Operations Management

MIT Sloan study finds holiday sales could increase with better staffing decisions by retailers


Research shows impact of understaffing on revenue and profitability

CAMBRIDGE, Mass., Dec. 4, 2017––During the holiday season, staffing decisions by retailers are more important than ever. Staffing levels affect store performance, yet labor-related expenses constitute one of the largest components of retailers’ operating costs. To help retailers improve staffing decisions and store performance, MIT Sloan School of Management Visiting Prof. Rogelio Oliva and his colleagues developed a data-driven method that helps determine optimal staffing levels. Testing the method with real data, they found that it could increase sales performance for an apparel retailer by approximately 10%.“The ability to efficiently match store labor with incoming customer traffic is crucial, especially during the holidays when stores expect increased traffic and often rely on year-end sales. But optimizing staffing levels is very challenging, as retail environments are characterized by volatile store traffic, which makes it difficult to provide consistent service quality,” says Oliva.

Traditionally, staffing decisions depend on store budget allocation. He explains, “A typical sales-based staffing rule is to match a constant ratio of expected store sales to the number of store associates. However, this ignores the fact that retail sales are also affected by store traffic and might result in labor-to-traffic mismatches, which can negatively impact revenue. The scheduled labor may not be enough to meet customer traffic flows.”

Further, Oliva points out that this practice relies on past sales to predict future traffic flow. Those sales only include customers who made a purchase, leaving out those who left the store due to lack of service. Prior research shows that 33% of customers who experienced a problem were not able to locate sales help when they needed assistance, and 6% of all possible sales are lost because of lack of service, he notes.

Their new method is unique, as it goes beyond the focus on past sales at individual stores to leverage performance data across different stores within a retail chain. The approach enables retailers to derive aggregate labor requirements by utilizing traffic data, point-of-sale data, and labor data across stores with similar attributes like store format, product mix, and market demographics. 

Using this method to analyze data from an apparel retail chain’s stores in the U.S., the researchers found that store managers were systematically understaffing their stores. “If they even slightly increased staffing levels, they would generate incremental sales that would outweigh the labor costs. This highlights that employees are an important contributor to the sales process. When there aren’t enough sales associates, sales don’t reach their potential,” says Oliva.

The study revealed that the retailer’s stores were achieving 85-95% of their potential sales with current staffing levels. If they used the proposed heuristic to determine staffing levels, they could achieve 99% of potential sales. “We’re talking about a possible 10% increase in sales, which is quite significant,” he says.

Oliva points out that there is a limit to the number of staff that can be added before stores reach a point of diminishing returns. The goal is to use retailers’ data to find that “sweet spot” in the ratio of sales people to customers.

He adds, “The big takeaway is that retailers need to move past the inclination to minimize cost by understaffing stores because it has a big impact on profitability. They could be generating a lot more sales if they staff at the correct level. Stores should staff to maximize sales and profit, not to minimize cost.”

Oliva is a coauthor of “Traffic-based Labor Planning in Retail Stores.”

About the MIT Sloan School of Management

The MIT Sloan School of Management is where smart, independent leaders come together to solve problems, create new organizations, and improve the world. Learn more at