer base, builds focused campaigns for different customer profiles and presents offerings via its various sales channels, including stores, call centers and websites.
The marketing team then analyzes how customers respond to these campaigns to project financial returns and fine-tune the offers. To do that, it feeds data into the Hana real-time data analytics appliance from SAP, which uses in-memory computing to perform rapid analytics on large data sets. This allows statistics modelers and business analysts to query the data and -- if they find something unexpected -- query further, without involving IT.
"You don't have to pre-think what types of analytics you're going to do or pre-build the aggregation tables that you build with traditional BI solutions," Bessho says. Plus, the data can be loaded more quickly into the appliance than it can with traditional analytics platforms, and the queries run 55 times faster than with a traditional database. That speed encourages analysts to explore creatively, she says.
"A lot of the benefit is finding the unknown," Bessho says. "They get a surprising result, and they want to drill down into the data in ways they never anticipated. So it's important that the tool is responsive and cuts through rows of data quickly."
Analysts can now determine the types of campaigns that work best for various customer groups. "We now know how to go to different customers with [different] offers," Bessho says. For instance, one way to segment customers is by how close they are to the end of their contracts. Knowing this -- as well as what type of plans they have, what their credit scores are, and where they live -- T-Mobile can, for example, send phone upgrade offers to long-term customers and offers for different rate plans to newer ones.
These offers can go out via text message, email, the call center or physical stores. "When the customer is on the phone or walks in the store, we get more fresh data about them to help reps select the best offer at that specific time," Bessho says. "We can take advantage of historical data, as well as dynamic data, to create personalized, focused offers based on customer trends and behaviors."
T-Mobile also uses tools from Business Objects to produce dashboards and detailed operational reports for marketing leaders. It will soon launch a mobile business-intelligence capability so marketing execs can view the current performance of marketing campaigns on their tablets.
T-Mobile still faces challenges, including the need to recover from its failed buyout deal with AT&T and the June departure of its CEO. But the company is betting on customer insights to bolster its future prospects. It plans to add 300 more customer data attributes to the system to deepen and broaden its analytic capabilities, and it will add input from social media as well. In the first quarter of 2012, T-Mobile saw 187,000 net customer additions, compared with 99,000 net customer losses in the first quarter of 2011. "Our goal is to reduce churn, enhance loyalty, upsell and cross-sell new devices and rate plans, and make customers happier, while achieving better financial results," Bessho says.
SuperValu: A New Approach to Loyalty
For grocers, the concept of "loyalty" has historically been tied to the "loyalty card" -- those ubiquitous laminated cards that give shoppers automatic discounts. But market forces are driving grocers like SuperValu to kick their customer loyalty games up a notch. According to Wesley Story, group vice president of consumer insights and loyalty at SuperValu, competition is heating up, especially as more types of retailers -- from big-box stores to discounters -- add grocery items to their shelves. About two years ago, SuperValu launched an effort to become more customer-centric by creating a hassle-free shopping environment, offering more freshly prepared foods and matching product lineups to local tastes.
Customer data gathered from loyalty cards is key to this strategy, Story says, because it reveals buying trends and demographic shifts. "If you're not careful, all of a sudden the customer that was your target no longer lives around you," he says.
According to a study by RIS News and IDC Retail Insights, localizing merchandise and personalizing interactions has pushed business intelligence and analytics -- in particular, in-memory, data appliance and grid computing capabilities -- to the top of the priority list for grocers concerned about customer loyalty.
SuperValu has long used a Teradata data warehouse and traditional BI tools to analyze transaction and customer data. But it recently set up a big data analytics lab to accommodate faster, more complex, ad hoc queries against all types of data, including unstructured data from social media. The lab's tools include Teradata's Aster appliance, which collects data from operational systems and puts it in a nonproduction database optimized for analysis; Hadoop, an open-source analytics platform that uses parallel processing to quickly analyze large data volumes; and a visualization tool from Tableau Software designed to rapidly deploy dashboards that mash up various types of data, including information from external sources.
With this setup, SuperValu no longer needs to know how data will be structured or what questions it needs to ask. "If a query doesn't work, we can just throw it away because the investment is minimal versus weeks and months of development," Story says.
