Mismatch

The same day is staffed wrongly twice, thin at the peak and ample around it, and staff costs and waiting time arise side by side. In one retail chain, all 41 stores studied were systematically understaffed during the same three-hour peak, while over the rest of the day the plan provided more staff than the sales floor needed. How much revenue footfall turns into depends on the staffing in exactly that hour, as has been measured in retail.

The plan is set on Monday and does not follow the pattern when it shifts. Footfall explains only 30 percent of the fluctuation in weekly hours, the rest comes from the organization, from incorrect delivery information, late promotion changes and visits from headquarters. The morning absence changes capacity before the day begins, and for 36 percent of managers in German chain-store retail, short-notice rescheduling is the most time-consuming task, a quarter still plan with Excel or on paper.

Lead time

Cortex projects footfall, queue and dwell time per zone for the next hour and places them next to the staffing that the plan provides for that hour. The area that is getting tight is thus known before it is tight, and so is the zone in which customers wait for advice before they leave the area without buying.

What belongs in the troughs are breaks, shelf stocking and checks, what belongs in the peak are checkout, advice and sales floor. Cortex places the break windows around the three hours that carry the day and holds back the tasks that can wait for the hours in which no one is waiting.

Allocation

Cortex compares the demand per zone with attendance, roles and skills and names the bottleneck with the highest effect, one task, one role, one reason. After an absence, this yields a sequence, what is secured, what waits and who takes it on, and duplicated tasks are resolved while other bottlenecks remain open.

Tasks reach the right person bundled and not as a shout across the sales floor, so that paths get shorter and workflows are not torn apart. Newcomers receive the task that matches their stage of onboarding, and the duty roster itself stays as it is. Cortex shifts capacity for a time window and then returns it to the plan.

Open and close

Before opening, prices, checkouts, goods, technology and cleanliness are a list with status, not a walk-through. Cortex tracks readiness per store as a metric and presents the open items that someone still has to take on before opening, in the order in which they affect the first customer.

The last hour belongs to the customers. Cortex starts shelf stocking, cleaning and area closures only when footfall in the area has ebbed, and carries open cases into the handover, so that the next shift does not duplicate work and nothing is left undone.

Impact

Revenue per staff hour, service level, overtime and open tasks are measured in the same grid in which planning took place. That staffing and demand belong together has been proven, because reliable schedules raised store productivity by 5.1 percent in a randomized field experiment, and reallocating the existing staff budget between stores leads a retailer with over 500 stores to expect two to three percent more revenue at no additional cost.

Cortex writes no duty roster and monitors no one. The approval of every reallocation lies with operations, the set of rules belongs to the location, and across the network it becomes clear which store holds its peaks and where capacity is structurally lacking. Staff deployment thus turns from a weekly plan into a figure that can be managed hour by hour.