124 Harvard Business Review January–February 2012

124 Harvard Business Review January–February 2012 HBR.ORG

Zeynep Ton is a visiting assistant professor in the operations management group at MIT Sloan School of Management. PHOTOGRAPHY: STEVE WEBSTER PHOTOGRAPHY:

lmost one-fifth of American workers Some companies are have bad jobs. They endure low wages, poor bene ts, schedules that change investing in their workers with little—if any—notice, and few opportunities for advancement. and reaping healthy proﬁ ts. The conventional wisdom is that many companies have no choice but to o er bad jobs—especially re- by Zeynep Ton tailers whose business models entail competing on low prices. If retailers invest more in employees, customers will have to pay more, the assumption

January–February 2012 Harvard Business Review 125 WHY “GOOD JOBS” ARE GOOD FOR RETAILERS

goes. Indeed, it is easy to conclude that employee- retailers—that currently don’t invest adequately in friendly Wegmans and the Container Store can o er their workers could be part of the solution. great jobs only because their customers are willing to pay higher prices. Why Do Retailers I have studied retail operations for more than 10 Underinvest in Labor? years and have found that the presumed trade-o be- If investing in retail labor is such a good idea, as my tween investment in employees and low prices can be research suggests, why isn’t everybody doing it? The broken. Highly successful retail chains—such as Quik- main reason is that labor is often a retailer’s largest Trip convenience stores, Mercadona and Trader Joe’s controllable expense and can account for more than supermarkets, and Costco wholesale clubs—not only 10% of revenues—a considerable level in an indus- invest heavily in store employees but also have the try with low profit margins. In addition, many re- lowest prices in their industries, solid nancial perfor- tailers see labor as a cost driver rather than a sales mance, and better customer service than their com- driver and therefore focus on minimizing its costs. petitors. They have demonstrated that, even in the Accordingly, they often evaluate store managers lowest-price segment of retail, bad jobs are not a cost- on whether they meet monthly (or weekly) targets driven necessity but a choice. And they have proven for payroll as a percentage of sales. These manag- that the key to breaking the trade-o is a combination ers don’t have much control over sales (they almost of investment in the workforce and operational prac- never make decisions on merchandise mix, layout, tices that bene t employees, customers, and the com- price, or promotions), but they do have a fair amount pany. This article explains those practices. of control over payroll. So when sales decrease, they Although my research has focused on retailing, I immediately reduce sta ng levels. The pressure to believe that the model these retailers have created can reduce payroll expenses is so high that store manag- be applied in other service organizations where there ers at several large chains, including Walmart, have are large uctuations in customer tra c and employ- been widely reported to have forced employees to ees perform both production and customer service work off-the-clock, paying them for fewer hours tasks. These include hospitals, restaurants, banks, than they put in. and hotels. Moreover, the nancial bene ts of cutting em- The United States needs better jobs, not just ployees are direct, immediate, and easy to measure, more jobs. Service businesses—including low-price whereas the less-desirable e ects are indirect, long

TOO MANY U.S. RETAIL WORKERS HAVE BAD JOBS WAGES FULL TIME PART TIME In 2010, an At But there is no guarantee American cashier that store employees are % made an average of even that fortunate:

hours a week, that of retail workers are translates to only % $ part-timers. Their hourly . wages are 35% lower of retailers call anyone an hour, $ , than those of full-time a year. working more than employees. Moreover, they often do % not receive health beneﬁ ts hours and are scheduled too few a week a full-timer. below the average hours to earn a living. for all occupations. SOURCE FRANCOIS J. CARRÉ, CHRIS TILLY, AND LAUREN D. APPLEBAUM

126 Harvard Business Review January–February 2012 HBR.ORG

Idea in Brief Retailers have long believed that the only But a growing body of research— OFFER FEWER SKUs and promo- way to compete on price is to off er work- including the author’s studies of tions in order to reduce complexity. ers low wages, poor benefi ts, constantly Borders, Home Depot, QuikTrip CROSS TRAIN WORKERS so that changing schedules, and little opportu- convenience stores, Mercadona they can perform multiple tasks nity for advancement. and Trader Joe’s supermarkets, and instead of varying the number of Costco wholesale clubs—suggests employees to match changes in that there is an alternative to provid- customer traffi c. ing “bad jobs.” ELIMINATE WASTE everywhere Retailers can break the trade-off except in staffi ng in order to increase between low prices and investing labor productivity. in employees by adopting a set of EMPOWER EMPLOYEES to make operational practices: small on-the-spot decisions.

