Cover Story: Are You Ready for Big Data?
Direct marketers have been managing consumer data for decades. But as technology has increased the ways consumers interact with brands, marketers are confronting the prospect of "Big Data"—data sets that have grown so large and complex that they've become increasingly difficult to analyze, but offer greater rewards and competitive advantage for those who do.
"Big Data is a problem that many firms are facing today—data are just piling up everywhere," says Rio Longacre, a consultant at Capgemini Consulting in New York City. "The issue is, now that they have all this data, how do they combine it together, make sense out of it, and what do they do with it?"
Big Data can be overwhelming for marketers trying to understand just who their customers really are and when they are likely to purchase a particular product, says Tim Suther, chief marketing and strategy officer for the marketing and technology services firm Acxiom in Little Rock, Ark. It's a daunting task, and most organizations don't have a handle on it yet.
"A good portion of marketing is ineffective—of the nearly $500 billion spent on marketing worldwide, about $200 billion is suboptimally invested," Suther says. "We're helping organizations understand the impact of what they do on the ultimate measurement—did someone buy it or not?"
1. What's the first task for a marketer trying to leverage Big Data?
Answer: c. Determine the exact marketing questions that need to be answered by the data.
"The main advice for dealing with the Big Data problem: Companies need to envision what they want to get out of the information," Longacre says.
Mark Michaels, senior vice president of strategic development at the San Francisco Federal Credit Union, wants to know which of the credit union's members were sending money from their checking accounts to brokerages such as Charles Schwab so he can market the credit union's own investment products to those members. Michaels and his team also want to learn which customers are sending money monthly to mortgage services to offer them refinancing at lower rates.
To do this, the institution will begin offering members an online personal financial management tool through Geezeo Interactive in July. This will allow the institution to analyze the transaction data of not only all the deposit, loan and investment accounts those customers have with that institution, but also all the transaction data they might have with other institutions, brokerages and credit card companies.
"We already have a marketing information database which gives us various tools to run queries," Michaels says, "but that's not as important as the data going into it."
2. What's the next task for a marketer trying to leverage Big Data?
Answer: d. Figure out how to eliminate all the "noise" in the data to find the "signals" that will actually answer the questions.
"While we do have a lot of data on consumers out there that we can take advantage of, there is still a lot of garbage that is not very useful for firms trying to do target marketing," says Shawndra Hill, assistant professor at the Wharton School of the University of Pennsylvania in Philadelphia.
Hill and her team recently conducted research linking TV advertisements during the Super Bowl to what people were saying online. She says, "What we found was that, when people were talking about commercials, mostly they were talking about the creativity or celebrity in the ad, as opposed to the brand or product itself." Still, she says there were "signals" in the data for market research purposes. Some ads targeted specific demographics, such as women. As a result, the response from those demographic categories varied.
In another study, Hill and her team were able to link social media strategies used on a TV reality singing show to the amount of engagement on social media and sales of products associated with the show.
"We are very excited about this work because we can quantify the impact of different social TV strategies the networks are using on viewer engagement and sentiment in real time," Hill says.
Data also needs to be consistently collected, says Geoff Wolf, executive vice president of client strategy at the direct marketing agency J. Schmid & Associates in Mission, Kan.
For example, Wolf describes two customer service representatives who ask "How did you hear of us?" A customer answers, "I saw your ad in a magazine, but don't remember which one." The first representative puts down "unknown." The second picks a random magazine code to show it was a magazine.
"Now there's inconsistency with the data," says Wolf, "even though both sources were actually 'known.'"
Another aspect of the noise-to-signal ratio Suther points out is that companies need to find the most valuable customers in the data and allocate marketing dollars accordingly.
3. What can you learn about customers from Big Data?
Answer: a. Which customers and prospects offer the best marketing ROI
"For some organizations," explains Wolf, "a top customer can be worth 5, 10 or 15 times as much as the average customer."
Slavi Samardzija, executive vice president of marketing intelligence at the Richardson, Texas-based marketing solutions provider KBM Group, says marketers can spot these top customers by identifying "champions," customers actively engaged with the brand online.
