Business Review

Trina Dalziel/Getty Images

The relentless rise of online retailers has led to deep soul searching among brick-and-mortar retailers to find ways to compete. The traditional methods of competing through convenience, assortment, and pricing are largely ineffective against online retailers who outperform brick-and-mortar retailers in these dimensions. The last arrow in the quiver is to use service as a way to distinguish themselves from online retailers. Yet, research suggests that retailers tend to view store associates as an expense to be controlled rather than as a medium to provide better service for customers.

Practices such as having barebones staff in stores and unstable scheduling (schedules that vary on a day-to-day basis) have flourished in the guise of enabling greater profits for retailers. In study after study for over a decade, operations researchers have found that retailers understaff during peak hours. Increasing staffing, they found, could increase sales and profits. And yet this message on the costs of lean scheduling fell on deaf ears.

Our goal, in a randomized controlled experiment at Gap, was to shift retail associates to more-stable schedules and study the business results. The interdisciplinary team was led by Principal Investigator Joan C. Williams, Co-PI Susan Lambert of the University of Chicago, and Co-PI Saravanan Kesavan of the Kenan-Flagler Business School, University of North Carolina at Chapel Hill.

In retail today, most associates are part time, and part-time schedules typically change every day and every week, often with only three days’ notice of next week’s schedule. This was the way scheduling worked at Gap when we began working with them. But after our eight-month pretest in three stores, Gap made two important changes in all U.S. stores:

• Eliminating “on-calls.” On-calls are when employees are scheduled to work shifts that can be canceled anytime up until two hours before they are scheduled to begin.
• Requiring employee schedules to be posted two weeks in advance.

Once the full experiment was launched in November 2015, 28 stores in the San Francisco Bay Area and Chicago were randomly assigned to control and intervention groups. In intervention stores, store managers who chose to participate committed to trying out the two changes above, plus five additional changes:

• Giving a core team of associates a “soft guarantee” or 20 or more hours a week.
• Establishing standard start and end times for shifts.
• Giving more associates a stable core schedule (meaning that associates will have a consistent schedule from week-to-week).
• Using the mobile app Shift Messenger, in which associates could swap shifts on their own without getting their supervisor’s approval. Managers could also use the app to post additional shifts.
• Receiving additional staffing during understaffed periods, which were identified based on analysis of store traffic and conversion rate data. These extra hours were not part of the manager’s labor budget and were only given to the stores that were identified as likely to increase their sales by Co-PI Kesavan’s analysis.

More-stable scheduling increased sales and labor productivity

The results were striking. Sales in stores with more stable scheduling increased by 7%, an impressive number in an industry in which companies work hard to achieve increases of 1–2%. Labor productivity increased by 5%, in an industry where productivity grew by only 2.5% per year between 1987 and 2014. Our estimate is that Gap earned $2.9 million as a result of more-stable scheduling during the 35 weeks the experiment was in the field. Given that out-of-pocket expenses were small ($31,200), our data suggest that return on investment was very high. (If stable scheduling were adopted enterprise-wide, transition costs might well entail the costs of upgrading or replacing existing software systems.)

Unlike the typical way of driving sales through increase in traffic, the sales increase from our intervention occurred due to higher conversion rates and basket values made possible through better service from associates. In other words, the sales increase did not occur due to syphoning of traffic and sales from other Gap stores as the stores that had adopted the new policies became attractive to shoppers. So, this augurs well for the rollout of this initiative to the entire chain as a similar increase can be expected as the risk of cannibalization is low.

The conventional wisdom is that lean, unstable scheduling is inevitable in today’s fast-paced, low-profit, brick-and-mortar environment. It isn’t. Our store-level intervention produced a modest shift towards more stable schedules that increased three dimensions of schedule stability: schedules in intervention stores became more consistent and predictable, and employees gained more control over when they worked. Conventional wisdom also suggests that today’s retail environment requires the use of on-calls. Not so: while some store managers expressed initial anxiety about the elimination of on-calls, virtually all managed just fine without them. One manager noted, “I think some managers saw [on-calls] as a safety net to handle business and workload — but we learned we really didn’t need it.” Similarly, while some managers expressed anxiety about having to post schedules two weeks in advance, 90% managed to do so (excluding three outliers where constant changes in store leadership subverted stores’ ability to comply with company policies).

A related belief is that unstable scheduling is inevitable due to fluctuations in customer demand. In fact, only 30% of fluctuation in schedule instability stemmed from customer traffic.

So more stable scheduling is feasible, and holds the potential for improved sales and productivity. Will these findings on stable scheduling meet the same fate as the decade’s worth of findings on lean scheduling? We hope not. Our study also helped pinpoint the reasons why companies have found it challenging to be able to hear the message that lean and unstable scheduling is not the best business model.

Unstable schedules look good because retailers count their benefits and ignore their hidden costs

Unstable scheduling allows headquarters to point to the money saved on decreasing labor costs, which is important because labor accounts for 85% of controllable costs in retail stores. However, the money saved on achieving a tight fit between labor supply and labor demand is visible but back-end costs to achieve this fit remain invisible.