term, and difficult to measure. Home Depot is a that this could hurt sales and profits. Indeed, my well-known example. When former GE executive research suggests that understaffing retail stores Robert Nardelli became CEO, at the end of 2000, he amounts to a missed opportunity: In my analysis of cut sta ng levels and increased the percentage of data from 1999 through 2002 from more than 250 part-timers to reduce costs and boost pro ts. Those stores of Borders, a major bookstore chain at the moves achieved both goals immediately, but they time, I found that a one-standard-deviation increase eventually caused Home Depot’s excellent cus- in labor levels at a store increased pro t margins by tomer service—the company’s claim to fame and, 10% over the course of a year. Research by Marshall arguably, primary source of competitive advan- Fisher, Serguei Netessine, and Jayanth Krishnan tage—to su er, customer satisfaction to plunge, and supports my ndings: Their analysis of 17 months same-store sales growth to drop and even go nega- of data from a large retailer shows that for every $1 tive in some years. increase in payroll, a store could see a $4 to $28 in- What happened to Home Depot is common. crease in monthly sales. Many store managers at various retailers told me Of course the relationship between sta ng levels that the pressure to meet short-term performance and pro tability is not linear: After a certain point, in- targets led them to reduce employees even though creasing the former will reduce the latter. But instead they knew that the workers who remained would of responding to short-term pressures by automati- cut corners and make mistakes. And they suspected cally cutting labor, stores should strive to find the

TOO MANY U.S. RETAIL WORKERS HAVE BAD JOBS SCHEDULES EFFECT ON SOCIETY In addition to poor wages and Employers’ underinvestment benefi ts, retail employees have in retail employees is costly unpredictable work schedules. for society. Retail employees - to - receive disproportionately In an eff ort to match staffi ng hour shifts are common, and more public assistance than levels to customer traffi c, retail employees are often asked employees in other industries. chains schedule employees only to be on call. They are clearly on the losing Changing shifts on such short end of the large income gap in week in notice can make it diffi cult, the United States. advance, if not impossible, for workers to meet family commitments, and even those schedules can change at the last minute. perform other jobs, and arrange for child care. It also can wreak havoc on increas- ingly fragile family budgets.

January–February 2012 Harvard Business Review 127 WHY “GOOD JOBS” ARE GOOD FOR RETAILERS

