Insights

Fewer hours, same customers: what research says about understaffing in retail

When labour costs rise, store hours are usually the first line to shrink. Operations research shows that an hour cut at the peak costs more than one cut in a quiet spell, and that the cost differs from store to store.

5 min read

A single sales assistant at the till while a queue forms in a busy clothing store
On this page

In February 2026 the British Retail Consortium asked retail finance chiefs how they planned to handle rising employment costs. More than half said they would cut staff hours and overtime over the next year, and labour costs had become a top-three concern for 84 per cent of them, up from 21 per cent the previous July. Retailers in other markets face the same arithmetic: wages rise, sales do not keep pace, and store hours are the cost line that can be changed fastest.

The question operations research has studied for more than a decade is a narrower one. If hours have to go, which hours, and in which stores?

Understaffing hides in the peak

Sales data makes understaffing hard to see. A store that was too thin on the floor at 5 pm still records the sales it made, and the visitors who left without being served leave no trace in the till.

Vidya Mani, Saravanan Kesavan and Jayashankar Swaminathan tackled this by combining three hourly data series from 41 stores of a large retail chain: store traffic, sales and labour. Using a model built on queueing theory, they calculated a benchmark staffing level for each store and compared it with what was actually scheduled. Every one of the 41 stores was systematically understaffed during a three-hour peak period, and the authors found that removing that understaffing could significantly increase both sales and profitability. Forecasting errors and scheduling constraints explained part of the gap.

If the peak is already short of staff, an across-the-board reduction takes hours from the period where they earn the most.

More visitors do not automatically mean more sales

An earlier study by Olga Perdikaki and the same two co-authors used data from an apparel chain to split sales into conversion rate (transactions per visitor) and basket value. Sales rose with traffic, but with diminishing returns, and the level of staffing changed how much of the extra traffic turned into sales. Conversion fell as traffic rose. Lower conversion was also associated with weaker traffic growth later on, which suggests that visitors who leave unserved are less likely to come back.

That gives a practical warning sign. When conversion drops in the busiest hours of the week, the store is probably turning visitors away, whatever the sales total says. Footfall and sales tell different stories, and the gap between them is where staffing shows up.

The same hour is worth more in some stores

Many chains set labour budgets in proportion to forecast sales. Marshall Fisher, Santiago Gallino and Serguei Netessine questioned that rule. Working with a large specialty retailer, they estimated how revenue responded to staffing in each store, using random gaps between planned and actual labour, and then put the store-specific staffing levels into practice. Across 168 stores over six months, revenue rose by 4.5 per cent.

The revenue effect of labour varied considerably between stores, and was largest where potential demand was high, competition was close by and the store manager was experienced. A uniform cut of ten per cent per store therefore costs very different amounts in different stores.

But do not chase every peak

Two studies set limits on how closely the roster should track the footfall curve.

Kesavan, Bradley Staats and Wendell Gilland studied two years of data from 445 stores and found an inverted U between the share of temporary or part-time staff and sales and profit. Some flexibility helped, but too much of it hurt.

At Gap, Kesavan and colleagues ran a randomised field experiment in 28 stores in San Francisco and Chicago, making schedules more consistent, predictable and adequate, and giving employees more control over them. Store productivity rose by 5.1 per cent, from 3.3 per cent higher sales and 1.8 per cent fewer labour hours. The authors trace the effect to better store execution. Store work is demanding, and a roster that changes every week adds to the strain.

What the research adds up to

Taken together, the studies argue against cutting the same share of hours everywhere. Hours are best moved before they are removed, because the peak is where understaffing costs most and the quiet hours are where wage cost is easiest to recover. Cuts work better store by store, since the same hour of labour earns more in some stores than in others. A steady roster can cover the quiet floor, with the flex planned around peaks that recur instead of every new forecast. And a falling conversion rate in the busiest hours is the clearest sign that a cut has gone too far.

