Hunkered down in a cinder block room, binders and papers scattered about the table, the general manager and his scouts are trying to put together a baseball team. Things aren’t going well. The team has lost some key players, including star Jason Giambi, and the scouts and GM are divided on how to replace them.
Collectively, the scouts represent decades of experience in baseball. They know the game, or so they think. “You have a lot of experience and wisdom in this room,” one scout tells the GM. “You need to have a little faith and let us do the job of replacing Giambi.”
The GM, though, wants to try a different approach to evaluate players, one based not so much on the intuition of scouts as on the power of data. His team, the Oakland A’s, is one of the poorest in baseball. It can’t spend tons of money on salaries. Instead, the GM wants to rely on statistics to uncover solid players who have been overlooked. “Guys, you’re just talking,” he tells the scouts. “Talking … like this is business as usual. It’s not.”
This battle between the grizzled scouts and the numbers-crunching GM, between gut instinct and hard data, is a scene from Moneyball, a movie that’s evocative of an age brimming with information on seemingly everyone and everything. These are the days of Big Data, and in the world of business, whether one is running a baseball team or a retail outlet, the collection and analysis of information is crucial. Tom Davenport, the President’s Distinguished Professor of Information Technology, thinks this movie scene so well illustrates how revolutionary data analysis can be that he and a copresenter have acted it out as part of presentations. Playing a scout, Davenport even has worn a costume, donning an A’s jersey and a baseball cap that proved a bit small. “It didn’t fit,” he says. “I wore it tilted on my head.”
The co-founder and research director of the International Institute for Analytics, a research firm, Davenport is an expert on Big Data. A term that can be hard to define, Big Data basically refers to the unprecedented volume, variety, and velocity of data produced today and the efforts to utilize and make sense of it all. Davenport, however, doesn’t dwell on the size of this onslaught of information, the gigabytes upon gigabytes pouring in from smartphones, social media, Web searches, surveillance cameras, and sensors on everything from appliances to our own bodies. “It makes for good cocktail party conversation,” he says. “The more important thing is what you do with the data.”
Barack Obama used data to target specific groups of voters and win the 2012 presidential election. In the spying controversy swirling around the National Security Agency, the government monitored phone records and Internet history to thwart potential terrorist threats. In business, the process of analyzing data has gone by different names, says Davenport, including decision support, business intelligence, and now analytics. It is by no means a new idea, but the potential influence data analysis can have on businesses is immense. While it needs to be handled with care given privacy concerns, Big Data can help target customers, expand the role of employees, and give CEOs more insight to make decisions, assuming they can let go of their egos and, unlike those baseball scouts, remain open to what data can tell them.
“I would be hard pressed to think of an industry that wouldn’t be affected by Big Data,” Davenport says. “Any industry that moves things, has employees, has customers—that amounts to every industry.” That kind of impact can’t be ignored.
Sifting through the Mounds
In the business world, the amount of data produced and examined is a game changer, says Sal Parise, associate professor of information systems. Before, say in a retail store, analysts looked at transactions. “A customer bought something and it was recorded, and then you did the analysis,” says Parise. But now, beyond the cash register, companies can look at such details as tweets, Facebook comments, YouTube videos, and Pinterest pictures.
The amount of available data for analysis will only intensify as the tsunami of information produced continues to swell. One area of growth will be sensors, says Dan Vesset ’94, program vice president, business analytics and Big Data, at IDC, a Framingham, Mass.-based market research and advisory firm. Expect to see sensors placed in cars, airplanes, delivery trucks, medical devices, appliances, and our bodies (think wearable devices such as Google Glass), creating a world even more interconnected and dense with data. A refrigerator, for instance, will no longer be just a refrigerator. It’ll be a data generator. It might know you’re out of food and contact the supermarket to have some groceries delivered. Or the refrigerator might suggest recipes based on what food you have. Or it might alert the manufacturer if it needs preventive maintenance. “I think we’re only scratching the surface with sensors,” Vesset says.
Bala Iyer, professor of information technology management and the William D. Bygrave Term Chair, says: “If someone in business from the 1960s were around today, they would say, ‘Oh my God, you have all this information.’”
But unlike the rows and columns of clean numbers that fit nicely into traditional spreadsheets and databases, much of this data is unstructured. Text, not to mention videos, pictures, and podcasts, aren’t easily measured. How does one quantify or qualify the 140 characters of a tweet? “It’s hard to do,” Parise says. “What do people mean by a certain word? It’s not a hard science.”
On some level, technology can help, says Vesset. If a company wants to examine feedback on social media, software exists that can ascertain if a text is positive or negative. Other programs can examine a tweet and assess the influence it might have by the strength of the sender’s network. The stronger the network, the sooner the company knows it should reply. Companies also can find outside help from service providers that are popping up. Give them access to the information, Vesset says, and they’ll do the work for you. In India, Iyer visited a bunch of vendors ramping up capacity to offer analytics services to the U.S. and Europe.
The endgame, of course, is to discover better ways to reach customers. Jesse Weissman, MBA’05, spent three years as manager, digital strategy and analytics at Staples, the giant office supply retailer based in Framingham. As Weissman describes it, his responsibility was to “deliver the right offer to the right customer at the right time on Staples.com.” With customers’ shopping history acting as a guide, Weissman used analytics to promote products on the site. If the data suggested a person might want to purchase coffee, for instance, then that product was promoted somewhere on-screen without any prompting from the customer. “It’s figuring out innate customer preferences,” says Weissman, who recently started a new job working on data strategy at Harvard Business Publishing. “If we do our job properly, it’s not jarring. It’s seamless. It resonates.”