RETAILING’S VICIOUS CYCLE

LOW LABOR BUDGETS

LOW SALES sta ng level that maximizes pro ts on a sustained to make trade-offs between dimensions of perfor- AND LOW PROFITS QUALITY basis. In many cases, that will mean adding workers. mance: Should they answer a customer’s question if AND OR QUANTITY Retailers do not just underinvest in the quantity that keeps them from restocking a popular product? OF LABOR of labor. They treat the quality of labor the same Should they go looking for something a customer POOR OPERATIONAL way—paying low wages, offering insufficient ben- can’t nd if that prevents them from putting up next EXECUTION e ts, and providing inadequate training. The short- week’s promotion? term pressures are just too di cult to resist. The in- When these nitty-gritty, ongoing operational is- evitable consequences are understa ed stores with sues are handled by low-paid employees at under- high turnover of low-skilled employees who are of- sta ed stores, the consequences for operational ex- RETAILING’S VIRTUOUS CYCLE ten part-timers and have little or no commitment to ecution can be severe. their work. HIGH LABOR Retailing’s Vicious and Virtuous Cycles BUDGETS The Eﬀ ect on Store Operations Extensive research in operations management links HIGH Let’s look closer at what happens when, say, a su- employee turnover and poor training to poor perfor- SALES permarket manager cuts staffing to meet a payroll mance, especially in manufacturing settings. The AND GOOD PROFITS QUALITY or pro t target. A typical supermarket is a complex same is true in retailing. When my colleague Ananth AND QUANTITY operating environment. It carries close to 39,000 Raman of Harvard Business School and I rst started OF LABOR SKUs, ranging from an Idaho potato to a 6.4-ounce working with Borders, we found that there was a GOOD OPERATIONAL tube of Crest fluoride anticavity toothpaste with huge variation in operational performance among EXECUTION tartar protection. The store receives multiple de- stores that used the same information technology liveries every day from manufacturers and its own and o ered the same incentives to employees. The distribution centers, and store employees shelve performance of the best store was a whopping 43 much of the merchandise. It has about 100 promo- times better than that of the worst store. Part of this tions a week and serves close to 2,500 customers a variation, we found, could be explained by labor day. Customer tra c uctuates throughout the day practices. Stores in which employees had less train- and week and on holidays, but the uctuations are ing, greater workloads, and higher turnover per- fairly predictable. formed worse. In this environment, it takes a lot of operational That is not surprising. Operational execution re- expertise to get the right product on the right shelf quires people. So stores with a gap in people—too at the right time. Like most retailers, supermarkets few employees or unmotivated or incapable employ- carry more goods than they can display. So there is ees—will have a gap in operational execution. But not only a constant unloading of deliveries but also few retailers realize the seriousness of operational a constant shifting of items from backrooms to the problems and how much money they lose by under- selling oor and back again. It also takes a lot of logis- investing in employees. tical e ort to get the wrong product o the shelf: For In grocery retail, for example, close to a third of example, items move to storage when promotions stockouts are so-called phantom stockouts—the sup- end. Store employees are also supposed to remove ply chain gets the products to the right store, but cus- damaged or expired goods—generally more than 1% tomers cannot nd them because the products are in of what is on the shelves. the wrong place. For a specialty retail chain we looked In my eld visits, I consistently found that with at, the gure was 60%. Two surveys at Borders stores so many products, promotions, and storage areas, a showed that one out of six customers who asked a task that ought to be simple—such as shelving tooth- salesperson for help finding something had experi- paste—is not. Such a surprisingly complex operation enced a phantom stockout. Misplaced products obvi- requires something uniquely human: judgment. ously lead to lost sales. We estimated that the com- Poorly paid, poorly trained, and poorly motivated pany’s pro ts would have been 25% higher without employees have to monitor which products have sold, phantom stockouts. Misplaced products also frus- decide what to keep on the selling oor and what to trate customers and reduce employees’ productivity. move to and from backrooms, and remember which Underinvestment in employees also helps explain backrooms contain which items. retail stores’ poor compliance on agreements with All the while, shoppers are asking questions, manufacturers about promotions. Manufacturers which requires employees to use their judgment spend millions of dollars planning promotions, but