Fairytale Brownies' champions are medical and law practices, firms that rely heavily on referral business and hence try to sweeten their relationships with gifts. David Kravetz, cofounder and catalog- and Web-team leader of the Phoenix-based online and catalog food gift-giving firm, says the IT staff zeroed in on the company's champions by mining the firm's database for top customers, cross-referencing against standard industrial classification (SIC) codes to determine the type of firm. As a result, Fairytale Brownies now targets doctors and lawyers because they are such a lucrative niche.
Kravetz and his team are also mining the most loyal customers in the database to send them handwritten thank you notes. Fairytale Brownies competes with Harry & David, Mrs. Fields and Godiva, and he says "Many customers jump around looking for the best prices, so we want to compete on loyalty, product quality and making personal connections."
Most of Fairytale Brownies' sales have come from branded search terms, including the company's name and common misspellings—such as Fairy Brownies and Fairyland Brownies—Kravetz explains. He's been tinkering with the terms he is using in the firm's pay-per-click ad campaign on Google, and is fine-turning the non-branded search terms—such as "brownie gifts" and "cookie delivery"—to increase the number of new customer leads from that category.
Part of that entails excluding certain searches by people who Kravetz says, in reality, weren't interested in mail-order gifts.
"When people added cities to searches, such as 'cookie delivery Sarasota,' none of that was working for us, because these people were actually looking for a local place that would deliver cookies bought on the same day," he explains. "I added negative keywords to the campaign, so now if a person puts in a city name, our ad won't show to them. This excludes the noise from the signal, because these campaigns can get really expensive."
4. What does your "Master Data Management" strategy need to encompass?
Answer: d. All of the above
Fred Hagerman, senior vice president and chief marketing officer at Firstmark Credit Union in San Antonio, says that comprehensive data management is a "journey."
"If I just had to worry about a marketing campaign, we'd be very focused on just building a good marketing database, which allows us to do a good email or direct mail campaign," Hagerman says. "But from a higher perspective, we are also responsible for talking to members at the right place at the right time, which means bringing in more data that we have to share. So if members telephone the call center to ask their balance, we'll know to talk to them about an auto loan."
Comprehensive data management can also aid in the allocation of marketing dollars to the most appropriate channels, Wolf says, depending on how many channels a consumer visited and which one was last accessed before buying.
In addition to traditional marketing analytical tools, Firstmark is considering whether to invest in a predictive analytical tool by SAS. But first, Hagerman and his team are making sure the tool can access all of the databases: the core banking platform, which tracks checking and savings account balances and transaction histories; the lending platform, which tracks application and approval histories, annual salaries and credit scores; the credit card database, which tracks transactional histories, credit limits and available credit; and outside information from Geezeo's tool.
However, "machines" can only do so much, says Hagerman, and his team also plans to compare what the predictive software spits out with their "blue book" of past campaign metrics.
"We are fortunate in that we have a fairly compact marketing group, where we can all get together at the same table and figure out what the next step should be."
Longacre recommends marketers first get buy-in for an analytical project from their chief information officer, but hold off involving IT developers until a marketing strategy has been developed.
"Keep in mind, it's IT who will do the actual integration work," he says, "but this cannot start until the planning stages have been completed."
Part of that integration is making the data interaction happen in real time, so marketers can use the insights across channels.
"Many companies say they are multichannel marketers, but the problem is they are doing multichannel marketing one channel at a time," says Tom Young, KBM's executive vice president of client services. "All of the channels need to be synchronized in real time, so that appropriate and relevant messages can be delivered to consumers."
That's Firstmark's goal, Hagerman says. "The good thing about getting data from all the platforms is that we'll be able to see members who were declined for an auto loan and avoid marketing them a credit card."
5. Can Big Data help shape your big picture strategy?
Answer: a. Yes
While developing tactical campaigns is "absolutely" important, Samardzija says marketers should really be leveraging Big Data to drive innovation.
His firm recently analyzed unstructured data online to understand the emotions and behavior of people trying to lose weight, and found that they tended to play more Internet games and surf the Web for fun.
"The moment when they were dealing with weight issues also became the moment for escapism," he says. "This type of insight is strategic in nature and can impact the entire go-to market strategy for a given brand."
Samardzija says marketers "absolutely need to do day-to-day optimization—tactical efforts. But you also need to take Big Data and extract truly emotional insights that can lead to product innovation and new positioning in the marketplace."
Katie Kuehner-Hebert is a California-based journalist who has written for American Banker, HRO Today, Benefits Selling and many other business publications. >