It’s Hard to Change People’s Behavior

Rocket science retailing requires people to change their thinking and behavior, and doing that is always challenging and time consuming. Managers often underestimate the changes needed when adopting a new technology. Consider, for example, the experience of a consumer electronics retailer that had invested millions of dollars in purchasing and installing a software package for forecasting and inventory planning. About a year after the installation, company experts were overriding more than 80 percent of the software’s recommendations. (Incidentally, the software was intended to support the decisions of those same experts.) The company had done a good job of installing the software, but it had failed to convince its staffers that they should trust the results more often or to investigate why the software was generating forecasts and stocking quantities that needed to be changed so often. Thus the company’s attempts to make its forecasting and planning processes more analytical failed. This retailer, like many others, focused its effort on the technical aspects of improving forecasting and inventory planning while underinvesting in the managerial aspects pertaining to changing people’s behavior and mind-set. An alternate approach—that would have probably been more beneficial—would be to gradually build confidence among the staffers through pilots and a phase where the staffers’ overrides of the system’s recommendations, and the ensuing impact on forecast errors and inventory levels, were tracked. Over time, such steps would have led to a better algorithm and greater confidence among the human experts, who would now have been free to direct their effort to more challenging tasks.

A History of Exaggerated Benefit Promises Has Created Skepticism

Many managers and employees are jaded when it comes to new analytical approaches, having heard pitches in the past from software vendors and consultants that promised more than they delivered. At many companies, skepticism trumps even a well-documented pilot or proven results at another retailer.

New Skills Are Needed

Retailers may not have the right people to sustain the rocket science effort. Traditionally, retailers have emphasized product and industry knowledge over analytical skill when hiring. Thus their employees tend to rely on their instincts and judgment over data and analysis in making decisions. At times, organizational culture even explicitly discourages analysis. In an extreme case, a former executive at a large fashion retailer told us how a CEO, upon seeing her crunching numbers to forecast demand, grabbed the sheaf of papers on her desk and flung it across the room. As she tells the tale, the CEO was concerned that his organization would become overly dependent on the analysis and lose its feel for its market. Not surprisingly, analytical techniques still are shunned at this retailer.

Broad Organizational Changes May Be Required

Some executives complain that using rocket science techniques will require substantial changes in other parts of their companies. They say that they do not have good operational execution in their stores or distribution centers. “How do we apply analytical techniques when our data quality is so poor?” some of them have asked us. At retailers with long response times for replenishing products, we’ve heard arguments that the inability to respond to early sales data removes the primary benefit of being able to discern trends from it. Ironically, at one of these retailers (where merchants questioned the value of early sales data because their supply chain could not react soon enough), the logistics department argued to us that reducing lead times was not very useful because their merchants did not have the capability to discern trends and optimize purchase quantity from early sales data. This seemed like a question of which function—merchandising or logistics—was going to give in first.

Cross-Functional Collaboration Is Required

A senior executive, upon reading a draft of our chapter, noted that many of the suggestions in the book would require cross-functional projects. For example, improving supply chain responsiveness would require close collaboration among buyers and the supply chain and would possibly involve changes in the factory as well. Similarly, improving store execution would require coordination among store managers and human resources. Retailers are often reluctant to launch such cross-functional projects, presumably because such projects can be tough to implement. This executive noted that “the tendency in most retailers is to work on tactical, functional improvements.” Moreover, he noted that “the effort spent on a collection of these [tactical, functional] projects often eliminates organizational capacity for the kinds of broader changes” needed for rocket science retailing.

Analytics May Take the “Fun” Out of Retailing

Finally, some executives have even noted that rocket science retailing, with its emphasis on process discipline, data analysis, and mathematical approaches, could take the fun out of retailing. One executive at a fashion firm noted that people joined his company because they loved the catwalk, not because they loved crunching numbers.