128 Harvard Business Review January–February 2012 HBR.ORG

COSTCO MODEL PRACTICE Dead-end jobs are the norm at most retailers, but Costco a 2008 study by the In-Store Implementation Share- provides employ- Employees of these retailers have higher pay, group, an industry organization, found that about ees with advance- fuller training, better bene ts, and more-convenient half are executed either late or not at all. ment opportunities: schedules than their counterparts at the competition. Less obvious but hardly less serious is the way Around 98% of store Store employees earn about 40% more at Costco managers are pro- such problems distort point-of-sale data, which re- moted from within. than at its largest competitor, Walmart’s Sam’s Club. sults in poor inventory and promotion planning. For At Trader Joe’s, the starting wage for a full-time em- example, when a customer experiences a phantom Sales per employee ployee is $40,000 to $60,000 per year, more than stockout, inventory records show positive inven- at Costco are almost twice what some competitors o er. The wages and tory for the product, and point-of-sales data show double those at bene ts at QuikTrip are so good that the chain has Sam’s Club. that the product did not sell. The forecasting system been named one of Fortune’s “100 Best Companies then concludes that there is no demand and reduces to Work For” every year since 2003. All of Merca- the forecast of future demand, so the retailer will dona’s employees are permanent, and more than stock less or even none of that product. Today’s op- $ 85% are salaried full-timers. erational problems a ect not only today’s sales and COSTCO These model retailers make an e ort to provide SALES PER pro ts but also tomorrow’s. SQUARE FOOT advancement opportunities. For example, about Even Walmart, the poster child of big-box retail- 98% of store managers at Costco and all store man- ing and supply-chain management, has struggled agers at Mercadona, QuikTrip, and Trader Joe’s are with these problems, which is one reason it began $ promoted from within, and many executives at putting RFID tags on some merchandise. But such SAM’S CLUB these companies started out in the stores. SALES PER technologies are often an expensive way to solve SQUARE FOOT Not surprisingly, employee turnover is low. Quik- people and process problems. Because labor budgets SOURCE COSTCO AND WALMART Trip’s 13% turnover rate among full-time employees ANNUAL REPORTS; at many retail chains are set as a percentage of sales, WAYNE CASCIO, ACADEMY OF is substantially lower than the 59% average rate in they take a hit when sales drop. When the labor bud- MANAGEMENT PERSPECTIVES, the top quartile of the convenience store industry. At get is low, store managers cannot increase sta ng Trader Joe’s, turnover among full-time employees is levels, even when they know it will make the store less than 10%. At Mercadona, it’s a mere 4%. Turn- more pro table. And retail chain managers are hesi- over for employees who stay at Costco for more than tant to invest in employee training or other bene ts a year is 5.5%. that increase retention—and boost sales. The vicious In addition to o ering the lowest prices in their cycle continues. industries, these retailers also provide better customer service than their competitors. The University Operating in a Virtuous Cycle TRADER JOE’S of Michigan’s American Customer Satisfaction Index When retailers view labor not as a cost to be mini- MODEL PRACTICE ranks Costco as high as Nordstrom—a department mized but as a driver of sales and pro ts, they create The starting wage for store chain known for outstanding customer ser- full-time employees a virtuous cycle. Investment in employees allows at Trader Joe’s is vice—and consistently higher than Sam’s Club. Quik- for excellent operational execution, which boosts $40,000 to $60,000 Trip performs better than its competitors in evalu- sales and pro ts, which allows for a larger labor bud- a year, more than ations by mystery shoppers. Customers get in and get, which results in even more investment in store twice what some out of QuikTrip stores quickly because merchandise employees. competitors oﬀ er. is always where it is supposed to be, and employees I recently studied four low-price retailers that Sales per labor hour have been trained to ring up three customers per operate in a virtuous cycle: Mercadona, the largest are more than 40% minute (often by not having to scan merchandise supermarket chain in Spain, with more than 1,300 higher than those and by calculating change in their heads). stores and €16 billion in sales; QuikTrip, a U.S. con- of an average U.S. American retail customers have become resigned venience store chain with more than 540 stores and supermarket. to the notion that if they want the lowest prices, they $8 billion in sales; Trader Joe’s, an American super- TRADER JOE’S can’t expect much, if any, sales assistance. Trader SALES PER market chain with more than 340 stores and $8 bil- SQUARE FOOT ARE Joe’s and Mercadona don’t accept that. Their em- MORE THAN lion in sales; and Costco, a wholesale-club chain with ployees constantly engage shoppers in conversation more than 580 stores and $76 billion in sales. These and inform them about new products. Many Merca- retailers are highly regarded by customers and in- dona employees know customers by name and can xTHOSE AT AN AVERAGE dustry peers. In addition to healthy sales and pro t U.S. SUPERMARKET. make personal recommendations. Trader Joe’s em- growth, they have substantially higher asset and la- ployees are known for suggesting products and reci- bor productivity than their competitors. pes. In fact, Consumer Reports ranked Trader Joe’s

January–February 2012 Harvard Business Review 129 WHY “GOOD JOBS” ARE GOOD FOR RETAILERS