Using Big Data, companies are analyzing customer behavior in unprecedented ways, says Vesset. How does the weather affect customer response rates to phone calls? Do customers click on a red or green button more on a website? Where in a store do customers tend to walk? All these things and more can be tracked.
With so many ways to communicate and collect data from customers, companies can constantly check in and create relationships with them, says Parise. As an example, he cites a person interested in buying sneakers. From the moment the potential customer starts researching shoes on the Web, that person should be on the radar of the sneaker company. When he goes to the store, the company should send him promotions over his smartphone to close the deal. When he wears the sneakers for a run, he ideally should have the company’s app downloaded so he can chart his workouts and further build affinity with the brand. Ultimately, this ongoing contact allows the brand to become an ingrained part of the customer’s life. “That’s the Holy Grail,” Parise says. “You’re having a conversation with the customer.”
Big Data seemingly puts a lot of power in the hands of businesses, but it also gives customers a louder voice and, as a result, more sway over public perception of brands. Normally, a company controls its brand, buying ads and deciding on a marketing strategy. Now customers shape brands themselves via myriad social media sites. They post comments about products and make videos. “There’s always been word of mouth. Now it’s word of mouth on steroids,” Parise says. “You can’t control it. The customer has a voice now.”
The smart companies realize this and are listening and acting. They may find the major influencers on social media and give them product demos to receive early feedback, or they may reward a customer for promoting a product on his or her social networks. “You’ve got to turn a customer into an advocate,” Parise says. “People listen to their peers more than the brand.”
Beyond empowering customers, Big Data also changes the workplace dynamic by empowering employees, Iyer says. The data gives them knowledge. If they have access to it, they can become active rather than passive partners in the business, making informed recommendations about decisions. The clear divide that once existed in the workplace between those who are analysts and those who aren’t is dissolving, agrees Weissman. “We are now all expected to be analysts,” he says. “It’s the democratization of data. It puts a lot of responsibility on everybody to be aware and diligent.”
Some business leaders are embracing these changes, Iyer says, and some are not. Leaders who have made the transition don’t want employees to say “I think,” says Iyer. “They want you to say, ‘I know because of the data.’” In some circles, leaders who cling to their autonomy whatever the data may say are dismissively labeled HiPPOs, or highest paid person’s opinions.
The leader who uses her intuition together with whatever data is available, Davenport says, will be the most effective. He co-teaches an MBA course on decision-making, and he notices how confident the students seem. “We all think we’re good decisionmakers,” Davenport says. “I think a lot of students come in trusting their gut. We tell them to distrust their gut a bit.”
Data needs to be collected, analyzed, and passed on to leaders who should take it to heart. “The data is only good,” Parise says, “if you make better decisions from it.”
A Big Responsibility
As is often the case, however, with increased power comes increased responsibility. “You can dazzle people with data,” says Davenport, but he warns that numbers shouldn’t be used to prove a pre-existing point of view. If CEOs turn to data for insight, they should seek the truth that’s hidden in the facts and figures, even if it’s bad news.
Iyer raises another concern: information disparity. Companies collecting and analyzing data have a vast advantage over those that can’t. Big Data not only helps businesses reach customers more efficiently but also can lead to more innovation. “We need people to have access to information,” Iyer says. “If you are not participating in that conversation, you are falling behind.”
Jeff Takle, MBA’08, has seen how a lack of information plays out in South Sudan, in an area near the Congo border with no infrastructure, no electricity, and no roads for miles around. Takle is director of innovations at Abt Associates, a social policy consulting firm in Cambridge, Mass., that has provided policy recommendations to some 60 nations around the world. Takle has looked at how Big Data can track the spread of drug-resistant tuberculosis in Nigeria, and he came to South Sudan to help local farmers better sell their crops. With few information sources, the farmers don’t know the cost of goods beyond 10 miles, so vendors in pickup trucks frequently will drive to them and offer unfair prices.
On first blush, these African farmers may not seem to have anything to do with Big Data. But their predicament illustrates how disadvantageous an information disparity can be. Think of Google or Amazon or Walmart or other owners of huge data depositories. What kind of immeasurable advantages do these companies have over smaller, poorer enterprises, and how does this widening gap play out in the years to come? “How can any startup in Zimbabwe compete with Coca-Cola?” Takle asks. “The people who can aggregate the data and crunch it are in a disproportionate position of power.”
Privacy is obviously another area of concern. A company that knows a lot about its customers, for instance, should be careful about how it presents them with products. “We can do customer segmentation like we have never known before,” says Weissman, who strives to remember what the customer is thinking. “I want to put my customer’s experience first. You don’t want to come off as creepy.” Vesset recalls the now-infamous story of how, based on the items a teenager was buying, a well-known retailer knew she was pregnant even before her father. Furious, the father complained to the company after coupons for baby items addressed to his daughter arrived at the house, only to discover later that his daughter indeed was expecting.
While companies make privacy missteps, they keep pushing the boundaries of what is acceptable. Davenport is waiting for a backlash, for people to demand en masse that their privacy be protected, but it still hasn’t come. “I thought the NSA thing might trigger it,” he says, “but I don’t think it will.” People might say they are concerned about privacy, but ultimately their actions suggest otherwise. “People are willing to reveal private info for discounts and free goods,” Davenport says. “It’s a trade-off people are willing to make.” In the future, the trade-offs are sure to continue. Now, for instance, people might not like receiving targeted ads to their smartphones. “That feels quite creepy,” Davenport says. “Ten years from now, that might seem normal.”