Notwithstanding these doubts, complaints, and obstacles, many retailers have adopted at least some rocket science practices, and some of them have sustained large rocket science projects effectively. In other words, we’ve seen considerable variation in the rate of adoption. In fact, the most interesting question might be how some firms manage to implement rocket science when so many others seem to stumble with it. This chapter offers insights that we have gleaned from the successful cases that we have observed and participated in.

Recognize That There Are Multiple Customers, and Cater to All of Them

Projects that have successfully introduced analytic approaches have recognized the need to create champions at multiple levels. I2 Technologies, a provider of decision support software, embodied this principle in the late 1990s. I2 recognized that it needed to sell its software not only to “C-level executives” (who usually had the budgetary authority to approve purchases) but also to the staffers who would regularly use the software. The company’s attention to creating champions lower down in companies was and is relatively rare.

I2 created and sold a product to serve schedulers in factories. Schedulers at most factories determine which tasks should be performed by specific machines and operators at various points in time. Their decisions in large part determine not only the factory’s capacity utilization and inventory levels but also the level of service offered to specific customers. Yet in many factories, senior managers were unaware of the centrality of schedulers. I2’s founder, Sanjiv Sidhu, related to us one of the first sales calls that he and cofounder Ken Sharma made at The Timken Company, a steelmaker. Sharma asked the executive they were meeting with what it would take to get a $10 million investment at Timken. The manager had responded that it would take a meeting of the company’s board of directors. Sharma then pointed out that the production schedulers made $10 million investments in inventory every day with little oversight from senior management and few tools to support decision making. Not surprisingly, many schedulers felt they did not get much respect from colleagues. I2 recognized the schedulers’ key role and understood that they could be crucial allies when implementing advanced analytic approaches to improve factory planning.

Consequently, when implementing its “factory planner,” i2 emphasized the importance of the work done by schedulers and the tools they needed to do their jobs better. Equally important, i2 pushed to increase schedulers’ visibility within and outside their organizations. One industry analyst told us, “i2 created rock stars out of these schedulers.”² Following a successful implementation, i2 would invite the schedulers to make presentations to other companies and at i2’s annual conference, Planet, which by the late 1990s was attracting three thousand participants a year. As a consequence of the greater visibility, many schedulers were able to negotiate better wages or switch to other companies. Not surprisingly, the schedulers remained loyal to i2. Planet became an i2 lovefest or, as one analyst characterized it, “Woodstock for middle-aged i2 users.” In fact, as a consequence of the loyalty that i2 inspired among them, schedulers often offered suggestions to i2’s engineers to identify ways in which the software could be improved.

Why do managers seeking to implement rocket science retailing have to create champions at all levels? As with the United Nations Security Council, multiple stakeholders have effective veto power over rocket science retailing projects because the adoption of rocket science retailing tends to be discretionary. Lower-level staffers can stymie a project even after senior managers have decided to purchase a particular piece of software. It’s not uncommon, for example, to see companies where multiple software packages have been purchased and installed but never used. Lower-level staffers cannot necessarily decide whether to buy software but they can decide whether to use it.

What’s more, implementations of rocket science retailing usually require drawing upon knowledge that already exists in the firm (and they almost always perform better if they do). In one of our early projects at a skiwear manufacturer, users pointed out to us the vital role of production minimums in their supply chain. Without incorporating these minimums, our approach would have been useless to the manufacturer. When users get excited about an application, they’re more likely to offer up these sorts of nuggets of wisdom. Hence, catering to all levels in the company—not just senior executives—turns out to be critical during implementation.

Quantify the Benefits—Using Agreed-upon Assumptions—But Recognize That Precise Quantification Is Often Impossible

Consider our experience at the first two retailers where we worked on improving aspects of operational execution (such as inventory record inaccuracy and misplaced SKUs) in the stores. The two projects started more or less concurrently.