QUIKTRIP MODEL PRACTICE While most retailers try to make do with fewer workers, as the second-best supermarket chain in the United QuikTrip does the ries, both o er fewer choices within their chosen States after Wegmans, which is known for outstand- opposite and categories than rivals do. QuikTrip o ers only high- ing labor practices but does not compete on the basis maintains a force of demand products. hundreds of fl oaters of low prices. Do customers mind limited options? Sales per who can fi ll in for employees who get square foot at these stores suggest that they do not. Breaking the Trade-Off sick, are on vaca- With fewer products, employees can be familiar Mercadona, QuikTrip, Costco, and Trader Joe’s do tion, or have an with everything the store sells and make knowledge- not expect the virtuous cycle to operate on its own. emergency. able recommendations to customers. Trader Joe’s is They complement their investment in employees famous for this. At Mercadona stores, each section QuikTrip’s sales per with operational practices that make the execution labor hour are 66% is managed by a specialist who will gladly explain to of work more efficient and more fulfilling for em- higher than those of shoppers why Mercadona does or does not carry par- ployees, lower costs and improve service for custom- an average conve- ticular products. This is one reason the company felt ers, and boost sales and pro ts for the retailer. These nience store chain it could further reduce product variety to cope with and 50% higher practices allow retailers to break the presumed the recent economic crisis. Its con dence was borne than those of the top trade-o between investing in employees and main- quartile convenience out: Sales went up because simplifying operational taining low prices. store chain. execution allowed Mercadona to reduce prices even Off er less. In an e ort to o er more to custom- more and allowed employees to explain to customers, retailers tend to make choices that increase $ ers why they were getting a better deal. the complexity of their operations—sometimes Achieve flexibility by cross-training em- QUIKTRIP without realizing it. One such choice is high prod- SALES PER ployees. At retail chains that operate in a vicious SQUARE FOOT uct variety. A typical supermarket carries myriad cycle, changes in customer traffic lead to changes types and sizes of toothpaste, for example. From an in the number of employees. These retailers fol- operations perspective, high product variety adds $ low what Harvard Business School’s W. Earl Sasser costs up and down the supply chain. It increases Jr. dubbed a “chase-demand strategy” in a Novem- AVERAGE manufacturing costs and supply-demand mismatch CONVENIENCE ber 1976 HBR article, “Match Supply and Demand costs: The more types of toothpaste retailers stock, STORE CHAIN in Service Industries.” But although fluctuations SALES PER the harder it is to predict demand for any particular SQUARE FOOT in customer tra c tend to be fairly predictable (at type and the more inventory retailers end up hold- one retailer, I found that more than 90% could be ing. And as I mentioned above, high product variety explained by day of the week, time of day, weather, makes the operating environment more complex for $ and holidays), employees’ schedules are not. Work- store employees. For all that, it doesn’t necessarily TOP QUARTILE ers get very short notice of changes and are often increase sales. CONVENIENCE asked to shorten their shifts. STORE CHAIN Stores also can o er too many promotions. Varia- SALES PER A lot of retailers consider this to be an e cient SQUARE FOOT tions in customer demand caused by promotions approach but do not take into account its true costs. and forward buying (the purchasing by retailers or Home Depot’s “ exible” approach in the early 2000s wholesalers of promoted products that they hope to involved replacing knowledgeable full-time em- sell at full price, after the promotion has ended) lead ployees with part-timers who did not know as much to ine ciency and waste in the supply chain. Pro- about home improvement and the store’s products motions also increase employees’ workloads and and so could not help customers e ectively. reduce labor productivity, as we have seen. Not surprisingly, I found that unpredictable Retailers that operate in a virtuous cycle, by schedules, short shifts, and dead-end jobs take a contrast, make choices that simplify their op- toll on employees’ morale. When morale is low, ab- erations. They consistently offer “everyday low senteeism, tardiness, and turnover rise, increasing prices” rather than a kaleidoscope of promotions, the variability of the labor supply, which, of course, and they carry fewer products. Mercadona, for in- makes matching labor with customer traffic more stance, carries around 8,000 SKUs and Trader Joe’s difficult. In addition, retailers with high turnover and Costco only about 4,000 (compared with the cannot a ord to invest in employee training; aver- supermarket industry average of approximately age training per new retail employee is a mere seven 39,000). Although Mercadona has positioned itself hours in the United States. Untrained or poorly as a one-stop shop with a broad array of categories trained employees are less productive and make and Trader Joe’s competes only in certain catego- more errors.

