Tag Archives: Adam Galinsky

When the Boston Red Sox reversed their curse in 2004 by vanquishing the Yankees and going on to take the World Series, many fans and pundits were quick to give much of the credit to management’s decision to enlist sophisticated computerized analyses of player performance data to make staffing decisions. This year, the team’s all-too-familiar collapse left these same observers wondering how the numbers could have led the Red Sox astray. Einstein kept a sign in his office that read, “Not everything that counts can be counted, and not everything that can be counted counts.” Baseball fans are not the only ones being forced to consider that the best decisions aren’t necessarily the ones based on analyzing reams of data. Companies of all sorts are setting themselves up for the same hard lesson, thanks to the growing excitement over technology’s ability to place all manner of salient data at the fingertips of managers. Armed with so-called dashboard displays on their PCs, CEOs can effortlessly summon up a cornucopia of performance indicators. Call up this week’s sales by product line, throw in profit by region, cost per widget produced, and change in inventory levels. Compare it all with last week, with the same week last year, and with forecasts and goals. And there you have it: a comprehensive guide to the organization’s performance and to what you need to do to improve it. If only it were that easy. The fact is, this sort of data worship can provide a distorted, misleading view of what’s going on, and it can lead to flat-out bad decisions. Part of the problem is that some of the most important inputs aren’t easily quantifiable. (Einstein used to keep a sign in his office that read, “Not everything that counts can be counted, and not everything that can be counted counts.”) Adam Galinsky, an associate professor at Northwestern University’s Kellogg School of Management who studies decision making, notes that a bias toward hard data causes organizations real harm. “Things that are hard to quantify tend to get left out of the decision-making process,” he says. For business owners, there are all sorts of intangible or hard-to-quantify factors that can mean the difference between life or death: employee morale, the emergence of new technologies, changes in the competitive landscape, evolution in customer tastes. Think of executives at Ford (NYSE:F) and GM (NYSE:GM) endlessly twiddling with the easy-to-track impact of rebates, advertising, and health care costs, while remaining oblivious to the cultural and political forces–forces that helped trigger higher gas prices–that spelled doom for their obsession with SUVs. This isn’t a new problem. Managers have always had a misplaced confidence in numbers. But the ability to do more and more with the data in an increasingly high-tech fashion is making things worse. Indeed, the availability of slick new data-crunching tools, and the hype they’re receiving, leads executives to rely on them more. Good managers know the nonnumerical stuff matters, of course, but it sure must feel as if you’re being highly effective when you can dash off a few e-mail notes and then in a few days or even hours watch those real-time graphs leap skyward on your screen. And, of course, companies can’t help but treasure these data tools, given that they’ve spent a fortune investing in them, egged on by consultants and other experts. Thus, the CEO risks becoming Nero, fiddling with his keyboard, as it were, while the company burns. The problem goes beyond neglecting the intangibles. Sometimes the data that goes into a dashboard is incomplete or biased in subtle but significant ways that can lead a manager astray. Jorge Grau, the CIO at cell-phone-tower operator SBA Communications (NASDAQ:SBAC) in Boca Raton, Florida, helped build a system to place virtually any sort of real-time report in the hands of managers. But there have been occasions when managers, delighted by the sheer number of facts at their fingertips, were oblivious to problems with the data. “We’ve been bitten,” Grau says. In several cases, managers were making decisions based on summary performance for the entire company, not realizing that those numbers did not include data from a few key regions. When data is misread this way, it can lead to bad calls on anything from budgeting to promotions. “Reports can do more harm than good,” Grau says, which is why he now cautions managers about accepting a slick table of figures at face value. Even when the data is complete, there are plenty of ways to misinterpret it. One temptation for managers is to assume that when two sets of numbers go up and down together, changing one will lead to a change in the other–inspiring a CEO, for example, to order up a series of costly facility upgrades simply because the last two facility investments happened to be followed by sales spikes. “You can find correlations between the most improbable things, if you look for them,” says Luca Rigotti, a decision sciences researcher at Duke University’s Fuqua School of Business. “You’ll just end up doing silly stuff.” In fact, there’s a range of potentially costly psychological tricks that managers can inflict on themselves in the face of perfectly solid data, says Northwestern’s Galinsky. For example, the same set of numbers can provoke wildly different decisions depending on the order in which the numbers are presented. That’s because people tend to jump to conclusions based on the first numbers they hear, and they don’t allow the later figures to change their minds. The same numbers can also have different impacts depending on how they are framed–for example, stating that 10 percent of the customer base will defect if a product is changed, as opposed to stating that 90 percent will stay. “Managers need to protect themselves against falling for those sorts of biases,” says Galinsky. Even if you avoid false correlations and biases, you’re still in danger of getting smacked by the law of unintended consequences when you try to manage by dashboard, says Christopher Zappe, who studies decision sciences at Bucknell University. “People tend to respond to falling numbers with a knee-jerk reaction to try to halt the fall,” Zappe says. “But it often ends up aggravating the situation because they often don’t really understand the ways in which the different things going on at an organization affect each other.” Thus, the manager who sees the on-screen sales graph start to droop might order sales reps to increase the number of prospects they contact, not realizing that sales are falling because the reps are already stretched too thin. Or a jump in inventory might lead to a decision to slow production, even though much of the inventory is set to go out the door. None of this is to say that dashboards are the enemy here. Indeed, CEOs probably get into more trouble when they fail to bring enough data into the decision-making process. And companies that sell dashboard-related systems are themselves quick to point out some of the pitfalls of data-driven decision making. “The concept of data quality is a very unsexy part of our business,” says Kendall Collins, vice president of product marketing for Salesforce.com (NYSE:CRM). “But without it you’re not going to have the ability to make decisions or react in an intelligent way.” Dashboards also can be modified to deal with the problems. Howard Diamond, CEO of ePartners, an Irving, Texas, consultancy that installs information systems that often include executive dashboards, notes that even intangible aspects of a business can be brought into a dashboard if companies are clever. For example, part of the Cool Cuts 4 Kids chain of hairstyling shops, an ePartners client, wanted to track trends in customer satisfaction. The company’s system will chart the number of customers who fail to return within 10 weeks. But part of the key to success also has to be to retain a certain amount of healthy skepticism about the value of the information we get from our computers, no matter how slick or costly the system. That, and not trading Babe Ruth to the Yankees. Contributing editor David H. Freedman (whatsnext@inc.com) is a Boston-based author of several books about business and technology. His latest, A Perfect Mess, co-authored with Eric Abrahamson, will be published by Little, Brown in January.