At one retailer, we estimated the annual benefits (in dollars) that could accrue to the company from improving store execution. At the second retailer, where the opportunity was clearly much bigger, we resisted offering an estimate. We instead pointed out to the top managers that the opportunity was clearly huge and that quantifying the benefit was unnecessary. The first retailer launched its project to improve execution soon after our analysis, while the second one failed to take any serious steps for more than a year. The two retailers certainly differed in many ways, and thus you’d expect them to behave differently. Even so, we believe that our failure to quantify the opportunity at the second retailer caused the project to languish. Soon after we quantified the benefits there, it gained traction.

Put simply, you have to quantify the potential benefits of a project if you want it to succeed. I2 Technologies was passionate about quantifying benefits and even hired an independent auditing firm to interview past clients and collect data on performance improvements brought about by its projects.

Recognize, however, that quantifying benefits usually requires making assumptions that can be challenged. The supply chain vice president at a high-end jewelry retailer, in arguing for supply chain improvements, estimated the benefits from reducing stockouts in the company’s stores. To perform his analysis, he had to assume that stockouts led to lost sales, and his colleagues challenged him, arguing that consumers at this high-end retailer often substituted another product when faced with a stockout. The manager then redid his analysis, assuming various levels of consumer substitution and showing that the savings still could be substantial.

Why is it important for missionaries to quantify the benefits of rocket science retailing? For one thing, doing so ensures that they get recognition for their efforts. Often, once the benefits of rocket science retailing are realized, other people—including some who were originally opposed—are likely to want to claim credit for the benefit. As the old saying goes, failure is an orphan, but success has a thousand fathers.

Kevin Freeland, one of the architects of Best Buy’s turnaround and a believer in analytical approaches to retailing, advises missionaries to “own a benefit” with senior management.³ While Freeland worked at Best Buy, the company engaged a consultant to measure the impact of various activities within the firm, such as increasing inventory turns. According to the consultant’s analysis, Freeland and his team had generated close to $1 billion in value.

Win Quickly! Quick Wins Matter More Than Big Wins

A few years ago during a case discussion at an executive program, a participant asked us how to identify the projects with the highest return on investment. Given the political realities in many organizations, he should’ve been worrying about how to identify the projects with the quickest return on investment. Quick wins create and sustain enthusiasm for change.

Missionaries might consider the following framework in choosing “quick win” projects. You categorize potential projects along two dimensions: the potential benefit and the degree of technical and organizational difficulty. Low-difficulty/high-potential-benefit projects are most desirable but typically hard to find. High-difficulty/low-potential-benefit projects should be avoided. Managers then have a choice between low-difficulty/low-potential-benefit projects or high-difficulty/ high-potential-benefit projects. Low-difficulty/low-potential-benefit projects usually offer quicker wins, so managers should start with those before transitioning to the more challenging and rewarding projects (e.g., high-difficulty/high-potential-benefit projects).

Small wins can attract the resources needed for larger projects. Freeland says managers should test and experiment and “fail fast and quietly” so that they can refine their processes and create success stories.

Why do quick wins matter? Every rocket science retailing project that we have witnessed has had its share of naysayers and doubters. Doubters are typically willing to change their minds when they can see evidence of success. However, if that evidence fails to appear relatively quickly, they are likely to become naysayers as well. And naysayers rarely change their minds.

Don’t Be Sexy

When seeking a test case, choose the unglamorous tasks that other staffers do not like to do. In most fashion firms, for example, merchants have strong views on style and enjoy choosing cuts and colors for clothing. In contrast, they often find the task of choosing sizes tiresome. Thus missionaries seeking to implement rocket science retailing in this sector should apply their techniques to size optimization first. That way, they’ll face less resistance and can more easily score quick wins.

If you start off by messing with style, you’re more likely to set off a political squabble. Consequently, you might fail to get your test case, and even if you do get it, coworkers with valuable insider knowledge might not want to help you. On the other hand, if you take an unglamorous task, you may win support because the knowledgeable folks might welcome your help. What’s more, these sorts of tasks may not have attracted talented people or substantial resources. Consequently, the potential to generate savings is substantial. As i2 discovered with factory scheduling, you might find low-hanging fruit. Finally, if you fail, not as many people will notice. In Kevin Freeland’s words, you’ll fail quietly.