130 Harvard Business Review January–February 2012 HBR.ORG

MERCADONA MODEL PRACTICE Most retailers infl ict unpredictable schedules on their Instead of varying the number of employees to employees, but also has eld employees whose main job is to relay match traffic as much as other retailers do, Quik- Mercadona trains employee and customer feedback to purchasing and Trip and Mercadona vary what employees do. They workers in a variety marketing departments. of tasks (managing achieve this by training employees to perform a va- In contrast to retailers that constantly strive to a section, performing riety of tasks. At QuikTrip, part-time employees re- inventory checks, make do with fewer employees, retailers that oper- ceive 40 hours of training and full-time employees ordering merchandise, ate in a virtuous cycle often err on the side of over- receive two weeks—not just in checking out custom- replenishing prod- staffing. They want to make sure that employees ers but also in brewing co ee, ordering merchandise, ucts) so that they are not too rushed to serve customers well and n- can vary what they sweeping oors and the parking lot, cleaning bath- ish their logistics tasks. QuikTrip goes even further, do with changes in rooms, and stocking coolers, freezers, and grills. customer traffi c. maintaining a force of hundreds of employees who At Mercadona, every new employee receives four do not report to a speci c store but are ready to ll weeks of training, during which they learn how to Mercadona’s sales in for people who get sick, take a vacation, or have manage a particular section (meat or cosmetics, for per employee are 18% an emergency. higher than those example), perform inventory checks (for data accu- Let employees make small decisions. In most of other Spanish racy), order merchandise, replenish products from supermarkets and retail stores, merchandise planning is centralized backrooms, and check for product defects or other 46% higher than and only managers can make decisions about prod- problems. When customer traffic is high, employ- those of an average uct returns and customer complaints. But at com- ees at QuikTrip and Mercadona focus on customer- U.S. supermarket panies that operate in a virtuous cycle, employees in 2008. related tasks; when tra c is low, they focus on other constantly make decisions. QuikTrip, Trader Joe’s, tasks. QuikTrip employees also can move from one and Mercadona employees decide how many units store to another, because all stores have the same ~$ of each item to order for their stores. How can large design. , chains trust thousands of people to make inventory MERCADONA As a result of this cross-training, employees have SALES PER decisions? Every decision is small, corporate IT is de- more-predictable schedules and are always busy SQUARE FOOT signed to assist, and the decisions are monitored. Be- (that is, more productive), and customers get faster cause empowering employees in these ways makes service from more-knowledgeable employees. $ companies more responsive to local needs and pref- Eliminate waste in everything but staffi ng. erences, it increases customer as well as employee CARREFOUR Retailers that invest in employees are by no means SALES PER satisfaction. SQUARE FOOT easygoing about what people do. Rather, they are obsessed with eliminating waste and improving e - We’ve Seen This Before ciency. At Costco stores, products are shelved on pal- Several decades ago, there was an intense debate lets, which eliminates the need to unload and shelve about whether it was possible for low-cost products them. At Trader Joe’s, many perishable products are to be high quality. Many academics and practition- sold already packaged instead of loose, which speeds ers argued that investing in quality would increase up checkout. Costco and Trader Joe’s also work hard costs. But some companies, starting with Toyota, to eliminate waste in the supply chain—by, for exam- showed that this was a false trade-o : Investing in ple, purchasing most products directly from manu- people and processes actually drove quality up and facturers and moving them to retail stores via their costs down. own highly e cient distribution centers. Today many retail managers believe that there is QuikTrip and Mercadona apply world-class man- a trade-o between investing in employees and of- ufacturing practices to their store operations. Every fering the lowest prices. That, too, is false. Retailers in-store logistics process—from receiving merchan- that persist in believing in it forgo the opportunity dise to moving products within the store—is timed to improve their own performance and contribute and standardized, and compliance with the stan- the kind of jobs the U.S. economy urgently needs. dards is constantly monitored. Employee feedback is When backed up with a specific set of operating incorporated into process design and improvement. practices, investing in employees can boost cus- At QuikTrip, employees from every position regu- tomer experience and decrease costs. Companies larly discuss problems and identify opportunities can compete successfully on the basis of low prices for improvement. At Mercadona, managers at head- and simultaneously keep their customers and em- quarters in charge of specific processes routinely ployees happy. visit stores and talk to employees. The company HBR Reprint R L

January–February 2012 Harvard Business Review 131 Harvard Business Review Notice of Use Restrictions, May 2009

Harvard Business Review and Harvard Business Publishing Newsletter content on EBSCOhost is licensed for the private individual use of authorized EBSCOhost users. It is not intended for use as assigned course material in academic institutions nor as corporate learning or training materials in businesses. Academic licensees may not use this content in electronic reserves, electronic course packs, persistent linking from syllabi or by any other means of incorporating the content into course resources. Business licensees may not host this content on learning management systems or use persistent linking or other means to incorporate the content into learning management systems. Harvard Business Publishing will be pleased to grant permission to make this content available through such means. For rates and permission, contact [email protected].