Historical traffic goes a long way. In a study of an apparel chain, Hsiao-Hui Chuang, Rogelio Oliva and Olga Perdikaki built a staffing rule that used past traffic to set labour. In simulations it came within 0.5 per cent of the profit possible with perfect traffic information when traffic was stable, and within 2.5 per cent when it varied strongly.

Measuring visitors against staffed hours

All of these studies rest on the same data: visitors per hour next to staffed hours per hour, store by store. Most retailers have the second series and the sales, but not the first.

Anonymous footfall counting fills that gap. With hourly visitor counts per store, the ratio of staff to visitors becomes visible, the recurring peaks stand out from one-off spikes, and the effect of a roster change can be checked the following week. Staff planning with footfall data shows how that works in practice, and the weekly rhythm of footfall looks at the pattern behind the roster. The measurement uses anonymous, aggregated statistics from a method approved by a data protection authority.

Sources

  • Mani, V., Kesavan, S., & Swaminathan, J. M. (2015). Estimating the impact of understaffing on sales and profitability in retail stores. Production and Operations Management, 24(2), 201–218. https://doi.org/10.1111/poms.12237
  • Perdikaki, O., Kesavan, S., & Swaminathan, J. M. (2012). Effect of traffic on sales and conversion rates of retail stores. Manufacturing & Service Operations Management, 14(1), 145–162. https://doi.org/10.1287/msom.1110.0356
  • Fisher, M., Gallino, S., & Netessine, S. (2021). Setting retail staffing levels: A methodology validated with implementation. Manufacturing & Service Operations Management, 23(6), 1562–1579. https://doi.org/10.1287/msom.2020.0917
  • Kesavan, S., Staats, B. R., & Gilland, W. (2014). Volume flexibility in services: The costs and benefits of flexible labor resources. Management Science, 60(8), 1884–1906. https://doi.org/10.1287/mnsc.2013.1844
  • Kesavan, S., Lambert, S. J., Williams, J. C., & Pendem, P. K. (2022). Doing well by doing good: Improving retail store performance with responsible scheduling practices at the Gap, Inc. Management Science, 68(11), 7818–7836. https://doi.org/10.1287/mnsc.2021.4291
  • Chuang, H. H.-C., Oliva, R., & Perdikaki, O. (2016). Traffic-based labor planning in retail stores. Production and Operations Management, 25(1), 96–113. https://doi.org/10.1111/poms.12403
  • British Retail Consortium survey of retail finance directors, February 2026, as reported in Retail Gazette. https://www.retailgazette.co.uk/blog/2026/02/staffing-cull-on-the-horizon-for-brc

Curious what this looks like for your venue?

Book a demo See what data you get

Frequently asked questions

Does understaffing actually cost sales?

Research says yes, and mostly at the peaks. A study of 41 stores in one retail chain combined hourly traffic, sales and labour data and found that every store was systematically understaffed during a three-hour peak, and that removing the understaffing would raise both sales and profit. The hours that matter most are the ones where visitors are most numerous.

Should staffing follow footfall hour by hour?

Staffing to traffic helps, but research also shows limits. Too high a share of temporary and part-time labour was linked to lower sales and profit in a study of 445 stores, and a randomised trial at Gap found that more stable, predictable schedules raised store productivity. The practical answer is a stable base roster that covers the quiet floor, with planned flex on the recurring peaks.

Why not just staff in proportion to forecast sales?

Because the same hour of labour is worth different amounts in different stores. A study that rolled out store-specific staffing levels across 168 stores found that the revenue effect of labour varied considerably, and was largest in stores with high potential demand, nearby competition and experienced managers. A uniform rule, applied to every store, leaves revenue on the table in some and wastes wage cost in others.

What data do I need to see where hours are missing?

Visitors per hour next to staffed hours per hour, per store. Sales alone cannot show it, because the till only records the visitors who bought. Footfall shows the visitors who came in and left without buying, which is where understaffing hides.

See your peaks next to your roster

Book a 30-minute walkthrough and see how hourly footfall data shows where staffed hours and visitors fall out of step.

Book a demo