Structural Changes to Sustain Rocket Science Retailing

Organizational structure plays a vital role in sustaining rocket science retailing. In altering their organizations, retailers have found ways to allocate tasks to people with the appropriate skills and access to appropriate information.

Many retailers, for example, create separate departments for buying and planning within the merchandising function. A buyer and a planner might form a team to make merchandising plans for a specific product category (say, women’s swimsuits). Buyers typically handle duties such as spotting the next major trend, identifying the appropriate supplier for a product, and negotiating the best terms with the supplier. They tend to focus on longer-term decisions (for example, for the following season or year). Planners, on the other hand, tend to focus on the short term and make more detailed decisions, such as how many units of a particular SKU to carry at a store. Not surprisingly, buyers tend to rely more on intuition, while planners tend to stress detailed analysis of historical sales data.

The buyer-planner division recognizes that good merchandising requires a combination of big-picture creativity and detailed analysis. And it reflects the understanding that few people are capable of excelling at both.

Retailers also design their organizational structures to reflect the differences between top-down and bottom-up forecasting. Old Navy, a division of Gap, for example, has a company planner, who does a top-down forecast from economic factors and corporate growth goals. Merchandising teams, on the other hand, develop bottom-up forecasts based on their evaluation of individual SKUs. The approaches yield different forecasts, and managers from both groups meet to reconcile them.

Retailers have also sought to inject the top-down perspective through the budgeting process. Footwear seller Nine West, like many chains in the United States, required its “retail directors” (who combined the roles of buyers and planners) to seek to maximize gross margins and sales while operating within a budget.⁴ Retail directors had considerable leeway to determine how much of each SKU to carry at each store, but the budget for each director for each season was set cooperatively by that director and a division president. In these meetings, the director brought the bottom-up perspective, and the president offered the top-down view.

Cultural Changes to Sustain Rocket Science Retailing

Sustained improvement in operational execution requires a company to attend to its culture with active support from the CEO and other senior executives.

Our favorite example of this comes not from retailing but from a distributor of electronic components called Arrow Electronics, headquartered in Melville, New York. Soon after Steve Kaufman joined Arrow in 1982, he realized that Arrow’s practice of having a distribution center for each sales region was inefficient.⁵ A few well-run warehouses could service all the sales regions. But Arrow’s sales representatives often needed to know whether parts were available at a distribution center before they took customer orders, so they had to have real-time visibility of on-hand inventory, and the inventory records had to be nearly 100 percent accurate.

Kaufman determined that organizational culture was undermining data accuracy. He decided that he needed to create a culture that cared about maintaining accurate records, so he launched a four-pronged push to change the culture. He modified the physical audit process in the warehouses to focus on operational results, not just financial metrics, and also made the results a subject of discussion at weekly senior staff meetings. He lauded the work of Betty Jane Scheihing, an Arrow employee with a “missionary’s passion” for accuracy. He held warehouse-floor workers accountable for the accuracy of their counts. And he took personal actions to sustain the culture that he was working to create. His effort was so thorough and successful that we’ll walk you through each step below.

Inventory Accuracy Metrics

When Kaufman joined Arrow, the company operated thirty-seven warehouses, which were physically audited once a year. To underscore the importance of these audits, Kaufman made their results a feature of his Tuesday morning staff meetings. He usually discussed results from a particular audit on the next Tuesday. During the meeting, Kaufman would highlight discrepancies between the company’s records and the auditors’ counts.

Each audit team reported the percentage of bins with miscounts of more than a specified tolerance, usually ten pieces or $10, whichever was less. Bins outside the tolerance were called “dings.” If a warehouse had less than 2 percent dings, Kaufman rated its performance on inventory record accuracy as good. If it had between 2 percent and 5 percent dings, it needed “tuning up.” If it had more than 5 percent dings, it had a “serious problem” and needed a team from the corporate office to come in and suggest improvements. And if a warehouse had more than 10 percent dings, Kaufman deemed it “out of control” and transferred or terminated the manager.