The grocer is already better able to keep popular items in stock by studying out-of-stock data from its inventory management system, peak shopping times from its transaction data, staffing levels from the labor management system and customer perceptions from its "voice of the customer" system. It has determined that certain stores needed to add a midday restocking shift to accommodate the rush of traffic between 4 p.m. and 6 p.m. "Some of this is Retail 101. But before, we didn't know exactly what the staffing levels needed to be at what stores or what the customer perception levels were," Story says.
Analytics also enables SuperValu to engage with customers through the most effective medium, be it email, text messaging, mobile apps or social media, Story says. The old-school approach was to ask customers which channel they prefer; however, it's far more accurate to watch their behavior, he says. So, for a highly digital customer, you increase activity where they respond the most -- maybe text and social media -- and drop it in the media where they're less active, like email and snail mail.
Predictive analytics is the next step, Story explains. The grocer is experimenting with segmenting customers and predicting their behavior by overlaying loyalty-card data with demographic, psychographic, behavioral and economic information from external providers. By seeing, for instance, the effects of the recession on shopping patterns, SuperValu can better predict which customers will switch to lower-priced items during a downturn and proactively market store brands to them. The company is also reaching out to digitally savvy consumers via mobile apps and social media.
"That's the secret sauce," Story says. "Bringing it all together to understand what the redemptions are, how we offered them, through which vehicle, where they [were] redeemed, which [channels] customers are most active in -- and their social media influence if they are a highly connected consumer."
Case Study Overweis Dairy: Seeing through the customer's eyes
Gut-feel decisions are no longer enough for businesses today, even for a nearly 100-year-old, family-owned company like Oberweis Dairy. Based in North Aurora, Ill., Oberweis operates more than 40 ice cream/dairy stores, a wholesale distribution business and a home delivery business. In 2010, when the company needed to make some changes, it invested in a system from SAS to make sure its efforts would pay off.
So far, the system has helped Oberweis improve customer retention in its home delivery business and increase store profitability and service times, according to Bruce Bedford, vice president of marketing. "We're blessed with tremendous customers who are brand-loyal, but it's also because we maintain an emphasis on the highest-quality foods, listen to their needs and respond quickly," he says. "In that effort, analytics tools have been tremendous."
Oberweis turned to analytics when it discovered a customer attrition problem in its home delivery business. The company reaches out to customers through direct mail, door-to-door visits and the Internet. Bedford says that many customers who signed up for home delivery in response to direct mail campaigns and door-to-door visits canceled the service after 180 days, but that was not the case for those who responded to Internet campaigns. The Internet was the only channel through which the company did not offer a $100 discount in the form of free deliveries for six months. The marketing team hypothesized that attrition rates spiked at 180 days because the value of the free-delivery offer had been depleted at that point.
To counter this trend, Oberweis devised a new promotion that offered identical savings of $100 but through a yearlong reduced charge of 99 cents per delivery. After determining that the response rates for the two offers were the same, the company tested their respective effects on customer loyalty. The results were dramatic: Among customers who responded to the 99-cent offer, there was a 35% improvement in the retention rate at the nine-month mark, "which is worth millions of dollars in incremental revenue gain," Bedford says.
Analytics also enabled Oberweis to speed service in its stores. "Customers were getting up to the cashier and not knowing what they wanted to order," Bedford says. The culprit, the marketing team determined, was the menu board. "We never designed it with the intention of getting people through the line efficiently," he says.
So last fall, the marketing team came up with four designs that led customers through the decisions of ice cream serving size, flavor and cone type, and featured images of six popular sundaes. The designs also highlighted products with high profit margins. "We didn't want to guide someone toward a simple sundae or traditional ice cream cone instead of our waffle cone, which is an upsell," Bedford says.
Using SAS modeling, the company tested the designs in several stores. When the best one was rolled out, Oberweis saw an average profit increase of 3% on fountain purchases and an estimated 30% improvement in service time during peak hours. "It's good for the customer because it's an uncomplicated and quick experience, and we've been able to drive incremental profitability," Bedford says.
Through predictive analytics, Oberweis has also determined that store customers who intend to purchase just a bottle of milk are most receptive to offers of discounted quarts of ice cream. "Before, we had no idea that would be beneficial to do, but we saw a dramatic increase in quarts of ice cream sold when store staff was trained to offer a dollar discount," Bedford says. "The story was lying there in the data, and by combing through it with the right tools, we could draw it out."
- Mary Brandel
Brandel is a Computerworld contributing writer. You can contact her at firstname.lastname@example.org.
Read more about applications in Computerworld's Applications Topic Center.