Our deep reservoir of qualitative data — we did check-ins with participating store managers every two weeks, as well as surveys — gave us a detailed looked into the hidden costs of unstable scheduling, and how a move towards scheduling stability decreased them.

Thirty years of studies have documented poor execution in retail, and we found ample evidence linking poor execution to unstable scheduling. Lean staffing with short-hours part-time staff who have unstable schedules can jeopardize customer service in many different ways, most obviously by leading to long check-out lines or situations where customers can’t find someone to help them get a size or style they need.

Stable, predictable schedules improved customer service in these and other ways. One very simple improvement? Employees found it easier to show up on time, because they were able to more accurately predict their commuting time. Changing company practices so that employees could take the same bus at the same time of day turned out to be a powerful way to improve the customer experience.

Prior research also links unstable scheduling with lower “process conformance” — stores’ conformance to company processes. Store managers linked more stable schedules with a decrease in thefts and an increase in the orderliness of stock, which decrease “phantom stock-outs” (when a customer is told an item is out of stock when it in fact is just out of place). Another hidden back-end cost of unstable schedules was the inordinate amount of time managers spent on the schedule: one manager of a large store decreased time spent from three days to four hours.

No wonder unstable scheduling remains popular: companies count its (front-end) benefits and don’t count its (back-end) costs.

How cognitive bias and agency misattribution underpin unstable schedules

Retailers today invest millions of dollars on marketing campaigns to drive customers into stores but don’t invest in labor planning to ensure that traffic is converted into sales. They do so because poor execution in retail is treated as inevitable, or attributed to the limitations of store managers or staff. What it typically is not attributed to is headquarters.

Wharton Professor Marshall Fisher highlights the role of cognitive bias in sustaining retailers’ faith in unstable scheduling. Humans tend to overweight what’s immediate and easy to measure—the payroll checks a retailer write every week to its stores’ staffs—and to underweight what’s longer-term and harder to measure—the costs of unstable scheduling. “This opens the door to self-delusion,” notes Fisher. “Retailers can convince themselves that they can cut payroll by 5% in the last three weeks of a quarter to meet their profit promise to Wall Street and it really won’t impact customer service, because there’s probably people in the stores not doing anything anyway.”

Headquarters’ incentive is to make its numbers in the short term. Store managers’ incentive is to run their stores to sell more in the long term. When departments have different incentives, “agency misattribution” often results: coordination failures are attributed to personality drawbacks in another unit, instead of recognizing that the other department just has different information.

At the start of the study, we often heard HQ fault store managers for “emotional scheduling”—a script pushed by the purveyors of scheduling software. “In measuring customer experience and making decisions related to a labor model, retailers should rely solely on facts. Too often, changes are made because of an anecdotal or emotional response from the field,” notes a best practices guide from Kronos. However, our experiment shows that a hybrid approach of combining algorithms with manager intuition can lead to better staffing decisions. While our experiment provided guidelines for managers, it still allowed the managers to make the final decision on how much of the interventions to implement. The increase in sales and productivity witnessed at the Gap shows that retailers stand to benefit when they allow discretion to store managers.

What would an effective change initiative look like?

Some retailers have announced stable scheduling initiatives to great fanfare, only to have them flounder. Our three years of experience with Gap suggests that launching a stable scheduling initiative is a major undertaking. The good news, and the bad news, is that we found—contrary to popular wisdom—that a lot of schedule instability stems from headquarters.

Store managers consistently identified three ways HQ created schedule instability. Last-minute changes in shipment dates or inaccuracies in the number of units meant that a store received 5,000 units on Thursday instead of the 2,000 units it had been told to expect on Tuesday; this wreaked havoc with posted schedules. Last-minute changes in marketing meant that merchandise had to be marked down, tables set up, stock moved around, signs made, windows changed—and sometimes there were three different promotions in a single week. The third was leadership visits, which meant extra hours were spent ensuring the store looked perfect. To quote one manager, “If payroll is such a commodity, I feel it’s crucial to spend it on the business versus visits. This isn’t a museum. You know we aren’t curating a museum.” All meant that hours had to be added here and subtracted there, after the schedule was posted.

The bad news is that for retailers to implement more stable scheduling, they will need to involve several different complex business processes. The good news is that if headquarters is playing a major role in producing schedule instability, it can play a major role in reducing it.

But the important message is that an effective stable scheduling initiative is not a simple human resources issue. A truly effectively move will require a concerted strategy that starts at the top and includes not just HR but also operations, marketing, and supply chain.

Here’s the bottom line: a shift to more stable schedules is a win-win for retailers and their employees. During a challenging time in the retail industry, Gap made a commitment to its values and it paid off. Retailers would be well-advised to take the initiative to implement more stable scheduling for their associates and improve customer service to effectively combat the threat from online retailers.