The key takeaway in Kaufman’s approach was not the specific rules for accuracy that he created. More important was the quickness and clarity of his communication. He and his management team examined the results of audits promptly, shared their view of acceptable performance, and made clear the consequences of falling short. Just as critical, Kaufman did not state his goals in financial terms alone: a warehouse could have been classified as a “serious problem” or “out of control” even if it had the appropriate value of inventory in dollars but many inaccurate inventory records.

Identify and Empower Champions

Kaufman also got lucky. Someone who shared his zeal for accuracy had arrived even earlier than he had. Betty Jane Scheihing came to Arrow as an inventory clerk in 1968 after graduating from Philadelphia Bible College. Back then, the company stored its inventory information in a manual Cardex system. Such was Scheihing’s passion for ensuring accuracy that she wore the title of “Cardex Girl” with pride. Scheihing retired in 2004 as senior vice president of global operations and human resources, making her one of the most senior officers at the company. Throughout her career, she remained an accuracy maven.

Once he arrived, Kaufman not only publicly cheered Scheihing’s efforts, he also ensured that she had the backing she needed from him and saw to it that her ideas were implemented. Scheihing, for example, insisted that salespeople and sales managers not be allowed to interfere in the operations of the warehouse and that even their presence could compromise inventory accuracy. Salespeople—typically to meet a client’s urgent request—would be tempted to remove a few units of a product from the warehouse without going through the necessary and associated paperwork. Consequently, the company redesigned processes to bar sales staff from distribution centers. Scheihing also believed that all movement of products within the warehouse should not only be tracked but also directed by the company’s inventory control systems. Thus a salesperson could not generate a pick ticket, even for a big customer with an urgent need, without entering an order into the system. Finally, Scheihing believed that managers who failed to maintain accurate data should be fired. She once shut down a warehouse in Texas because inventory record accuracy was too low.

Setting a Personal Example

As is often the case with successful leaders, Kaufman’s actions spoke louder than his words. He made frequent trips to distribution centers. At a center, he would meet with the operating manager and review the results of the last count. He even would personally check the inventory accuracy for a common part number every time he visited a warehouse anywhere in the world. He also wrote frequent e-mails to employees emphasizing the importance of accurate inventory. Staffers with primary responsibility for this metric often received e-mails multiple times a year, praising their accuracy levels or urging improvements. The message, Kaufman believed, was clear: “Inventory record accuracy is so important even the chairman attends to it.”

Empowering the Shop Floor

Arrow gave relatively low-level warehouse operators the responsibility for finding and correcting errors. This approach was relatively rare outside of the company. Every firm Arrow had acquired over the prior twenty years (more than sixty) had followed the more traditional approach of assuming that warehouse personnel could not be trusted with this responsibility, and thus had not provided them with the ability to make changes to records in the inventory management systems. Instead, the prior owners, Scheihing said, “assumed that only college-educated corporate office managers or finance and accounting types could be trusted with this responsibility.” In one company, all new hires in the management-training program rotated through the department that did inventory adjustments, on the assumption that their analytical and financial skills uniquely qualified them for this role.

Scheihing believed that this approach failed on several counts. One, the work itself—that is, fixing physical count errors—required attention to detail but wasn’t difficult. Two, the people closest to the errors had the best chance of spotting, correcting, and fixing them and the underlying causes. They also had the greatest motivation since the errors slowed their work. Three, managers and corporate-office staff would never be enthusiastic or consistently motivated by what they saw as grunt work, nor would they give it the needed priority and focus. Four, the people closest to the work would take pride in keeping their own houses in order rather than being treated as human extensions of the conveyors.

Whenever Arrow acquired a company, it disbanded the headquarters group doing inventory adjustments and trained the warehouse workers to do it. In every case, this improved accuracy as shown by the cycle count statistics.